Search Results

Search found 2042 results on 82 pages for 'average'.

Page 76/82 | < Previous Page | 72 73 74 75 76 77 78 79 80 81 82  | Next Page >

• Debian virtual memory reaching limit

  - by Gregor

  As a relative newbie to systems, I inherited a Debian server and I've noticed that virtual memory is very high (around 95%!). The server has been running slow for around 6 months, and I was wondering if any of you had any tips on things I could try, particularly on freeing up memory. The server hosts various websites and also a Postit email server. Here are the details: Operating system Debian Linux 5.0 Webmin version 1.580 Time on system Thu Apr 12 11:12:21 2012 Kernel and CPU Linux 2.6.18-6-amd64 on x86_64 Processor information Intel(R) Core(TM)2 Duo CPU E7400 @ 2.80GHz, 2 cores System uptime 229 days, 12 hours, 50 minutes Running processes 138 CPU load averages 0.10 (1 min) 0.28 (5 mins) 0.36 (15 mins) CPU usage 14% user, 1% kernel, 0% IO, 85% idle Real memory 2.94 GB total, 1.69 GB used Virtual memory 3.93 GB total, 3.84 GB used Local disk space 142.84 GB total, 116.13 GB used Free m output: free -m total used free shared buffers cached Mem: 3010 2517 492 0 107 996 -/+ buffers/cache: 1413 1596 Swap: 4024 3930 93 Top output: top - 11:59:57 up 229 days, 13:38, 1 user, load average: 0.26, 0.24, 0.26 Tasks: 136 total, 2 running, 134 sleeping, 0 stopped, 0 zombie Cpu(s): 3.8%us, 0.5%sy, 0.0%ni, 95.0%id, 0.7%wa, 0.0%hi, 0.0%si, 0.0%st Mem: 3082544k total, 2773160k used, 309384k free, 111496k buffers Swap: 4120632k total, 4024712k used, 95920k free, 1036136k cached PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 28796 www-data 16 0 304m 68m 6188 S 8 2.3 0:03.13 apache2 1 root 15 0 10304 592 564 S 0 0.0 0:00.76 init 2 root RT 0 0 0 0 S 0 0.0 0:04.06 migration/0 3 root 34 19 0 0 0 S 0 0.0 0:05.67 ksoftirqd/0 4 root RT 0 0 0 0 S 0 0.0 0:00.00 watchdog/0 5 root RT 0 0 0 0 S 0 0.0 0:00.06 migration/1 6 root 34 19 0 0 0 S 0 0.0 0:01.26 ksoftirqd/1 7 root RT 0 0 0 0 S 0 0.0 0:00.00 watchdog/1 8 root 10 -5 0 0 0 S 0 0.0 0:00.12 events/0 9 root 10 -5 0 0 0 S 0 0.0 0:00.00 events/1 10 root 10 -5 0 0 0 S 0 0.0 0:00.00 khelper 11 root 10 -5 0 0 0 S 0 0.0 0:00.02 kthread 16 root 10 -5 0 0 0 S 0 0.0 0:15.51 kblockd/0 17 root 10 -5 0 0 0 S 0 0.0 0:01.32 kblockd/1 18 root 15 -5 0 0 0 S 0 0.0 0:00.00 kacpid 127 root 10 -5 0 0 0 S 0 0.0 0:00.00 khubd 129 root 10 -5 0 0 0 S 0 0.0 0:00.00 kseriod 180 root 10 -5 0 0 0 S 0 0.0 70:09.05 kswapd0 181 root 17 -5 0 0 0 S 0 0.0 0:00.00 aio/0 182 root 17 -5 0 0 0 S 0 0.0 0:00.00 aio/1 780 root 16 -5 0 0 0 S 0 0.0 0:00.00 ata/0 782 root 16 -5 0 0 0 S 0 0.0 0:00.00 ata/1 783 root 16 -5 0 0 0 S 0 0.0 0:00.00 ata_aux 802 root 10 -5 0 0 0 S 0 0.0 0:00.00 scsi_eh_0 803 root 10 -5 0 0 0 S 0 0.0 0:00.00 scsi_eh_1 804 root 10 -5 0 0 0 S 0 0.0 0:00.00 scsi_eh_2 805 root 10 -5 0 0 0 S 0 0.0 0:00.00 scsi_eh_3 1013 root 10 -5 0 0 0 S 0 0.0 49:27.78 kjournald 1181 root 15 -4 16912 452 448 S 0 0.0 0:00.05 udevd 1544 root 14 -5 0 0 0 S 0 0.0 0:00.00 kpsmoused 1706 root 13 -5 0 0 0 S 0 0.0 0:00.00 kmirrord 1995 root 18 0 193m 3324 1688 S 0 0.1 8:52.77 rsyslogd 2031 root 15 0 48856 732 608 S 0 0.0 0:01.86 sshd 2071 root 25 0 17316 1072 1068 S 0 0.0 0:00.00 mysqld_safe 2108 mysql 15 0 320m 72m 4368 S 0 2.4 1923:25 mysqld 2109 root 18 0 3776 500 496 S 0 0.0 0:00.00 logger 2180 postgres 15 0 99504 3016 2880 S 0 0.1 1:24.15 postgres 2184 postgres 15 0 99504 3596 3420 S 0 0.1 0:02.08 postgres 2185 postgres 15 0 99504 696 628 S 0 0.0 0:00.65 postgres 2186 postgres 15 0 99640 892 648 S 0 0.0 0:01.18 postgres

  Read the article

• File copying utility like rsync with error handling like ddrescue, for data recovery from a hard drive with bad sectors or hardware failure

  - by purefusion

  I have a hard drive with either bad blocks or sectors that are failing to read due to potential mechanical issues, such as a bad disk head, bad motor, or some other issue that is causing the hard drive to read data excruciatingly slowly and with lots of read errors. I'm seeing an average of 50 KB/sec, with some reads dropping below 10 KB/sec, and frequently it gets stuck on a file or sector altogether, usually for quite a long time—from 2-10 minutes or more (when using rsync, before it times out). Speed seems to vary wildly, and it gets stuck on files a lot, and when it finally gets "unstuck" it only seems to last for a short burst before it gets stuck again. The drive is also very quiet with only an occasional sound of files copying (usually when it gets stuck/unstuck for a brief time, before getting stuck again). Thus, there are none of those evil sounds that are normally associated with HDD death. Someone suggested that the problems sounded like they might be caused by a misaligned disk head, which requires a lot of re-reads before it finally reads data with success. Sounds plausible, but I digress... Anyway, the problem with rsync is that it seems to have no decent error handling support. Obviously, it wasn't meant for use in recovering data from failing hard drives, but all the so-called "data recovery" utilities out there that are meant for such use usually focus on recovery of deleted files or messed up partitions, rather than copying files off dying hard drives. Deleted file recovery is not what I need, obviously, so perhaps you can understand my disappointment in not being able to find what I'm after yet. Naturally, this is where you'd probably say "You should use ddrescue!" Well, that's all fine and dandy, but I've already got most of the data backed up, so I just want to recover certain files. I'm not concerned with trying to recover a full partition block-by-block as ddrescue does. I am only interested in rescuing just specific files and directories. Ideally, what I'd like is some sort of cross between rsync and ddrescue: something that lets me specify source and destination as directories of normal files like rsync (rather than two full partitions as ddrescue requires), with a way to skip files with errors in an initial run, and then allows me to attempt recovery of those files with errors in a later run (with a slightly altered command, of course), perhaps even offering an option to specify the number of retry attempts ...just like how ddrescue works with blocks, only I want a utility that works with specific files/directories like rsync does. So am I daydreaming here, or does something out there exist that can do this? Or, maybe even a way to make rsync or ddrescue work in such a way? I'm really open to whatever solutions might work, so long as they let me choose which files I want to "rescue", and can skip files with errors in the initial run, and try/retry those errors again later. So far I've tried rsync with the following options, but it often gets stuck on a file for longer than the timeout, and ideally I'd just like it to move on to the next file and come back later to the files it gets stuck on. I don't think that's possible though. Anyway, here's what I've been using up till now: rsync -avP --stats --block-size=512 --timeout=600 /path/to/source/* /path/to/destination/

  Read the article

• Where's my memory?! Nginx + PHP-FPM front end webserver slows to a crawl...

  - by incredimike

  I'm not sure if I have a problem with a memory leak (as my hosting company suggests), or if we both need to read http://linuxatemyram.com. Maybe you clever people can help us out? This is a front-end webserver VM running essentially only nginx & php-fpm on RHEL 5.5. This server is powering Magento, a PHP eCommerce thinggy. The server is running in a shared environment, but we're changing that soon. Anyway.. after a reboot the server runs just fine, but within a day it will grind itself into nothingness. Pages will take literally 2 minutes to load, CPU spikes like crazy, etc.. The console is even sluggish when I SSH in. It's like my whole server is being brought to its knees. I've also been monitoring the DB server via top and tcpdumping incoming traffic. The DB stays idle for a good portion of that "slow" load time. When i start seeing queries coming from the front-end server, the page loads soon afterward. Here are some stats after me logging in during a slow-down, after restarting php-fpm: [mike@front01 ~]$ free -m total used free shared buffers cached Mem: 5963 5217 745 0 192 314 -/+ buffers/cache: 4711 1252 Swap: 4047 4 4042 [mike@front01 ~]$ top top - 11:38:55 up 2 days, 1:01, 3 users, load average: 0.06, 0.17, 0.21 Tasks: 131 total, 1 running, 130 sleeping, 0 stopped, 0 zombie Cpu0 : 0.0%us, 0.3%sy, 0.0%ni, 99.3%id, 0.3%wa, 0.0%hi, 0.0%si, 0.0%st Cpu1 : 0.3%us, 0.0%sy, 0.0%ni, 99.7%id, 0.0%wa, 0.0%hi, 0.0%si, 0.0%st Cpu2 : 0.0%us, 0.0%sy, 0.0%ni,100.0%id, 0.0%wa, 0.0%hi, 0.0%si, 0.0%st Cpu3 : 0.0%us, 0.0%sy, 0.0%ni,100.0%id, 0.0%wa, 0.0%hi, 0.0%si, 0.0%st Mem: 6106800k total, 5361288k used, 745512k free, 199960k buffers Swap: 4144728k total, 4976k used, 4139752k free, 328480k cached PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 31806 apache 15 0 601m 120m 37m S 0.0 2.0 0:22.23 php-fpm 31805 apache 15 0 549m 66m 31m S 0.0 1.1 0:14.54 php-fpm 31809 apache 16 0 547m 65m 32m S 0.0 1.1 0:12.84 php-fpm 32285 apache 15 0 546m 63m 33m S 0.0 1.1 0:09.22 php-fpm 32373 apache 15 0 546m 62m 32m S 0.0 1.1 0:09.66 php-fpm 31808 apache 16 0 543m 60m 35m S 0.0 1.0 0:18.93 php-fpm 31807 apache 16 0 533m 49m 30m S 0.0 0.8 0:08.93 php-fpm 32092 apache 15 0 535m 48m 27m S 0.0 0.8 0:06.67 php-fpm 4392 root 18 0 194m 10m 7184 S 0.0 0.2 0:06.96 cvd 4064 root 15 0 154m 8304 4220 S 0.0 0.1 3:55.57 snmpd 4394 root 15 0 119m 5660 2944 S 0.0 0.1 0:02.84 EvMgrC 31804 root 15 0 519m 5180 932 S 0.0 0.1 0:00.46 php-fpm 4138 ntp 15 0 23396 5032 3904 S 0.0 0.1 0:02.38 ntpd 643 nginx 15 0 95276 4408 1524 S 0.0 0.1 0:01.15 nginx 5131 root 16 0 90128 3340 2600 S 0.0 0.1 0:01.41 sshd 28467 root 15 0 90128 3340 2600 S 0.0 0.1 0:00.35 sshd 32602 root 16 0 90128 3332 2600 S 0.0 0.1 0:00.36 sshd 1614 root 16 0 90128 3308 2588 S 0.0 0.1 0:00.02 sshd 2817 root 5 -10 7216 3140 1724 S 0.0 0.1 0:03.80 iscsid 4161 root 15 0 66948 2340 800 S 0.0 0.0 0:10.35 sendmail 1617 nicole 17 0 53876 2000 1516 S 0.0 0.0 0:00.02 sftp-server ... Is there anything else I should be looking at, or any more information that might be useful? I'm just a developer, but the slowdowns on this system worry me and make it hard to do my work.. Help me out, ServerFault!

  Read the article

• What seems to be plaguing my hard drives?

  - by Craig

  In a little bit of a tech nightmare here. I know oodles about software, not so much more than the above average user about hardware. I recently had to toss an old desktop of mine. It was gradually getting slower and slower, and after shutting the desktop down for long periods of time, it would choke up upon startup. Sometimes it'd give a disk read error, sometimes say no OS was found, etc. Restarting it about 5-15 times would eventually boot properly. Weird. I also noticed that startup programs were going missing, Dropbox was reindexing my entire folder, and Backblaze was backing up less than the number of files that it should. This lead me to believe it was probably a hard drive issue. I began to wonder why I'd have hard drive issues, and came to closure when I assumed it was because of recent power surges and outages. I'm sure that does a number on the drive. I bought a new desktop recently. It's not a beast or anything, but it's enough for what I do. It's an eMachines (I know, I know) Ultra-Slim (http://www.amazon.com/eMachines-Ultra-Slim-ER1401-57-Desktop-PC/dp/B00475OG9U). This is ideal for me because it's small and portable. It comes with an AC adapter and battery, like a laptop. Just to be safe, I bought an uninterruptable power supply on top of that. It's basically protected completely from any outages that might scramble the drive. I set this up a few days ago and for the past few days I've been perfecting settings, downloading the usual applications, etc. Two days ago, I noticed Dropbox was reindexing my entire Dropbox folder. I installed both Dropbox and Backblaze on this system, but it is very much more lightweight than the other. Only about 15 third-party applications installed. I thought that maybe Dropbox and Backblaze were stressing my system, so I turned off Backblaze. Still, Dropbox runs and comes across this infinite reindex issue. I noticed that upon a reboot recently, two applications did not start on startup either. Also, much like my old desktop, every 3rd or 4th reboot, I'll be forced into a chkdsk. This makes me incredibly nervous. What could possibly be going wrong with my old, years-old desktop that is immediately causing the same issues to my new one? I've considered all of the basics. I'm in a very air conditioned room. I take run routine virus scans. I'd like to think I take care of my systems very well. What is this issue that is haunting me? There's always the possibility that this new desktop has a junk hard drive, but it just seems way too coincidental.

  Read the article

• 1600+ 'postfix-queue' processes - OK to have this many?

  - by atomicguava

  I have a Plesk 9.5.4 CentOS server running Postfix. I had been having massive problems with the mailq being full of 'double-bounce' email messages containing errors relating to 'Queue File Write Error', but I believe these are now fixed thanks to this thread. My new problem is that when I run top, I can see lots of processes called 'postfix-queue' and have fairly high load: top - 13:59:44 up 6 days, 21:14, 1 user, load average: 2.33, 2.19, 1.96 Tasks: 1743 total, 1 running, 1742 sleeping, 0 stopped, 0 zombie Cpu(s): 5.1%us, 8.8%sy, 0.0%ni, 85.3%id, 0.8%wa, 0.0%hi, 0.0%si, 0.0%st Mem: 3145728k total, 1950640k used, 1195088k free, 0k buffers Swap: 0k total, 0k used, 0k free, 0k cached PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 1324 apache 16 0 344m 33m 5664 S 21.7 1.1 0:03.17 httpd 32443 apache 15 0 350m 36m 6864 S 14.4 1.2 0:13.83 httpd 1678 root 15 0 13948 2568 952 R 2.0 0.1 0:00.37 top 1890 mysql 15 0 689m 318m 7600 S 1.0 10.4 219:45.23 mysqld 1394 apache 15 0 352m 41m 5972 S 0.7 1.3 0:03.91 httpd 1369 apache 15 0 344m 33m 5444 S 0.3 1.1 0:02.03 httpd 1592 apache 15 0 349m 37m 5912 S 0.3 1.2 0:02.52 httpd 1633 apache 15 0 336m 20m 1828 S 0.3 0.7 0:00.01 httpd 1952 root 19 0 335m 28m 10m S 0.3 0.9 1:35.41 httpd 1 root 15 0 10304 732 612 S 0.0 0.0 0:04.41 init 1034 mhandler 15 0 11520 1160 884 S 0.0 0.0 0:00.00 postfix-queue 1036 mhandler 15 0 11516 1120 860 S 0.0 0.0 0:00.00 postfix-queue 1041 mhandler 17 0 11516 1156 884 S 0.0 0.0 0:00.00 postfix-queue 1043 mhandler 15 0 11512 1116 860 S 0.0 0.0 0:00.00 postfix-queue 1063 mhandler 16 0 11516 1160 884 S 0.0 0.0 0:00.00 postfix-queue 1068 mhandler 15 0 11516 1128 860 S 0.0 0.0 0:00.00 postfix-queue 1071 mhandler 17 0 11512 1152 884 S 0.0 0.0 0:00.00 postfix-queue 1072 mhandler 15 0 11512 1116 860 S 0.0 0.0 0:00.00 postfix-queue 1081 mhandler 16 0 11516 1156 884 S 0.0 0.0 0:00.00 postfix-queue 1082 mhandler 15 0 11512 1120 860 S 0.0 0.0 0:00.00 postfix-queue 1089 popuser 15 0 33892 1972 1200 S 0.0 0.1 0:00.02 pop3d 1116 mhandler 16 0 11516 1164 884 S 0.0 0.0 0:00.00 postfix-queue 1117 mhandler 15 0 11516 1124 860 S 0.0 0.0 0:00.00 postfix-queue 1120 mhandler 16 0 11516 1160 884 S 0.0 0.0 0:00.00 postfix-queue 1121 mhandler 15 0 11512 1120 860 S 0.0 0.0 0:00.00 postfix-queue 1130 mhandler 17 0 11516 1160 884 S 0.0 0.0 0:00.00 postfix-queue 1131 mhandler 15 0 11516 1120 860 S 0.0 0.0 0:00.00 postfix-queue 1149 root 17 -4 12572 680 356 S 0.0 0.0 0:00.00 udevd 1181 mhandler 16 0 11516 1160 884 S 0.0 0.0 0:00.00 postfix-queue 1183 mhandler 15 0 11512 1116 860 S 0.0 0.0 0:00.00 postfix-queue 1224 mhandler 16 0 11516 1160 884 S 0.0 0.0 0:00.00 postfix-queue 1225 mhandler 15 0 11516 1120 860 S 0.0 0.0 0:00.00 postfix-queue 1228 apache 15 0 345m 34m 5472 S 0.0 1.1 0:04.64 httpd 1241 mhandler 16 0 11516 1156 884 S 0.0 0.0 0:00.00 postfix-queue 1242 mhandler 15 0 11512 1120 860 S 0.0 0.0 0:00.00 postfix-queue 1251 mhandler 17 0 11516 1156 884 S 0.0 0.0 0:00.00 postfix-queue 1252 mhandler 15 0 11516 1120 860 S 0.0 0.0 0:00.00 postfix-queue 1258 apache 15 0 349m 37m 5444 S 0.0 1.2 0:01.28 httpd When I run ps -Al | grep -c postfix-queue it returns 1618! My question is this: is this normal or is there something else going wrong with Postfix? Right now, if I run mailq it is empty, and qshape deferred / qshape active are empty too. Thanks in advance for your help.

  Read the article

• RedHat 5.5 server does not show per processor memory utilization

  - by Mike S

  I have been searching all over internet but not finding any leads. I have a system with a memory leak that I am trying to troubleshoot. Unfortunately I am not able to see per processor memory utilization. Here are the outputs of TOP and PS commands. Linux SERVER_NAME 2.6.18-194.8.1.el5 #1 SMP Wed Jun 23 10:52:51 EDT 2010 x86_64 x86_64 x86_64 GNU/Linux top - 09:17:13 up 18:43, 3 users, load average: 0.00, 0.00, 0.00 Tasks: 375 total, 1 running, 373 sleeping, 0 stopped, 1 zombie Cpu(s): 0.0%us, 0.0%sy, 0.0%ni,100.0%id, 0.0%wa, 0.0%hi, 0.0%si, 0.0%st Mem: 32922828k total, 32776712k used, 146116k free, 267128k buffers Swap: 5245212k total, 0k used, 5245212k free, 32141044k cached PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 1 root 15 0 10348 744 620 S 0.0 0.0 0:05.65 init 2 root RT -5 0 0 0 S 0.0 0.0 0:00.05 migration/0 3 root 34 19 0 0 0 S 0.0 0.0 0:00.00 ksoftirqd/0 4 root RT -5 0 0 0 S 0.0 0.0 0:00.00 watchdog/0 5 root RT -5 0 0 0 S 0.0 0.0 0:00.00 migration/1 6 root 34 19 0 0 0 S 0.0 0.0 0:00.00 ksoftirqd/1 7 root RT -5 0 0 0 S 0.0 0.0 0:00.00 watchdog/1 8 root RT -5 0 0 0 S 0.0 0.0 0:00.00 migration/2 9 root 34 19 0 0 0 S 0.0 0.0 0:00.00 ksoftirqd/2 10 root RT -5 0 0 0 S 0.0 0.0 0:00.00 watchdog/2 11 root RT -5 0 0 0 S 0.0 0.0 0:00.01 migration/3 12 root 34 19 0 0 0 S 0.0 0.0 0:00.01 ksoftirqd/3 13 root RT -5 0 0 0 S 0.0 0.0 0:00.00 watchdog/3 14 root RT -5 0 0 0 S 0.0 0.0 0:00.00 migration/4 15 root 34 19 0 0 0 S 0.0 0.0 0:00.01 ksoftirqd/4 16 root RT -5 0 0 0 S 0.0 0.0 0:00.00 watchdog/4 17 root RT -5 0 0 0 S 0.0 0.0 0:00.00 migration/5 18 root 34 19 0 0 0 S 0.0 0.0 0:00.00 ksoftirqd/5 19 root RT -5 0 0 0 S 0.0 0.0 0:00.00 watchdog/5 20 root RT -5 0 0 0 S 0.0 0.0 0:00.00 migration/6 % ps -auxf | sort -nr -k 4 | head -10 Warning: bad syntax, perhaps a bogus '-'? See /usr/share/doc/procps-3.2.7/FAQ xfs 6205 0.0 0.0 23316 3892 ? Ss Aug19 0:00 xfs -droppriv -daemon uuidd 6101 0.0 0.0 60976 224 ? Ss Aug19 0:00 /usr/sbin/uuidd USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND smmsp 6130 0.0 0.0 57900 1784 ? Ss Aug19 0:00 sendmail: Queue runner@01:00:00 for /var/spool/clientmqueue rpc 5126 0.0 0.0 8052 632 ? Ss Aug19 0:00 portmap root 99 0.0 0.0 0 0 ? S< Aug19 0:00 [events/1] root 98 0.0 0.0 0 0 ? S< Aug19 0:00 [events/0] root 97 0.0 0.0 0 0 ? S< Aug19 0:00 [watchdog/31] root 96 0.0 0.0 0 0 ? SN Aug19 0:00 [ksoftirqd/31] root 95 0.0 0.0 0 0 ? S< Aug19 0:00 [migration/31] Any help with this is appretiate.

  Read the article

• Monitor randomly shutting down, computer accepting no input, need to restart to get working

  - by Sebastian Lamerichs

  First off, spec list: OS: Windows 7 Ultimate 64-bit SP1 CPU: i7-4820k @ 3.7GHz (stock) GPU: Two 3GB Radeon HD 7970s @ 1.05GHz Mobo: AsRock X79 Extreme6 HDD: 2TB Seagate Barracuda 7200rpm RAM: 16GB quad-channel Kingston 1600MHz PSU: Antec HCG 900W Monitors: Acer S220HQL 1920x1080 + ViewSonic VA2251 1920x1080. Plugged into different GPUs. My problem is that, on a daily-ish basis, my monitors will turn off and not turn back on. My computer will still be running, GPU/CPU/case fans all still going, but the monitors will not turn back on. Additionally, it seems to cease all network activity. It doesn't seem to log any errors at all. I've verified that this is not a monitor issue, as when I press the num/caps/scroll lock buttons on my keyboard, the lights don't change, so the computer is clearly not accepting input. I have noticed a few other people on the internet with this problem, and some have claimed that it was solved by disabling PCI-Express Link State Power Management, but the issue still occurs for me after this. Whilst my CPU and GPUs both run at 100% 24/7, the temperatures are certainly not at dangerous levels, with the CPU averaging 65°C and the GPUs at 70°C and 78°C average. All components are brand new. I have tried forcing MSI Afterburner to start when Windows starts and to force a constant voltage, as this fixed the issue for a few days for another user, but he reported back saying that it had stopped working properly again, so I'm not putting too much faith in this working. Many people have said to adjust display sleep mode settings, but this will clearly not work, as the keyboard lights would still work if the monitors were the issue. The closest I can get to a log file for this issue is the following Folding@Home logs: 14:45:21:WU01:FS00:0x17:Completed 1120000 out of 2000000 steps (56%) 14:46:43:WU00:FS01:0x17:Completed 480000 out of 2000000 steps (24%) 14:46:49:WU01:FS00:0x17:Completed 1140000 out of 2000000 steps (57%) 14:48:30:WU01:FS00:0x17:Completed 1160000 out of 2000000 steps (58%) 14:49:55:WU01:FS00:0x17:Completed 1180000 out of 2000000 steps (59%) As you can see, the second GPU (FS01) stops computation approximately three and a half minutes before the issue occurs (it should be completing 1% every 80-120 seconds), and the first GPU (FS00) continues for a few minutes more before the logs just end. As far as I can tell, the computer has a network failure at the time the first GPU stops working, the latest IRC message I received from this time was at 14:47:58. That being said, there could have just not been any messages between then and 14:50:00, so I'm going to be connecting a laptop to the same bouncer to double-check if it happens again. The GPUs functioned perfectly well in another computer for a significant period of time, so I'm fairly confident that they aren't the issue, which means that this is being caused by either software or the motherboard, or possibly RAM. I really hope it's software. I heard from a forum board that there was a patch from Microsoft that fixed this problem, but "I've forgot which KB it was or the google search terms I used to find the patch, LOL.", so that's not much help. Haven't seen it mentioned by anyone else on about a dozen threads about this issue either. The computer is plugged in via a surge-protected power board, and I've run several other computers and pieces of hardware through it with no issues, so that is not the cause. I have just set the hard disk to never turn off, although I don't believe that that will solve the issue. Strangely, this has only happened when I'm not at the computer (which is actually a minority of the time). Until today it had only happened when I had not been actively using the computer for 6 hours, but today it happened within 10-30 minutes of me last using the computer actively. I have enabled file logging from MSI Afterburner, so hopefully this will shed some light on the issue, but I'm not too optimistic. I've heard that it could be a motherboard problem, but I figured I should ask around before RMAing it. Any help?

  Read the article

• ActiveX component can't create Object Error? Check 64 bit Status

  - by Rick Strahl

  If you're running on IIS 7 and a 64 bit operating system you might run into the following error using ASP classic or ASP.NET with COM interop. In classic ASP applications the error will show up as: ActiveX component can't create object   (Error 429) (actually without error handling the error just shows up as 500 error page) In my case the code that's been giving me problems has been a FoxPro COM object I'd been using to serve banner ads to some of my pages. The code basically looks up banners from a database table and displays them at random. The ASP classic code that uses it looks like this: <% Set banner = Server.CreateObject("wwBanner.aspBanner") banner.BannerFile = "wwsitebanners" Response.Write(banner.GetBanner(-1)) %> Originally this code had no specific error checking as above so the ASP pages just failed with 500 error pages from the Web server. To find out what the problem is this code is more useful at least for debugging: <% ON ERROR RESUME NEXT Set banner = Server.CreateObject("wwBanner.aspBanner") Response.Write(err.Number & " - " & err.Description) banner.BannerFile = "wwsitebanners" Response.Write(banner.GetBanner(-1)) %> which results in: 429 - ActiveX component can't create object which at least gives you a slight clue. In ASP.NET invoking the same COM object with code like this: <% dynamic banner = wwUtils.CreateComInstance("wwBanner.aspBanner") as dynamic; banner.cBANNERFILE = "wwsitebanners"; Response.Write(banner.getBanner(-1)); %> results in: Retrieving the COM class factory for component with CLSID {B5DCBB81-D5F5-11D2-B85E-00600889F23B} failed due to the following error: 80040154 Class not registered (Exception from HRESULT: 0x80040154 (REGDB_E_CLASSNOTREG)). The class is in fact registered though and the COM server loads fine from a command prompt or other COM client. This error can be caused by a COM server that doesn't load. It looks like a COM registration error. There are a number of traditional reasons why this error can crop up of course. The server isn't registered (run regserver32 to register a DLL server or /regserver on an EXE server) Access permissions aren't set on the COM server (Web account has to be able to read the DLL ie. Network service) The COM server fails to load during initialization ie. failing during startup One thing I always do to check for COM errors fire up the server in a COM client outside of IIS and ensure that it works there first - it's almost always easier to debug a server outside of the Web environment. In my case I tried the server in Visual FoxPro on the server with: loBanners = CREATEOBJECT("wwBanner.aspBanner") loBanners.cBannerFile = "wwsitebanners" ? loBanners.GetBanner(-1) and it worked just fine. If you don't have a full dev environment on the server you can also use VBScript do the same thing and run the .vbs file from the command prompt: Set banner = Server.CreateObject("wwBanner.aspBanner") banner.BannerFile = "wwsitebanners" MsgBox(banner.getBanner(-1)) Since this both works it tells me the server is registered and working properly. This leaves startup failures or permissions as the problem. I double checked permissions for the Application Pool and the permissions of the folder where the DLL lives and both are properly set to allow access by the Application Pool impersonated user. Just to be sure I assigned an Admin user to the Application Pool but still no go. So now what? 64 bit Servers Ahoy A couple of weeks back I had set up a few of my Application pools to 64 bit mode. My server is Server 2008 64 bit and by default Application Pools run 64 bit. Originally when I installed the server I set up most of my Application Pools to 32 bit mainly for backwards compatibility. But as more of my code migrates to 64 bit OS's I figured it'd be a good idea to see how well code runs under 64 bit code. The transition has been mostly painless. Until today when I noticed the problem with the code above when scrolling to my IIS logs and noticing a lot of 500 errors on many of my ASP classic pages. The code in question in most of these pages deals with this single simple COM object. It took a while to figure out that the problem is caused by the Application Pool running in 64 bit mode. The issue is that 32 bit COM objects (ie. my old Visual FoxPro COM component) cannot be loaded in a 64 bit Application Pool. The ASP pages using this COM component broke on the day I switched my main Application Pool into 64 bit mode but I didn't find the problem until I searched my logs for errors by pure chance. To fix this is easy enough once you know what the problem is by switching the Application Pool to Enable 32-bit Applications: Once this is done the COM objects started working correctly again. 64 bit ASP and ASP.NET with DCOM Servers This is kind of off topic, but incidentally it's possible to load 32 bit DCOM (out of process) servers from ASP.NET and ASP classic even if those applications run in 64 bit application pools. In fact, in West Wind Web Connection I use this capability to run a 64 bit ASP.NET handler that talks to a 32 bit FoxPro COM server which allows West Wind Web Connection to run in native 64 bit mode without custom configuration (which is actually quite useful). It's probably not a common usage scenario but it's good to know that you can actually access 32 bit COM objects this way from ASP.NET. For West Wind Web Connection this works out well as the DCOM interface only makes one non-chatty call to the backend server that handles all the rest of the request processing. Application Pool Isolation is your Friend For me the recent incident of failure in the classic ASP pages has just been another reminder to be very careful with moving applications to 64 bit operation. There are many little traps when switching to 64 bit that are very difficult to track and test for. I described one issue I had a couple of months ago where one of the default ASP.NET filters was loading the wrong version (32bit instead of 64bit) which was extremely difficult to track down and was caused by a very sneaky configuration switch error (basically 3 different entries for the same ISAPI filter all with different bitness settings). It took me almost a full day to track this down). Recently I've been taken to isolate individual applications into separate Application Pools rather than my past practice of combining many apps into shared AppPools. This is a good practice assuming you have enough memory to make this work. Application Pool isolate provides more modularity and allows me to selectively move applications to 64 bit. The error above came about precisely because I moved one of my most populous app pools to 64 bit and forgot about the minimal COM object use in some of my old pages. It's easy to forget. To 64bit or Not Is it worth it to move to 64 bit? Currently I'd say -not really. In my - admittedly limited - testing I don't see any significant performance increases. In fact 64 bit apps just seem to consume considerably more memory (30-50% more in my pools on average) and performance is minimally improved (less than 5% at the very best) in the load testing I've performed on a couple of sites in both modes. The only real incentive for 64 bit would be applications that require huge data spaces that exceed the 32 bit 4 gigabyte memory limit. However I have a hard time imagining an application that needs 4 gigs of memory in a single Application Pool :-). Curious to hear other opinions on benefits of 64 bit operation. © Rick Strahl, West Wind Technologies, 2005-2011Posted in COM   ASP.NET  FoxPro  

  Read the article

• Tips on installing Visual Studio 2010 SP1

  - by Jon Galloway

  Visual Studio SP1 went up on MSDN downloads (here) on March 8, and will be released publicly on March 10 here. Release announcements: Soma: Visual Studio 2010 enhancements Jason Zander: Announcing Visual Studio 2010 Service Pack 1 I started on this post with tips on installing VS2010 SP1 when I realized I’ve been writing these up for Visual Studio and .NET framework SP releases for a while (e.g. VS2008 / .NET 3.5 SP1 post, VS2005 SP1 post). Looking back the years of Visual Studio SP installs (and remembering when we’d get up to SP6 for a Visual Studio release), I’m happy to see that it just keeps getting easier. Service Packs are a lot less finicky about requiring beta software to be uninstalled, install more quickly, and are just generally a lot less scary. If I can’t have a jetpack, at least my future provided me faster, easier service packs. Disclaimer: These tips are just general things I've picked up over the years. I don't have any inside knowledge here. If you see anything wrong, be sure to let me know in the comments. You may want to check the readme file before installing - it's short, and it's in that new-fangled HTML format. On with the tips! Before starting, uninstall Visual Studio features you don't use Visual Studio service packs (and other Microsoft service packs as well) install patches for the specific features you’ve got installed. This is a big reason to always do a custom install when you first install Visual Studio, but it’s not difficult to update your existing installation. Here’s the quick way to do that: Tap the windows key and type “add or remove programs” and press enter (or click on the “Add or remove programs” link if you must).   Type “Visual Studio 2010” in the search box in the upper right corner, click on the Visual Studio program (the one with the VS infinity looking logo) and click on Uninstall/Change. Click on Add or Remove Features The next part’s up to you – what features do you actually use? I’ve been doing primarily ASP.NET MVC development in C# lately, so I selected Visual C# and Visual Web Developer. Remember that you can install features later if needed, and can also install the express versions if you want. Selecting everything just because it’s there - or you paid for it – means that you install updates for everything, every time. When you’ve made your changes, click on the Update button to uninstall unused features. Shut down all instances of Visual Studio It probably goes without saying that you should close a program down before installing it, partly to avoid the file-in-use-reboot-after-install horror. Additional "hunch / works on my machine" quality tip: On one computer I saw a note in the setup log about Visual Studio a prompt for user input to close Visual Studio, although I never saw the prompt. Just to  be sure, I'd personally open up Task Manager and kill any devenv.exe processes I saw running, as it couldn't hurt. Use the web installer I use the Web Installers whenever possible. There’s no point in downloading the DVD unless you’re doing multiple installs or won’t have internet access. The DVD IS is 1.5GB, since it needs to be able to service every possible supported installation option on both x86 and x64. The web installer is 776 KB (smaller than calc.exe), so you can start the installation right away. Like other web installers, the real benefit is that it only installs the updates you need (hence the reason for step 1 – uninstalling unused components). Instead of 1.5GB, my download was roughly 530MB. If you’re installing from MSDN (this link takes you right to the Visual Studio installs), select the first one on the list: The first step in the installation process is to analyze the machine configuration and tell you what needs to be installed. Since I've trimmed down my features, that's a pretty short list. The time's not far off where I may not install SQL Server on my dev machines, just using SQL Server Compact - that would shorten the list further. When I hit next, you can see that the download size has shrunk considerably. When I start the install, note that the installation begins while other components are downloading - another benefit of the web install. On my mid-range desktop machine, the install took 25 minutes. What if it takes longer? According to Heath Stewart (Visual Studio installer guru), average SP1 installs take roughly 45 minutes. An installation which takes hours to complete may be a sign of a problem: see his post Visual Studio 2010 Service Pack 1 installing for over 2 hours could be a sign of a problem. Why so long? Yes, even 25 minutes is a while. Heath's got another blog post explaining why the update can take longer than the initial install (see: A patch may take as long or longer to install than the target product) which explains all the additional steps and complexities a patch needs to deal with, as well as some mitigation steps that deployment authors can take to mitigate the impact. Other things to know about Visual Studio 2010 SP1 Installs over Visual Studio 2010 SP1 Beta That's nice. Previous Visual Studio versions did a number of annoying things when you installed SP's over beta's - fail with weird errors, get part way through and tell you needed to cancel and uninstall first, etc. I've installed this on two machines that had random beta stuff installed without tears. That Readme file you didn't read I mentioned the readme file earlier (http://go.microsoft.com/fwlink/?LinkId=210711 ). Some interesting things I picked up in there: 2.1.3. Visual Studio 2010 Service Pack 1 installation may fail when a USB drive or other removeable drive is connected 2.1.4. Visual Studio must be restarted after Visual Studio 2010 SP1 tooling for SQL Server Compact (Compact) 4.0 is installed 2.2.1. If Visual Studio 2010 Service Pack 1 is uninstalled, Visual Studio 2010 must be reinstalled to restore certain components 2.2.2. If Visual Studio 2010 Service Pack 1 is uninstalled, Visual Studio 2010 must be reinstalled before SP1 can be installed again 2.4.3.1. Async CTP If you installed the pre-SP1 version of Async CTP but did not uninstall it before you installed Visual Studio 2010 SP1, then your computer will be in a state in which the version of the C# compiler in the .NET Framework does not match the C# compiler in Visual Studio. To resolve this issue: After you install Visual Studio 2010 SP1, reinstall the SP1 version of the Async CTP from here. Hardware acceleration for Visual Studio is disabled on Windows XP Visual Studio 2010 SP1 disables hardware acceleration when running on Windows XP (only on XP). You can turn it back on in the Visual Studio options, under Environment / General, as shown below. See Jason Zander's post titled Performance Troubleshooting Article and VS2010 SP1 Change.

  Read the article

• AutoMapper MappingFunction from Source Type of NameValueCollection

  - by REA_ANDREW

  I have had a situation arise today where I need to construct a complex type from a source of a NameValueCollection.  A little while back I submitted a patch for the Agatha Project to include REST (JSON and XML) support for the service contract.  I realized today that as useful as it is, it did not actually support true REST conformance, as REST should support GET so that you can use JSONP from JavaScript directly meaning you can query cross domain services.  My original implementation for POX and JSON used the POST method and this immediately rules out JSONP as from reading, JSONP only works with GET Requests. This then raised another issue.  The current operation contract of Agatha and one of its main benefits is that you can supply an array of Request objects in a single request, limiting the about of server requests you need to make.  Now, at the present time I am thinking that this will not be the case for the REST imlementation but will yield the benefits of the fact that : The same Request objects can be used for SOAP and RST (POX, JSON) The construct of the JavaScript functions will be simpler and more readable It will enable the use of JSONP for cross domain REST Services The current contract for the Agatha WcfRequestProcessor is at time of writing the following: [ServiceContract] public interface IWcfRequestProcessor { [OperationContract(Name = "ProcessRequests")] [ServiceKnownType("GetKnownTypes", typeof(KnownTypeProvider))] [TransactionFlow(TransactionFlowOption.Allowed)] Response[] Process(params Request[] requests); [OperationContract(Name = "ProcessOneWayRequests", IsOneWay = true)] [ServiceKnownType("GetKnownTypes", typeof(KnownTypeProvider))] void ProcessOneWayRequests(params OneWayRequest[] requests); }   My current proposed solution, and at the very early stages of my concept is as follows: [ServiceContract] public interface IWcfRestJsonRequestProcessor { [OperationContract(Name="process")] [ServiceKnownType("GetKnownTypes", typeof(KnownTypeProvider))] [TransactionFlow(TransactionFlowOption.Allowed)] [WebGet(UriTemplate = "process/{name}/{*parameters}", BodyStyle = WebMessageBodyStyle.WrappedResponse, ResponseFormat = WebMessageFormat.Json)] Response[] Process(string name, NameValueCollection parameters); [OperationContract(Name="processoneway",IsOneWay = true)] [ServiceKnownType("GetKnownTypes", typeof(KnownTypeProvider))] [WebGet(UriTemplate = "process-one-way/{name}/{*parameters}", BodyStyle = WebMessageBodyStyle.WrappedResponse, ResponseFormat = WebMessageFormat.Json)] void ProcessOneWayRequests(string name, NameValueCollection parameters); }   Now this part I have not yet implemented, it is the preliminart step which I have developed which will allow me to take the name of the Request Type and the NameValueCollection and construct the complex type which is that of the Request which I can then supply to a nested instance of the original IWcfRequestProcessor  and work as it should normally.  To give an example of some of the urls which you I envisage with this method are: http://www.url.com/service.svc/json/process/getweather/?location=london http://www.url.com/service.svc/json/process/getproductsbycategory/?categoryid=1 http://www.url.om/service.svc/json/process/sayhello/?name=andy Another reason why my direction has gone to a single request for the REST implementation is because of restrictions which are imposed by browsers on the length of the url.  From what I have read this is on average 2000 characters.  I think that this is a very acceptable usage limit in the context of using 1 request, but I do not think this is acceptable for accommodating multiple requests chained together.  I would love to be corrected on that one, I really would but unfortunately from what I have read I have come to the conclusion that this is not the case. The mapping function So, as I say this is just the first pass I have made at this, and I am not overly happy with the try catch for detecting types without default constructors.  I know there is a better way but for the minute, it escapes me.  I would also like to know the correct way for adding mapping functions and not using the anonymous way that I have used.  To achieve this I have used recursion which I am sure is what other mapping function use. As you do have to go as deep as the complex type is. public static object RecurseType(NameValueCollection collection, Type type, string prefix) { try { var returnObject = Activator.CreateInstance(type); foreach (var property in type.GetProperties()) { foreach (var key in collection.AllKeys) { if (String.IsNullOrEmpty(prefix) || key.Length > prefix.Length) { var propertyNameToMatch = String.IsNullOrEmpty(prefix) ? key : key.Substring(property.Name.IndexOf(prefix) + prefix.Length + 1); if (property.Name == propertyNameToMatch) { property.SetValue(returnObject, Convert.ChangeType(collection.Get(key), property.PropertyType), null); } else if(property.GetValue(returnObject,null) == null) { property.SetValue(returnObject, RecurseType(collection, property.PropertyType, String.Concat(prefix, property.PropertyType.Name)), null); } } } } return returnObject; } catch (MissingMethodException) { //Quite a blunt way of dealing with Types without default constructor return null; } }   Another thing is performance, I have not measured this in anyway, it is as I say the first pass, so I hope this can be the start of a more perfected implementation.  I tested this out with a complex type of three levels, there is no intended logical meaning to the properties, they are simply for the purposes of example.  You could call this a spiking session, as from here on in, now I know what I am building I would take a more TDD approach.  OK, purists, why did I not do this from the start, well I didn’t, this was a brain dump and now I know what I am building I can. The console test and how I used with AutoMapper is as follows: static void Main(string[] args) { var collection = new NameValueCollection(); collection.Add("Name", "Andrew Rea"); collection.Add("Number", "1"); collection.Add("AddressLine1", "123 Street"); collection.Add("AddressNumber", "2"); collection.Add("AddressPostCodeCountry", "United Kingdom"); collection.Add("AddressPostCodeNumber", "3"); AutoMapper.Mapper.CreateMap<NameValueCollection, Person>() .ConvertUsing(x => { return(Person) RecurseType(x, typeof(Person), null); }); var person = AutoMapper.Mapper.Map<NameValueCollection, Person>(collection); Console.WriteLine(person.Name); Console.WriteLine(person.Number); Console.WriteLine(person.Address.Line1); Console.WriteLine(person.Address.Number); Console.WriteLine(person.Address.PostCode.Country); Console.WriteLine(person.Address.PostCode.Number); Console.ReadLine(); }   Notice the convention that I am using and that this method requires you do use.  Each property is prefixed with the constructed name of its parents combined.  This is the convention used by AutoMapper and it makes sense. I can also think of other uses for this including using with ASP.NET MVC ModelBinders for creating a complex type from the QueryString which is itself is a NameValueCollection. Hope this is of some help to people and I would welcome any code reviews you could give me. References: Agatha : http://code.google.com/p/agatha-rrsl/ AutoMapper : http://automapper.codeplex.com/   Cheers for now, Andrew   P.S. I will have the proposed solution for a more complete REST implementation for AGATHA very soon. 

  Read the article

• Adding DTrace Probes to PHP Extensions

  - by cj

  The powerful DTrace tracing facility has some PHP-specific probes that can be enabled with --enable-dtrace. DTrace for Linux is being created by Oracle and is currently in tech preview. Currently it doesn't support userspace tracing so, in the meantime, Systemtap can be used to monitor the probes implemented in PHP. This was recently outlined in David Soria Parra's post Probing PHP with Systemtap on Linux. My post shows how DTrace probes can be added to PHP extensions and traced on Linux. I was using Oracle Linux 6.3. Not all Linux kernels are built with Systemtap, since this can impact stability. Check whether your running kernel (or others installed) have Systemtap enabled, and reboot with such a kernel: # grep CONFIG_UTRACE /boot/config-`uname -r` # grep CONFIG_UTRACE /boot/config-* When you install Systemtap itself, the package systemtap-sdt-devel is needed since it provides the sdt.h header file: # yum install systemtap-sdt-devel You can now install and build PHP as shown in David's article. Basically the build is with: $ cd ~/php-src $ ./configure --disable-all --enable-dtrace $ make (For me, running 'make' a second time failed with an error. The workaround is to do 'git checkout Zend/zend_dtrace.d' and then rerun 'make'. See PHP Bug 63704) David's article shows how to trace the probes already implemented in PHP. You can also use Systemtap to trace things like userspace PHP function calls. For example, create test.php: <?php $c = oci_connect('hr', 'welcome', 'localhost/orcl'); $s = oci_parse($c, "select dbms_xmlgen.getxml('select * from dual') xml from dual"); $r = oci_execute($s); $row = oci_fetch_array($s, OCI_NUM); $x = $row[0]->load(); $row[0]->free(); echo $x; ?> The normal output of this file is the XML form of Oracle's DUAL table: $ ./sapi/cli/php ~/test.php <?xml version="1.0"?> <ROWSET> <ROW> <DUMMY>X</DUMMY> </ROW> </ROWSET> To trace the PHP function calls, create the tracing file functrace.stp: probe process("sapi/cli/php").function("zif_*") { printf("Started function %s\n", probefunc()); } probe process("sapi/cli/php").function("zif_*").return { printf("Ended function %s\n", probefunc()); } This makes use of the way PHP userspace functions (not builtins) like oci_connect() map to C functions with a "zif_" prefix. Login as root, and run System tap on the PHP script: # cd ~cjones/php-src # stap -c 'sapi/cli/php ~cjones/test.php' ~cjones/functrace.stp Started function zif_oci_connect Ended function zif_oci_connect Started function zif_oci_parse Ended function zif_oci_parse Started function zif_oci_execute Ended function zif_oci_execute Started function zif_oci_fetch_array Ended function zif_oci_fetch_array Started function zif_oci_lob_load <?xml version="1.0"?> <ROWSET> <ROW> <DUMMY>X</DUMMY> </ROW> </ROWSET> Ended function zif_oci_lob_load Started function zif_oci_free_descriptor Ended function zif_oci_free_descriptor Each call and return is logged. The Systemtap scripting language allows complex scripts to be built. There are many examples on the web. To augment this generic capability and the PHP probes in PHP, other extensions can have probes too. Below are the steps I used to add probes to OCI8: I created a provider file ext/oci8/oci8_dtrace.d, enabling three probes. The first one will accept a parameter that runtime tracing can later display: provider php { probe oci8__connect(char *username); probe oci8__nls_start(); probe oci8__nls_done(); }; I updated ext/oci8/config.m4 with the PHP_INIT_DTRACE macro. The patch is at the end of config.m4. The macro takes the provider prototype file, a name of the header file that 'dtrace' will generate, and a list of sources files with probes. When --enable-dtrace is used during PHP configuration, then the outer $PHP_DTRACE check is true and my new probes will be enabled. I've chosen to define an OCI8 specific macro, HAVE_OCI8_DTRACE, which can be used in the OCI8 source code: diff --git a/ext/oci8/config.m4 b/ext/oci8/config.m4 index 34ae76c..f3e583d 100644 --- a/ext/oci8/config.m4 +++ b/ext/oci8/config.m4 @@ -341,4 +341,17 @@ if test "$PHP_OCI8" != "no"; then PHP_SUBST_OLD(OCI8_ORACLE_VERSION) fi + + if test "$PHP_DTRACE" = "yes"; then + AC_CHECK_HEADERS([sys/sdt.h], [ + PHP_INIT_DTRACE([ext/oci8/oci8_dtrace.d], + [ext/oci8/oci8_dtrace_gen.h],[ext/oci8/oci8.c]) + AC_DEFINE(HAVE_OCI8_DTRACE,1, + [Whether to enable DTrace support for OCI8 ]) + ], [ + AC_MSG_ERROR( + [Cannot find sys/sdt.h which is required for DTrace support]) + ]) + fi + fi In ext/oci8/oci8.c, I added the probes at, for this example, semi-arbitrary places: diff --git a/ext/oci8/oci8.c b/ext/oci8/oci8.c index e2241cf..ffa0168 100644 --- a/ext/oci8/oci8.c +++ b/ext/oci8/oci8.c @@ -1811,6 +1811,12 @@ php_oci_connection *php_oci_do_connect_ex(char *username, int username_len, char } } +#ifdef HAVE_OCI8_DTRACE + if (DTRACE_OCI8_CONNECT_ENABLED()) { + DTRACE_OCI8_CONNECT(username); + } +#endif + /* Initialize global handles if they weren't initialized before */ if (OCI_G(env) == NULL) { php_oci_init_global_handles(TSRMLS_C); @@ -1870,11 +1876,22 @@ php_oci_connection *php_oci_do_connect_ex(char *username, int username_len, char size_t rsize = 0; sword result; +#ifdef HAVE_OCI8_DTRACE + if (DTRACE_OCI8_NLS_START_ENABLED()) { + DTRACE_OCI8_NLS_START(); + } +#endif PHP_OCI_CALL_RETURN(result, OCINlsEnvironmentVariableGet, (&charsetid_nls_lang, 0, OCI_NLS_CHARSET_ID, 0, &rsize)); if (result != OCI_SUCCESS) { charsetid_nls_lang = 0; } smart_str_append_unsigned_ex(&hashed_details, charsetid_nls_lang, 0); + +#ifdef HAVE_OCI8_DTRACE + if (DTRACE_OCI8_NLS_DONE_ENABLED()) { + DTRACE_OCI8_NLS_DONE(); + } +#endif } timestamp = time(NULL); The oci_connect(), oci_pconnect() and oci_new_connect() calls all use php_oci_do_connect_ex() internally. The first probe simply records that the PHP application made a connection call. I already showed a way to do this without needing a probe, but adding a specific probe lets me record the username. The other two probes can be used to time how long the globalization initialization takes. The relationships between the oci8_dtrace.d names like oci8__connect, the probe guards like DTRACE_OCI8_CONNECT_ENABLED() and probe names like DTRACE_OCI8_CONNECT() are obvious after seeing the pattern of all three probes. I included the new header that will be automatically created by the dtrace tool when PHP is built. I did this in ext/oci8/php_oci8_int.h: diff --git a/ext/oci8/php_oci8_int.h b/ext/oci8/php_oci8_int.h index b0d6516..c81fc5a 100644 --- a/ext/oci8/php_oci8_int.h +++ b/ext/oci8/php_oci8_int.h @@ -44,6 +44,10 @@ # endif # endif /* osf alpha */ +#ifdef HAVE_OCI8_DTRACE +#include "oci8_dtrace_gen.h" +#endif + #if defined(min) #undef min #endif Now PHP can be rebuilt: $ cd ~/php-src $ rm configure && ./buildconf --force $ ./configure --disable-all --enable-dtrace \ --with-oci8=instantclient,/home/cjones/instantclient $ make If 'make' fails, do the 'git checkout Zend/zend_dtrace.d' trick I mentioned. The new probes can be seen by logging in as root and running: # stap -l 'process.provider("php").mark("oci8*")' -c 'sapi/cli/php -i' process("sapi/cli/php").provider("php").mark("oci8__connect") process("sapi/cli/php").provider("php").mark("oci8__nls_done") process("sapi/cli/php").provider("php").mark("oci8__nls_start") To test them out, create a new trace file, oci.stp: global numconnects; global start; global numcharlookups = 0; global tottime = 0; probe process.provider("php").mark("oci8-connect") { printf("Connected as %s\n", user_string($arg1)); numconnects += 1; } probe process.provider("php").mark("oci8-nls_start") { start = gettimeofday_us(); numcharlookups++; } probe process.provider("php").mark("oci8-nls_done") { tottime += gettimeofday_us() - start; } probe end { printf("Connects: %d, Charset lookups: %ld\n", numconnects, numcharlookups); printf("Total NLS charset initalization time: %ld usecs/connect\n", (numcharlookups 0 ? tottime/numcharlookups : 0)); } This calculates the average time that the NLS character set lookup takes. It also prints out the username of each connection, as an example of using parameters. Login as root and run Systemtap over the PHP script: # cd ~cjones/php-src # stap -c 'sapi/cli/php ~cjones/test.php' ~cjones/oci.stp Connected as cj <?xml version="1.0"?> <ROWSET> <ROW> <DUMMY>X</DUMMY> </ROW> </ROWSET> Connects: 1, Charset lookups: 1 Total NLS charset initalization time: 164 usecs/connect This shows the time penalty of making OCI8 look up the default character set. This time would be zero if a character set had been passed as the fourth argument to oci_connect() in test.php.

  Read the article

• Tools to Help Post Content On Your WordPress Blog

  - by Matthew Guay

  Now that you’ve got a nice blog, you want to do more with it and start posting content.  Here we look at some tools that will allow you to post directly to your WordPress blog. Writing a new blog post is easy with WordPress as we saw in our previous post about Starting your own WordPress blog.  The web editor gives you a lot of features and even lets you edit your post’s source code if you enjoy hacking HTML.  There are other tools that will allow you to post content, here we look at how you can post with dedicated apps, browser plugins, and even by email. Windows Live Writer Windows Live Writer (part of the Windows Live Essentials Suite) is a great app for posting content to your blog.  This free program for Microsoft lets you post content to a variety of blogging services, including Blogger, Typepad, LiveJournal, and of course WordPress.  You can write blog posts directly from its Word-like editor, complete with pictures and advanced formatting.  Even if you’re offline, you can still write posts and save them for when you’re online again. For more information about installing Live writer, check out our article on how to Install Windows Live Essentials In Windows 7. Once Live Writer is installed, open it to add your blog.  If you already had Live Writer installed and configured for a blog, you can add your new blog, too.  Just click your blog’s name in the top right corner, and select “Add blog account”. Select “Other blog service” to add your WordPress blog to Writer, and click Next.   Enter your blog’s web address, and your username and password.  Check Remember my password so you don’t have to enter it every time you write something. Writer will analyze your blog and setup your account. During the setup process it may ask to post a temporary post.  This will let you preview blog posts using your blog’s real theme, which is helpful, so click Yes. Finally, add your Blog’s name, and click Finish. You can now use the rich editor to write and add content to a new blog post.   Select the Preview tab to see how your post will look on your blog… Or, if you’re a HTML geek, select the Source tab to edit the code of your blog post. From the bottom of the window, you can choose categories, insert tags, and even schedule the post to publish on a different day.  Live Writer is fully integrated with WordPress; you’re not missing anything by using the desktop editor. If you want to edit a post you’ve already published, click the Open button and select the post.  You can chose and edit any post, including ones you published via the web interface or other editors. Add Multimedia Content to your Posts with Live Writer Back in the Edit tab, you can add pictures, videos and more from the sidebar.  Select what you want to insert. Pictures If you insert a picture, you can add many nice borders and designs to it. Or, you can even add artistic effects from the Effects tab in the sidebar. Photo Gallery If you want to post several pictures, say some of your vacation shots, then inserting a picture gallery may be the best option.  Select Insert Photo Gallery in the sidebar, and then choose the pictures you want in the gallery. Once the gallery is inserted, you can choose from several styles to showcase your pictures. When you post the blog, you will be asked to sign in with your Windows Live ID as the gallery pictures will be stored in the free Skydrive storage service. Your blog readers can see the preview of your pictures directly on your blog, and then can view each individual picture, download them, or see a slideshow online via the link. Video If you want to add a video to your blog post, select Video from the sidebar as above.  You can select a video that’s already online, or you can choose a new video from file and upload it via YouTube directly from Windows Live Writer.   Note that you will have to sign in with your YouTube account to upload videos to YouTube, so if you’re not logged in you’ll be prompted to do so when you click Insert. Geek Tip:  If you ever want to copy your Live Writer settings to another computer, check out our article on how to Backup Your Windows Live Writer Settings. Microsoft Office Word Word 2007 and 2010 also let you post content directly to your blog.  This is especially nice if you’ve already typed up a document and think it would be good on your Blog as well.  Check out our in-depth tutorial on posting blog posts via Word 2007 using Word 2007 as a blogging tool. This works in Word 2010 too, except the Office Orb has been replaced by the new Backstage view.  So, in Word 2010, to start a new blog post, click File \ New then select Blog post.  Proceed as you would in Word 2007 to add your blog settings and post the content you want. Or, if you’ve already written a document and want to post it, select File \ Share (or Save and Send in the final version of Word 2010), and then click Publish as Blog Post.  If you haven’t setup your blog account yet, set it up as shown in the Word 2007 article. Post Via Email Most of us use email daily, and already have our favorite email app or service.  Whether on your desktop or mobile phone, it’s easy to create rich emails and add content.  WordPress lets you generate a unique email address that you can use to easily post content and email to your blog.  Just compose your email with the subject as the title of your post, and send it to this unique address.  Your new post will be up in minutes. To active this feature, click the My Account button in the top menu bar in your WordPress.com account, and select My Blogs. Click the Enable button under Post by Email beside your blog’s name.   Now you’ll have a private email you can use to post to your blog.  Anything you send to this email will be posted as a new post.  If you think your email may be compromised, click Regenerate to get a new publishing email address. Any email program or webapp now is a blog post editor.  Feel free to use rich formatting or insert pictures; it all comes through great.  This is also a great way to post to your blog from your mobile device.  Whether you’re using webmail or a dedicated email client on your phone, you can now blog from anywhere.   Mobile Applications WordPress also offer dedicated applications for blogging directly from your mobile device.  You can write new posts, edit existing ones, and manage comments all from your Smartphone.  Currently they offer apps for iPhone, Android, and Blackberry.  Check them out at the link below. Conclusion Whether you want to write from your browser or email a post to your blog, WordPress is flexible enough to work right along with your preferences.  However you post, you can be sure that it will look professional and be easily accessible with your WordPress blog. Download Windows Live Writer Download WordPress apps for your mobile device Similar Articles Productive Geek Tips Quick Tip: Set a Future Date for a Post in WordPressAdd Social Bookmarking (Digg This!) Links to your Wordpress BlogFuture Date a Post in Windows Live WriterHow To Start Your Own Professional Blog with WordPressUsing Word 2007 as a Blogging Tool TouchFreeze Alternative in AutoHotkey The Icy Undertow Desktop Windows Home Server – Backup to LAN The Clear & Clean Desktop Use This Bookmarklet to Easily Get Albums Use AutoHotkey to Assign a Hotkey to a Specific Window Latest Software Reviews Tinyhacker Random Tips Acronis Online Backup DVDFab 6 Revo Uninstaller Pro Registry Mechanic 9 for Windows Fun with 47 charts and graphs Tomorrow is Mother’s Day Check the Average Speed of YouTube Videos You’ve Watched OutlookStatView Scans and Displays General Usage Statistics How to Add Exceptions to the Windows Firewall Office 2010 reviewed in depth by Ed Bott

  Read the article

• Oracle Data Mining a Star Schema: Telco Churn Case Study

  - by charlie.berger

  There is a complete and detailed Telco Churn case study "How to" Blog Series just posted by Ari Mozes, ODM Dev. Manager.  In it, Ari provides detailed guidance in how to leverage various strengths of Oracle Data Mining including the ability to: mine Star Schemas and join tables and views together to obtain a complete 360 degree view of a customer combine transactional data e.g. call record detail (CDR) data, etc. define complex data transformation, model build and model deploy analytical methodologies inside the Database  His blog is posted in a multi-part series.  Below are some opening excerpts for the first 3 blog entries.  This is an excellent resource for any novice to skilled data miner who wants to gain competitive advantage by mining their data inside the Oracle Database.  Many thanks Ari! Mining a Star Schema: Telco Churn Case Study (1 of 3) One of the strengths of Oracle Data Mining is the ability to mine star schemas with minimal effort.  Star schemas are commonly used in relational databases, and they often contain rich data with interesting patterns.  While dimension tables may contain interesting demographics, fact tables will often contain user behavior, such as phone usage or purchase patterns.  Both of these aspects - demographics and usage patterns - can provide insight into behavior.Churn is a critical problem in the telecommunications industry, and companies go to great lengths to reduce the churn of their customer base.  One case study1 describes a telecommunications scenario involving understanding, and identification of, churn, where the underlying data is present in a star schema.  That case study is a good example for demonstrating just how natural it is for Oracle Data Mining to analyze a star schema, so it will be used as the basis for this series of posts...... Mining a Star Schema: Telco Churn Case Study (2 of 3) This post will follow the transformation steps as described in the case study, but will use Oracle SQL as the means for preparing data.  Please see the previous post for background material, including links to the case study and to scripts that can be used to replicate the stages in these posts.1) Handling missing values for call data recordsThe CDR_T table records the number of phone minutes used by a customer per month and per call type (tariff).  For example, the table may contain one record corresponding to the number of peak (call type) minutes in January for a specific customer, and another record associated with international calls in March for the same customer.  This table is likely to be fairly dense (most type-month combinations for a given customer will be present) due to the coarse level of aggregation, but there may be some missing values.  Missing entries may occur for a number of reasons: the customer made no calls of a particular type in a particular month, the customer switched providers during the timeframe, or perhaps there is a data entry problem.  In the first situation, the correct interpretation of a missing entry would be to assume that the number of minutes for the type-month combination is zero.  In the other situations, it is not appropriate to assume zero, but rather derive some representative value to replace the missing entries.  The referenced case study takes the latter approach.  The data is segmented by customer and call type, and within a given customer-call type combination, an average number of minutes is computed and used as a replacement value.In SQL, we need to generate additional rows for the missing entries and populate those rows with appropriate values.  To generate the missing rows, Oracle's partition outer join feature is a perfect fit.  select cust_id, cdre.tariff, cdre.month, minsfrom cdr_t cdr partition by (cust_id) right outer join     (select distinct tariff, month from cdr_t) cdre     on (cdr.month = cdre.month and cdr.tariff = cdre.tariff);   ....... Mining a Star Schema: Telco Churn Case Study (3 of 3) Now that the "difficult" work is complete - preparing the data - we can move to building a predictive model to help identify and understand churn.The case study suggests that separate models be built for different customer segments (high, medium, low, and very low value customer groups).  To reduce the data to a single segment, a filter can be applied: create or replace view churn_data_high asselect * from churn_prep where value_band = 'HIGH'; It is simple to take a quick look at the predictive aspects of the data on a univariate basis.  While this does not capture the more complex multi-variate effects as would occur with the full-blown data mining algorithms, it can give a quick feel as to the predictive aspects of the data as well as validate the data preparation steps.  Oracle Data Mining includes a predictive analytics package which enables quick analysis. begin  dbms_predictive_analytics.explain(   'churn_data_high','churn_m6','expl_churn_tab'); end; /select * from expl_churn_tab where rank <= 5 order by rank; ATTRIBUTE_NAME       ATTRIBUTE_SUBNAME EXPLANATORY_VALUE RANK-------------------- ----------------- ----------------- ----------LOS_BAND                                      .069167052          1MINS_PER_TARIFF_MON  PEAK-5                   .034881648          2REV_PER_MON          REV-5                    .034527798          3DROPPED_CALLS                                 .028110322          4MINS_PER_TARIFF_MON  PEAK-4                   .024698149          5From the above results, it is clear that some predictors do contain information to help identify churn (explanatory value > 0).  The strongest uni-variate predictor of churn appears to be the customer's (binned) length of service.  The second strongest churn indicator appears to be the number of peak minutes used in the most recent month.  The subname column contains the interior piece of the DM_NESTED_NUMERICALS column described in the previous post.  By using the object relational approach, many related predictors are included within a single top-level column. .....   NOTE:  These are just EXCERPTS.  Click here to start reading the Oracle Data Mining a Star Schema: Telco Churn Case Study from the beginning.    

  Read the article

• How to get SQL Railroad Diagrams from MSDN BNF syntax notation.

  - by Phil Factor

  pre {margin-bottom:.0001pt; font-size:8.0pt; font-family:"Courier New"; margin-left: 0cm; margin-right: 0cm; margin-top: 0cm; } On SQL Server Books-On-Line, in the Transact-SQL Reference (database Engine), every SQL Statement has its syntax represented in  ‘Backus–Naur Form’ notation (BNF)  syntax. For a programmer in a hurry, this should be ideal because It is the only quick way to understand and appreciate all the permutations of the syntax. It is a great feature once you get your eye in. It isn’t the only way to get the information;  You can, of course, reverse-engineer an understanding of the syntax from the examples, but your understanding won’t be complete, and you’ll have wasted time doing it. BNF is a good start in representing the syntax:  Oracle and SQLite go one step further, and have proper railroad diagrams for their syntax, which is a far more accessible way of doing it. There are three problems with the BNF on MSDN. Firstly, it is isn’t a standard version of  BNF, but an ancient fork from EBNF, inherited from Sybase. Secondly, it is excruciatingly difficult to understand, and thirdly it has a number of syntactic and semantic errors. The page describing DML triggers, for example, currently has the absurd BNF error that makes it state that all statements in the body of the trigger must be separated by commas.  There are a few other detail problems too. Here is the offending syntax for a DML trigger, pasted from MSDN. Trigger on an INSERT, UPDATE, or DELETE statement to a table or view (DML Trigger) CREATE TRIGGER [ schema_name . ]trigger_name ON { table | view } [ WITH <dml_trigger_option> [ ,...n ] ] { FOR | AFTER | INSTEAD OF } { [ INSERT ] [ , ] [ UPDATE ] [ , ] [ DELETE ] } [ NOT FOR REPLICATION ] AS { sql_statement [ ; ] [ ,...n ] | EXTERNAL NAME <method specifier [ ; ] > }   <dml_trigger_option> ::=     [ ENCRYPTION ]     [ EXECUTE AS Clause ]   <method_specifier> ::=  This should, of course, be /* Trigger on an INSERT, UPDATE, or DELETE statement to a table or view (DML Trigger) */ CREATE TRIGGER [ schema_name . ]trigger_name ON { table | view } [ WITH <dml_trigger_option> [ ,...n ] ] { FOR | AFTER | INSTEAD OF } { [ INSERT ] [ , ] [ UPDATE ] [ , ] [ DELETE ] } [ NOT FOR REPLICATION ] AS { {sql_statement [ ; ]} [ ...n ] | EXTERNAL NAME <method_specifier> [ ; ] }   <dml_trigger_option> ::=     [ ENCRYPTION ]     [ EXECUTE AS CLAUSE ]   <method_specifier> ::=     assembly_name.class_name.method_name I’d love to tell Microsoft when I spot errors like this so they can correct them but I can’t. Obviously, there is a mechanism on MSDN to get errors corrected by using comments, but that doesn’t work for me (*Error occurred while saving your data.”), and when I report that the comment system doesn’t work to MSDN, I get no reply. I’ve been trying to create railroad diagrams for all the important SQL Server SQL statements, as good as you’d find for Oracle, and have so far published the CREATE TABLE and ALTER TABLE railroad diagrams based on the BNF. Although I’ve been aware of them, I’ve never realised until recently how many errors there are. Then, Colin Daley created a translator for the SQL Server dialect of  BNF which outputs standard EBNF notation used by the W3C. The example MSDN BNF for the trigger would be rendered as … /* Trigger on an INSERT, UPDATE, or DELETE statement to a table or view (DML Trigger) */ create_trigger ::= 'CREATE TRIGGER' ( schema_name '.' ) ? trigger_name 'ON' ( table | view ) ( 'WITH' dml_trigger_option ( ',' dml_trigger_option ) * ) ? ( 'FOR' | 'AFTER' | 'INSTEAD OF' ) ( ( 'INSERT' ) ? ( ',' ) ? ( 'UPDATE' ) ? ( ',' ) ? ( 'DELETE' ) ? ) ( 'NOT FOR REPLICATION' ) ? 'AS' ( ( sql_statement ( ';' ) ? ) + | 'EXTERNAL NAME' method_specifier ( ';' ) ? )   dml_trigger_option ::= ( 'ENCRYPTION' ) ? ( 'EXECUTE AS CLAUSE' ) ?   method_specifier ::= assembly_name '.' class_name '.' method_name Colin’s intention was to allow anyone to paste SQL Server’s BNF notation into his website-based parser, and from this generate classic railroad diagrams via Gunther Rademacher's Railroad Diagram Generator.  Colin's application does this for you: you're not aware that you are moving to a different site.  Because Colin's 'translator' it is a parser, it will pick up syntax errors. Once you’ve fixed the syntax errors, you will get the syntax in the form of a human-readable railroad diagram and, in this form, the semantic mistakes become flamingly obvious. Gunter’s Railroad Diagram Generator is brilliant. To be able, after correcting the MSDN dialect of BNF, to generate a standard EBNF, and from thence to create railroad diagrams for SQL Server’s syntax that are as good as Oracle’s, is a great boon, and many thanks to Colin for the idea. Here is the result of the W3C EBNF from Colin’s application then being run through the Railroad diagram generator. create_trigger: dml_trigger_option: method_specifier:   Now that’s much better, you’ll agree. This is pretty easy to understand, and at this point any error is immediately obvious. This should be seriously useful, and it is to me. However  there is that snag. The BNF is generally incorrect, and you can’t expect the average visitor to mess about with it. The answer is, of course, to correct the BNF on MSDN and maybe even add railroad diagrams for the syntax. Stop giggling! I agree it won’t happen. In the meantime, we need to collaboratively store and publish these corrected syntaxes ourselves as we do them. How? GitHub?  SQL Server Central?  Simple-Talk? What should those of us who use the system  do with our corrected EBNF so that anyone can use them without hassle?

  Read the article

• JMX Based Monitoring - Part Four - Business App Server Monitoring

  - by Anthony Shorten

  In the last blog entry I talked about the Oracle Utilities Application Framework V4 feature for monitoring and managing aspects of the Web Application Server using JMX. In this blog entry I am going to discuss a similar new feature that allows JMX to be used for management and monitoring the Oracle Utilities business application server component. This feature is primarily focussed on performance tracking of the product. In first release of Oracle Utilities Customer Care And Billing (V1.x I am talking about), we used to use Oracle Tuxedo as part of the architecture. In Oracle Utilities Application Framework V2.0 and above, we removed Tuxedo from the architecture. One of the features that some customers used within Tuxedo was the performance tracking ability. The idea was that you enabled performance logging on the individual Tuxedo servers and then used a utility named txrpt to produce a performance report. This report would list every service called, the number of times it was called and the average response time. When I worked a performance consultant, I used this report to identify badly performing services and also gauge the overall performance characteristics of a site. When Tuxedo was removed from the architecture this information was also lost. While you can get some information from access.log and some Mbeans supplied by the Web Application Server it was not at the same granularity as txrpt or as useful. I am happy to say we have not only reintroduced this facility in Oracle Utilities Application Framework but it is now accessible via JMX and also we have added more detail into the performance tracking. Most of this new design was working with customers around the world to make sure we introduced a new feature that not only satisfied their performance tracking needs but allowed for finer grained performance analysis. As with the Web Application Server, the Business Application Server JMX monitoring is enabled by specifying a JMX port number in RMI Port number for JMX Business and initial credentials in the JMX Enablement System User ID and JMX Enablement System Password configuration options. These options are available using the configureEnv[.sh] -a utility. These credentials are shared across the Web Application Server and Business Application Server for authorization purposes. Once this is information is supplied a number of configuration files are built (by the initialSetup[.sh] utility) to configure the facility: spl.properties - contains the JMX URL, the security configuration and the mbeans that are enabled. For example, on my demonstration machine: spl.runtime.management.rmi.port=6750 spl.runtime.management.connector.url.default=service:jmx:rmi:///jndi/rmi://localhost:6750/oracle/ouaf/ejbAppConnector jmx.remote.x.password.file=scripts/ouaf.jmx.password.file jmx.remote.x.access.file=scripts/ouaf.jmx.access.file ouaf.jmx.com.splwg.ejb.service.management.PerformanceStatistics=enabled ouaf.jmx.* files - contain the userid and password. The default configuration uses the JMX default configuration. You can use additional security features by altering the spl.properties file manually or using a custom template. For more security options see JMX Security for more details. Once it has been configured and the changes reflected in the product using the initialSetup[.sh] utility the JMX facility can be used. For illustrative purposes I will use jconsole but any JSR160 complaint browser or client can be used (with the appropriate configuration). Once you start jconsole (ensure that splenviron[.sh] is executed prior to execution to set the environment variables or for remote connection, ensure java is in your path and jconsole.jar in your classpath) you specify the URL in the spl.runtime.management.connnector.url.default entry. For example: You are then able to track performance of the product using the PerformanceStatistics Mbean. The attributes of the PerformanceStatistics Mbean are counts of each object type. This is where this facility differs from txrpt. The information that is collected includes the following: The Service Type is captured so you can filter the results in terms of the type of service. For maintenance type services you can even see the transaction type (ADD, CHANGE etc) so you can see the performance of updates against read transactions. The Minimum and Maximum are also collected to give you an idea of the spread of performance. The last call is recorded. The date, time and user of the last call are recorded to give you an idea of the timeliness of the data. The Mbean maintains a set of counters per Service Type to give you a summary of the types of transactions being executed. This gives you an overall picture of the types of transactions and volumes at your site. There are a number of interesting operations that can also be performed: reset - This resets the statistics back to zero. This is an important operation. For example, txrpt is restricted to collecting statistics per hour, which is ok for most people. But what if you wanted to be more granular? This operation allows to set the collection period to anything you wish. The statistics collected will represent values since the last restart or last reset. completeExecutionDump - This is the operation that produces a CSV in memory to allow extraction of the data. All the statistics are extracted (see the Server Administration Guide for a full list). This can be then loaded into a database, a tool or simply into your favourite spreadsheet for analysis. Here is an extract of an execution dump from my demonstration environment to give you an idea of the format: ServiceName, ServiceType, MinTime, MaxTime, Avg Time, # of Calls, Latest Time, Latest Date, Latest User ... CFLZLOUL, EXECUTE_LIST, 15.0, 64.0, 22.2, 10, 16.0, 2009-12-16::11-25-36-932, ASHORTEN CILBBLLP, READ, 106.0, 1184.0, 466.3333333333333, 6, 106.0, 2009-12-16::11-39-01-645, BOBAMA CILBBLLP, DELETE, 70.0, 146.0, 108.0, 2, 70.0, 2009-12-15::12-53-58-280, BPAYS CILBBLLP, ADD, 860.0, 4903.0, 2243.5, 8, 860.0, 2009-12-16::17-54-23-862, LELLISON CILBBLLP, CHANGE, 112.0, 3410.0, 815.1666666666666, 12, 112.0, 2009-12-16::11-40-01-103, ASHORTEN CILBCBAL, EXECUTE_LIST, 8.0, 84.0, 26.0, 22, 23.0, 2009-12-16::17-54-01-643, LJACKMAN InitializeUserInfoService, READ_SYSTEM, 49.0, 962.0, 70.83777777777777, 450, 63.0, 2010-02-25::11-21-21-667, ASHORTEN InitializeUserService, READ_SYSTEM, 130.0, 2835.0, 234.85777777777778, 450, 216.0, 2010-02-25::11-21-21-446, ASHORTEN MenuLoginService, READ_SYSTEM, 530.0, 1186.0, 703.3333333333334, 9, 530.0, 2009-12-16::16-39-31-172, ASHORTEN NavigationOptionDescriptionService, READ_SYSTEM, 2.0, 7.0, 4.0, 8, 2.0, 2009-12-21::09-46-46-892, ASHORTEN ... There are other operations and attributes available. Refer to the Server Administration Guide provided with your product to understand the full et of operations and attributes. This is one of the many features I am proud that we implemented as it allows flexible monitoring of the performance of the product.

  Read the article

• Observations in Migrating from JavaFX Script to JavaFX 2.0

  - by user12608080

  Observations in Migrating from JavaFX Script to JavaFX 2.0 Introduction Having been available for a few years now, there is a decent body of work written for JavaFX using the JavaFX Script language. With the general availability announcement of JavaFX 2.0 Beta, the natural question arises about converting the legacy code over to the new JavaFX 2.0 platform. This article reflects on some of the observations encountered while porting source code over from JavaFX Script to the new JavaFX API paradigm. The Application The program chosen for migration is an implementation of the Sudoku game and serves as a reference application for the book JavaFX – Developing Rich Internet Applications. The design of the program can be divided into two major components: (1) A user interface (ideally suited for JavaFX design) and (2) the puzzle generator. For the context of this article, our primary interest lies in the user interface. The puzzle generator code was lifted from a sourceforge.net project and is written entirely in Java. Regardless which version of the UI we choose (JavaFX Script vs. JavaFX 2.0), no code changes were required for the puzzle generator code. The original user interface for the JavaFX Sudoku application was written exclusively in JavaFX Script, and as such is a suitable candidate to convert over to the new JavaFX 2.0 model. However, a few notable points are worth mentioning about this program. First off, it was written in the JavaFX 1.1 timeframe, where certain capabilities of the JavaFX framework were as of yet unavailable. Citing two examples, this program creates many of its own UI controls from scratch because the built-in controls were yet to be introduced. In addition, layout of graphical nodes is done in a very manual manner, again because much of the automatic layout capabilities were in flux at the time. It is worth considering that this program was written at a time when most of us were just coming up to speed on this technology. One would think that having the opportunity to recreate this application anew, it would look a lot different from the current version. Comparing the Size of the Source Code An attempt was made to convert each of the original UI JavaFX Script source files (suffixed with .fx) over to a Java counterpart. Due to language feature differences, there are a small number of source files which only exist in one version or the other. The table below summarizes the size of each of the source files. JavaFX Script source file Number of Lines Number of Character JavaFX 2.0 Java source file Number of Lines Number of Characters ArrowKey.java 6 72 Board.fx 221 6831 Board.java 205 6508 BoardNode.fx 446 16054 BoardNode.java 723 29356 ChooseNumberNode.fx 168 5267 ChooseNumberNode.java 302 10235 CloseButtonNode.fx 115 3408 CloseButton.java 99 2883 ParentWithKeyTraversal.java 111 3276 FunctionPtr.java 6 80 Globals.java 20 554 Grouping.fx 8 140 HowToPlayNode.fx 121 3632 HowToPlayNode.java 136 4849 IconButtonNode.fx 196 5748 IconButtonNode.java 183 5865 Main.fx 98 3466 Main.java 64 2118 SliderNode.fx 288 10349 SliderNode.java 350 13048 Space.fx 78 1696 Space.java 106 2095 SpaceNode.fx 227 6703 SpaceNode.java 220 6861 TraversalHelper.fx 111 3095 Total 2,077 79,127 2531 87,800 A few notes about this table are in order: The number of lines in each file was determined by running the Unix ‘wc –l’ command over each file. The number of characters in each file was determined by running the Unix ‘ls –l’ command over each file. The examination of the code could certainly be much more rigorous. No standard formatting was performed on these files.  All comments however were deleted. There was a certain expectation that the new Java version would require more lines of code than the original JavaFX script version. As evidenced by a count of the total number of lines, the Java version has about 22% more lines than its FX Script counterpart. Furthermore, there was an additional expectation that the Java version would be more verbose in terms of the total number of characters.  In fact the preceding data shows that on average the Java source files contain fewer characters per line than the FX files.  But that's not the whole story.  Upon further examination, the FX Script source files had a disproportionate number of blank characters.  Why?  Because of the nature of how one develops JavaFX Script code.  The object literal dominates FX Script code.  Its not uncommon to see object literals indented halfway across the page, consuming lots of meaningless space characters. RAM consumption Not the most scientific analysis, memory usage for the application was examined on a Windows Vista system by running the Windows Task Manager and viewing how much memory was being consumed by the Sudoku version in question. Roughly speaking, the FX script version, after startup, had a RAM footprint of about 90MB and remained pretty much the same size. The Java version started out at about 55MB and maintained that size throughout its execution. What About Binding? Arguably, the most striking observation about the conversion from JavaFX Script to JavaFX 2.0 concerned the need for data synchronization, or lack thereof. In JavaFX Script, the primary means to synchronize data is via the bind expression (using the “bind” keyword), and perhaps to a lesser extent it’s “on replace” cousin. The bind keyword does not exist in Java, so for JavaFX 2.0 a Data Binding API has been introduced as a replacement. To give a feel for the difference between the two versions of the Sudoku program, the table that follows indicates how many binds were required for each source file. For JavaFX Script files, this was ascertained by simply counting the number of occurrences of the bind keyword. As can be seen, binding had been used frequently in the JavaFX Script version (and does not take into consideration an additional half dozen or so “on replace” triggers). The JavaFX 2.0 program achieves the same functionality as the original JavaFX Script version, yet the equivalent of binding was only needed twice throughout the Java version of the source code. JavaFX Script source file Number of Binds JavaFX Next Java source file Number of “Binds” ArrowKey.java 0 Board.fx 1 Board.java 0 BoardNode.fx 7 BoardNode.java 0 ChooseNumberNode.fx 11 ChooseNumberNode.java 0 CloseButtonNode.fx 6 CloseButton.java 0 CustomNodeWithKeyTraversal.java 0 FunctionPtr.java 0 Globals.java 0 Grouping.fx 0 HowToPlayNode.fx 7 HowToPlayNode.java 0 IconButtonNode.fx 9 IconButtonNode.java 0 Main.fx 1 Main.java 0 Main_Mobile.fx 1 SliderNode.fx 6 SliderNode.java 1 Space.fx 0 Space.java 0 SpaceNode.fx 9 SpaceNode.java 1 TraversalHelper.fx 0 Total 58 2 Conclusions As the JavaFX 2.0 technology is so new, and experience with the platform is the same, it is possible and indeed probable that some of the observations noted in the preceding article may not apply across other attempts at migrating applications. That being said, this first experience indicates that the migrated Java code will likely be larger, though not extensively so, than the original Java FX Script source. Furthermore, although very important, it appears that the requirements for data synchronization via binding, may be significantly less with the new platform.

  Read the article

• How to Easily Put a Windows PC into Kiosk Mode With Assigned Access

  - by Chris Hoffman

  Windows 8.1′s Assigned Access feature allows you to easily lock a Windows PC to a single application, such as a web browser. This feature makes it easy for anyone to configure Windows 8.1 devices as point-of-sale or other kiosk systems. In the past, setting up a Windows PC in kiosk mode involved much more work, requiring the use of third-party software, group policy, or Linux distributions designed around kiosk mode. Assigned Access is available on Windows 8.1 RT, Windows 8.1 Professional, and Windows 8.1 Enterprise. The standard edition of Windows 8.1 doesn’t support Assigned Access. Create a User Account for Assigned Access Rather than turn your entire computer into a locked-down kiosk system, Assigned Access allows you to create a separate user account that can only launch a single app — such as a web browser. To set this up, you must be logged into Windows as a user with administrator permissions. First, open the PC settings app — swipe in from the right or press Windows Key + C to open the charms bar, tap Settings, and tap Change PC settings. In the PC settings app, select Accounts and select Other accounts. Use the Add an account button to create a new Windows account. Select  the “Sign in without a Microsoft account” option and select Local account to create a local user account. You could also create a Microsoft account, but you may not want to do this if you just want a locked-down account with only browser access. If you need to install apps from the Windows Store to use in Assigned Access mode, you’ll have to set up a Microsoft account instead of a local account. A local account will still allow you access to the preinstalled apps, such as Internet Explorer. You may want to create a user account with a blank password. This would make it simple for anyone to access kiosk mode, even if the system becomes locked or needs to be rebooted. The account will be created as a standard user account with limited permissions. Leave it as a standard user account — don’t make it an administrator account. Set Up Assigned Access Once you’ve created an account, you’ll first need to sign into it. If you don’t, you’ll see a “This account has no apps” message when trying to enable Assigned Access. Go back to the welcome screen, log in to the new account you created, and allow Windows to go through the first-time account setup process. If you want to use a non-default app in kiosk mode, install it while logged in as that user account. Once you’re done, log out of the other account, log back in as your administrator account, and go back to the Other accounts screen. Click the Set up an account for assigned access option to continue. Select the user account you created and select the app you want to limit the account to. For a web-based kiosk, this can be a web browser such as the Modern version of Internet Explorer. Businesses can also create their own Modern apps and set them to run in kiosk mode in this way. Note that Microsoft’s documentation says “web browsers are not good choices for assigned access” because they require more permissions than average Modern (or “Windows Store”) apps. However, if you want to provide a kiosk for web-browsing, using Assigned Access is a much better option than using Guest Mode and offering up a full Windows desktop. When you’re done, restart your PC and log in as the Assigned Access account. Windows will automatically open the app you chose and won’t allow a user to leave that app. Standard Windows 8 features like the charms bar, app switcher, and Start screen won’t appear. Pressing the Windows key once will do nothing. To sign out of Assigned Access mode, press the Windows key five times — quickly — while signed in. You’ll be sent back to the standard login screen. The account will actually still be logged in and the app will remain running — this method just “locks” the screen and allows another user to log in. Automatically Log Into Assigned Access Whenever your Windows device boots, you can log into the Assigned Access account and turn it into a kiosk system. While this isn’t ideal for all kiosk systems, you may want the device to automatically launch the specific app when it boots without requiring any login process. To do so, you’ll just need to have Windows automatically log into the Assigned Access account when it boots. This option is hidden and not available in the standard Control Panel. You’ll need to use the hidden netplwiz Control Panel tool to set up automatic login on boot. If you didn’t create a password for the user account, leave the Password field empty while configuring this. Security Considerations If you’re using this feature to turn a Windows 8.1 system into a kiosk and leaving it open to the public, remember to consider security. Anyone could come up to the system, press the Windows key five times, and try to log into your standard administrator user account. Ensure the administrator user account has a strong password so people won’t be able to get past the kiosk system’s limitations and tamper with the system. Even Windows 8′s detractors have to admit that it’s an ideal system for a touch-screen kiosk device, running either a browser or another specific application. Assigned Access finally makes this easy to set up on Windows systems in the real world — no IT experience, third-party software, or Linux distributions necessary.     

  Read the article

• C#: LINQ vs foreach - Round 1.

  - by James Michael Hare

  So I was reading Peter Kellner's blog entry on Resharper 5.0 and its LINQ refactoring and thought that was very cool.  But that raised a point I had always been curious about in my head -- which is a better choice: manual foreach loops or LINQ?    The answer is not really clear-cut.  There are two sides to any code cost arguments: performance and maintainability.  The first of these is obvious and quantifiable.  Given any two pieces of code that perform the same function, you can run them side-by-side and see which piece of code performs better.   Unfortunately, this is not always a good measure.  Well written assembly language outperforms well written C++ code, but you lose a lot in maintainability which creates a big techncial debt load that is hard to offset as the application ages.  In contrast, higher level constructs make the code more brief and easier to understand, hence reducing technical cost.   Now, obviously in this case we're not talking two separate languages, we're comparing doing something manually in the language versus using a higher-order set of IEnumerable extensions that are in the System.Linq library.   Well, before we discuss any further, let's look at some sample code and the numbers.  First, let's take a look at the for loop and the LINQ expression.  This is just a simple find comparison:       // find implemented via LINQ     public static bool FindViaLinq(IEnumerable<int> list, int target)     {         return list.Any(item => item == target);     }         // find implemented via standard iteration     public static bool FindViaIteration(IEnumerable<int> list, int target)     {         foreach (var i in list)         {             if (i == target)             {                 return true;             }         }           return false;     }   Okay, looking at this from a maintainability point of view, the Linq expression is definitely more concise (8 lines down to 1) and is very readable in intention.  You don't have to actually analyze the behavior of the loop to determine what it's doing.   So let's take a look at performance metrics from 100,000 iterations of these methods on a List<int> of varying sizes filled with random data.  For this test, we fill a target array with 100,000 random integers and then run the exact same pseudo-random targets through both searches.                       List<T> On 100,000 Iterations     Method      Size     Total (ms)  Per Iteration (ms)  % Slower     Any         10       26          0.00046             30.00%     Iteration   10       20          0.00023             -     Any         100      116         0.00201             18.37%     Iteration   100      98          0.00118             -     Any         1000     1058        0.01853             16.78%     Iteration   1000     906         0.01155             -     Any         10,000   10,383      0.18189             17.41%     Iteration   10,000   8843        0.11362             -     Any         100,000  104,004     1.8297              18.27%     Iteration   100,000  87,941      1.13163             -   The LINQ expression is running about 17% slower for average size collections and worse for smaller collections.  Presumably, this is due to the overhead of the state machine used to track the iterators for the yield returns in the LINQ expressions, which seems about right in a tight loop such as this.   So what about other LINQ expressions?  After all, Any() is one of the more trivial ones.  I decided to try the TakeWhile() algorithm using a Count() to get the position stopped like the sample Pete was using in his blog that Resharper refactored for him into LINQ:       // Linq form     public static int GetTargetPosition1(IEnumerable<int> list, int target)     {         return list.TakeWhile(item => item != target).Count();     }       // traditionally iterative form     public static int GetTargetPosition2(IEnumerable<int> list, int target)     {         int count = 0;           foreach (var i in list)         {             if(i == target)             {                 break;             }               ++count;         }           return count;     }   Once again, the LINQ expression is much shorter, easier to read, and should be easier to maintain over time, reducing the cost of technical debt.  So I ran these through the same test data:                       List<T> On 100,000 Iterations     Method      Size     Total (ms)  Per Iteration (ms)  % Slower     TakeWhile   10       41          0.00041             128%     Iteration   10       18          0.00018             -     TakeWhile   100      171         0.00171             88%     Iteration   100      91          0.00091             -     TakeWhile   1000     1604        0.01604             94%     Iteration   1000     825         0.00825             -     TakeWhile   10,000   15765       0.15765             92%     Iteration   10,000   8204        0.08204             -     TakeWhile   100,000  156950      1.5695              92%     Iteration   100,000  81635       0.81635             -     Wow!  I expected some overhead due to the state machines iterators produce, but 90% slower?  That seems a little heavy to me.  So then I thought, well, what if TakeWhile() is not the right tool for the job?  The problem is TakeWhile returns each item for processing using yield return, whereas our for-loop really doesn't care about the item beyond using it as a stop condition to evaluate. So what if that back and forth with the iterator state machine is the problem?  Well, we can quickly create an (albeit ugly) lambda that uses the Any() along with a count in a closure (if a LINQ guru knows a better way PLEASE let me know!), after all , this is more consistent with what we're trying to do, we're trying to find the first occurence of an item and halt once we find it, we just happen to be counting on the way.  This mostly matches Any().       // a new method that uses linq but evaluates the count in a closure.     public static int TakeWhileViaLinq2(IEnumerable<int> list, int target)     {         int count = 0;         list.Any(item =>             {                 if(item == target)                 {                     return true;                 }                   ++count;                 return false;             });         return count;     }     Now how does this one compare?                         List<T> On 100,000 Iterations     Method         Size     Total (ms)  Per Iteration (ms)  % Slower     TakeWhile      10       41          0.00041             128%     Any w/Closure  10       23          0.00023             28%     Iteration      10       18          0.00018             -     TakeWhile      100      171         0.00171             88%     Any w/Closure  100      116         0.00116             27%     Iteration      100      91          0.00091             -     TakeWhile      1000     1604        0.01604             94%     Any w/Closure  1000     1101        0.01101             33%     Iteration      1000     825         0.00825             -     TakeWhile      10,000   15765       0.15765             92%     Any w/Closure  10,000   10802       0.10802             32%     Iteration      10,000   8204        0.08204             -     TakeWhile      100,000  156950      1.5695              92%     Any w/Closure  100,000  108378      1.08378             33%     Iteration      100,000  81635       0.81635             -     Much better!  It seems that the overhead of TakeAny() returning each item and updating the state in the state machine is drastically reduced by using Any() since Any() iterates forward until it finds the value we're looking for -- for the task we're attempting to do.   So the lesson there is, make sure when you use a LINQ expression you're choosing the best expression for the job, because if you're doing more work than you really need, you'll have a slower algorithm.  But this is true of any choice of algorithm or collection in general.     Even with the Any() with the count in the closure it is still about 30% slower, but let's consider that angle carefully.  For a list of 100,000 items, it was the difference between 1.01 ms and 0.82 ms roughly in a List<T>.  That's really not that bad at all in the grand scheme of things.  Even running at 90% slower with TakeWhile(), for the vast majority of my projects, an extra millisecond to save potential errors in the long term and improve maintainability is a small price to pay.  And if your typical list is 1000 items or less we're talking only microseconds worth of difference.   It's like they say: 90% of your performance bottlenecks are in 2% of your code, so over-optimizing almost never pays off.  So personally, I'll take the LINQ expression wherever I can because they will be easier to read and maintain (thus reducing technical debt) and I can rely on Microsoft's development to have coded and unit tested those algorithm fully for me instead of relying on a developer to code the loop logic correctly.   If something's 90% slower, yes, it's worth keeping in mind, but it's really not until you start get magnitudes-of-order slower (10x, 100x, 1000x) that alarm bells should really go off.  And if I ever do need that last millisecond of performance?  Well then I'll optimize JUST THAT problem spot.  To me it's worth it for the readability, speed-to-market, and maintainability.

  Read the article

• Using Node.js as an accelerator for WCF REST services

  - by Elton Stoneman

  Node.js is a server-side JavaScript platform "for easily building fast, scalable network applications". It's built on Google's V8 JavaScript engine and uses an (almost) entirely async event-driven processing model, running in a single thread. If you're new to Node and your reaction is "why would I want to run JavaScript on the server side?", this is the headline answer: in 150 lines of JavaScript you can build a Node.js app which works as an accelerator for WCF REST services*. It can double your messages-per-second throughput, halve your CPU workload and use one-fifth of the memory footprint, compared to the WCF services direct.   Well, it can if: 1) your WCF services are first-class HTTP citizens, honouring client cache ETag headers in request and response; 2) your services do a reasonable amount of work to build a response; 3) your data is read more often than it's written. In one of my projects I have a set of REST services in WCF which deal with data that only gets updated weekly, but which can be read hundreds of times an hour. The services issue ETags and will return a 304 if the client sends a request with the current ETag, which means in the most common scenario the client uses its local cached copy. But when the weekly update happens, then all the client caches are invalidated and they all need the same new data. Then the service will get hundreds of requests with old ETags, and they go through the full service stack to build the same response for each, taking up threads and processing time. Part of that processing means going off to a database on a separate cloud, which introduces more latency and downtime potential.   We can use ASP.NET output caching with WCF to solve the repeated processing problem, but the server will still be thread-bound on incoming requests, and to get the current ETags reliably needs a database call per request. The accelerator solves that by running as a proxy - all client calls come into the proxy, and the proxy routes calls to the underlying REST service. We could use Node as a straight passthrough proxy and expect some benefit, as the server would be less thread-bound, but we would still have one WCF and one database call per proxy call. But add some smart caching logic to the proxy, and share ETags between Node and WCF (so the proxy doesn't even need to call the servcie to get the current ETag), and the underlying service will only be invoked when data has changed, and then only once - all subsequent client requests will be served from the proxy cache.   I've built this as a sample up on GitHub: NodeWcfAccelerator on sixeyed.codegallery. Here's how the architecture looks:     The code is very simple. The Node proxy runs on port 8010 and all client requests target the proxy. If the client request has an ETag header then the proxy looks up the ETag in the tag cache to see if it is current - the sample uses memcached to share ETags between .NET and Node. If the ETag from the client matches the current server tag, the proxy sends a 304 response with an empty body to the client, telling it to use its own cached version of the data. If the ETag from the client is stale, the proxy looks for a local cached version of the response, checking for a file named after the current ETag. If that file exists, its contents are returned to the client as the body in a 200 response, which includes the current ETag in the header. If the proxy does not have a local cached file for the service response, it calls the service, and writes the WCF response to the local cache file, and to the body of a 200 response for the client. So the WCF service is only troubled if both client and proxy have stale (or no) caches.   The only (vaguely) clever bit in the sample is using the ETag cache, so the proxy can serve cached requests without any communication with the underlying service, which it does completely generically, so the proxy has no notion of what it is serving or what the services it proxies are doing. The relative path from the URL is used as the lookup key, so there's no shared key-generation logic between .NET and Node, and when WCF stores a tag it also stores the "read" URL against the ETag so it can be used for a reverse lookup, e.g:   Key Value /WcfSampleService/PersonService.svc/rest/fetch/3 "28cd4796-76b8-451b-adfd-75cb50a50fa6" "28cd4796-76b8-451b-adfd-75cb50a50fa6" /WcfSampleService/PersonService.svc/rest/fetch/3    In Node we read the cache using the incoming URL path as the key and we know that "28cd4796-76b8-451b-adfd-75cb50a50fa6" is the current ETag; we look for a local cached response in /caches/28cd4796-76b8-451b-adfd-75cb50a50fa6.body (and the corresponding .header file which contains the original service response headers, so the proxy response is exactly the same as the underlying service). When the data is updated, we need to invalidate the ETag cache – which is why we need the reverse lookup in the cache. In the WCF update service, we don't need to know the URL of the related read service - we fetch the entity from the database, do a reverse lookup on the tag cache using the old ETag to get the read URL, update the new ETag against the URL, store the new reverse lookup and delete the old one.   Running Apache Bench against the two endpoints gives the headline performance comparison. Making 1000 requests with concurrency of 100, and not sending any ETag headers in the requests, with the Node proxy I get 102 requests handled per second, average response time of 975 milliseconds with 90% of responses served within 850 milliseconds; going direct to WCF with the same parameters, I get 53 requests handled per second, mean response time of 1853 milliseconds, with 90% of response served within 3260 milliseconds. Informally monitoring server usage during the tests, Node maxed at 20% CPU and 20Mb memory; IIS maxed at 60% CPU and 100Mb memory.   Note that the sample WCF service does a database read and sleeps for 250 milliseconds to simulate a moderate processing load, so this is *not* a baseline Node-vs-WCF comparison, but for similar scenarios where the  service call is expensive but applicable to numerous clients for a long timespan, the performance boost from the accelerator is considerable.     * - actually, the accelerator will work nicely for any HTTP request, where the URL (path + querystring) uniquely identifies a resource. In the sample, there is an assumption that the ETag is a GUID wrapped in double-quotes (e.g. "28cd4796-76b8-451b-adfd-75cb50a50fa6") – which is the default for WCF services. I use that assumption to name the cache files uniquely, but it is a trivial change to adapt to other ETag formats.

  Read the article

• In Case You Weren’t There: Blogwell NYC

  - by Mike Stiles

  0 0 1 1009 5755 Vitrue 47 13 6751 14.0 Normal 0 false false false EN-US JA X-NONE /* Style Definitions */ table.MsoNormalTable {mso-style-name:"Table Normal"; mso-tstyle-rowband-size:0; mso-tstyle-colband-size:0; mso-style-noshow:yes; mso-style-priority:99; mso-style-parent:""; mso-padding-alt:0in 5.4pt 0in 5.4pt; mso-para-margin:0in; mso-para-margin-bottom:.0001pt; mso-pagination:widow-orphan; font-size:12.0pt; font-family:"Times New Roman";} Your roving reporter roved out to another one of Socialmedia.org’s fantastic Blogwell events, this time in NYC. As Central Park and incredible weather beckoned, some of the biggest brand names in the world gathered to talk about how they’re incorporating social into marketing and CRM, as well as extending social across their entire organizations internally. Below we present a collection of the live tweets from many of the key sessions GE @generalelectricJon Lombardo, Leader of Social Media COE How GE builds and extends emotional connections with consumers around health and reaps the benefits of increased brand equity in the process. GE has a social platform around Healthyimagination to create better health for people. If you and a friend are trying to get healthy together, you’ll do better. Health is inherently. Get health challenges via Facebook and share with friends to achieve goals together. They’re creating an emotional connection around the health context. You don’t influence people at large. Your sphere of real influence is around 5-10 people. They find relevant conversations about health on Twitter and engage sounding like a friend, not a brand. Why would people share on behalf of a brand? Because you tapped into an activity and emotion they’re already having. To create better habits in health, GE gave away inexpensive, relevant gifts related to their goals. Create the context, give the relevant gift, get social acknowledgment for giving it. What you get when you get acknowledgment for your engagement and gift is user generated microcontent. GE got 12,000 unique users engaged and 1400 organic posts with the healthy gift campaign. The Dow Chemical Company @DowChemicalAbby Klanecky, Director of Digital & Social Media Learn how Dow Chemical is finding, training, and empowering their scientists to be their storytellers in social media. There are 1m jobs coming open in science. Only 200k are qualified for them. Dow Chemical wanted to use social to attract and talk to scientists. Dow Chemical decided to use real scientists as their storytellers. Scientists are incredibly passionate, the key ingredient of a great storyteller. Step 1 was getting scientists to focus on a few platforms, blog, Twitter, LinkedIn. Dow Chemical social flow is Core Digital Team - #CMs – ambassadors – advocates. The scientists were trained in social etiquette via practice scenarios. It’s not just about sales. It’s about growing influence and the business. Dow Chemical trained about 100 scientists, 55 are active and there’s a waiting list for the next sessions. In person social training produced faster results and better participation. Sometimes you have to tell pieces of the story instead of selling your execs on the whole vision. Social Media Ethics Briefing: Staying Out of TroubleAndy Sernovitz, CEO @SocialMediaOrg How do we get people to share our message for us? We have to have their trust. The difference between being honest and being sleazy is disclosure. Disclosure does not hurt the effectiveness of your marketing. No one will get mad if you tell them up front you’re a paid spokesperson for a company. It’s a legal requirement by the FTC, it’s the law, to disclose if you’re being paid for an endorsement. Require disclosure and truthfulness in all your social media outreach. Don’t lie to people. Monitor the conversation and correct misstatements. Create social media policies and training programs. If you want to stay safe, never pay cash for social media. Money changes everything. As soon as you pay, it’s not social media, it’s advertising. Disclosure, to the feds, means clear, conspicuous, and understandable to the average reader. This phrase will keep you in the clear, “I work for ___ and this is my personal opinion.” Who are you? Were you paid? Are you giving an honest opinion based on a real experience? You as a brand are responsible for what an agency or employee or contactor does in your behalf. SocialMedia.org makes available a Disclosure Best Practices Toolkit. Socialmedia.org/disclosure. The point is to not ethically mess up and taint social media as happened to e-mail. Not only is the FTC cracking down, so is Google and Facebook. Visa @VisaNewsLucas Mast, Senior Business Leader, Global Corporate Social Media Visa built a mobile studio for the Olympics for execs and athletes. They wanted to do postcard style real time coverage of Visa’s Olympics sponsorships, and on a shoestring. Challenges included Olympic rules, difficulty getting interviews, time zone trouble, and resourcing. Another problem was they got bogged down with their own internal approval processes. Despite all the restrictions, they created and published a variety of and fair amount of content. They amassed 1000+ views of videos posted to the Visa Communication YouTube channel. Less corporate content yields more interest from media outlets and bloggers. They did real world video demos of how their products work in the field vs. an exec doing a demo in a studio. Don’t make exec interview videos dull and corporate. Keep answers short, shoot it in an interesting place, do takes until they’re comfortable and natural. Not everything will work. Not everything will get a retweet. But like the lottery, you can’t win if you don’t play. Promoting content is as important as creating it. McGraw-Hill Companies @McGrawHillCosPatrick Durando, Senior Director of Global New Media McGraw-Hill has 26,000 employees. McGraw-Hill created a social intranet called Buzz. Intranets create operational efficiency, help product dev, facilitate crowdsourcing, and breaks down geo silos. Intranets help with talent development, acquisition, retention. They replaced the corporate directory with their own version of LinkedIn. The company intranet has really cut down on the use of email. Long email threats become organized, permanent social discussions. The intranet is particularly useful in HR for researching and getting answers surrounding benefits and policies. Using a profile on your company intranet can establish and promote your internal professional brand. If you’re going to make an intranet, it has to look great, work great, and employees are going have to want to go there. You can’t order them to like it. 

  Read the article

• Pet Peeves with the Windows Phone 7 Marketplace

  - by Bil Simser

  Have you ever noticed how something things just gnaw at your very being. This is the case with the WP7 marketplace, the Zune software, and the things that drive me batshit crazy with a side of fries. To go. I wanted to share. XBox Live is Not the Centre of the Universe Okay, it’s fine that the Zune software has an XBox live tag for games so can see them clearly but do we really need to have it shoved down our throats. On every click? Click on Games in the marketplace: The first thing that it defaults to on the filters on the right is XBox Live: Okay. Fine. However if you change it (say to Paid) then click onto a title when you come back from that title is the filter still set to Paid? No. It’s back to XBox Live again. Really? Give us a break. If you change to any filter on any other genre then click on the selected title, it doesn’t revert back to anything. It stays on the selection you picked. Let’s be fair here. The Games genre should behave just like every other one. If I pick Paid then when I come back to the list please remember that. Double Dipping On the subject of XBox Live titles, Microsoft (and developers who have an agreement with Microsoft to produce Live titles, which generally rules out indie game developers) is double dipping with regards to exposure of their titles. Here’s the Puzzle and Trivia Game section on the Marketplace for XBox Live titles: And here’s the same category filtered on Paid titles: See the problem? Two indie titles while the rest are XBox Live ones. So while XBL has it’s filter, they also get to showcase their wares in the Paid and Free filters as well. If you’re going to have an XBox Live filter then use it and stop pushing down indie titles until they’re off the screen (on some genres this is already the case). Free and Paid titles should be just that and not include XBox Live ones. If you’re really stoked that people can’t find the Free XBox Live titles vs. the paid ones, then create a Free XBox Live filter and a Paid XBox Live filter. I don’t think we would mind much. Whose Trial is it Anyways? You might notice apps in the marketplace with titles like “My Fart App Professional Lite” or “Silicon Lamb Spleen Builder Free”. When you submit and app to the marketplace it can either be free or paid. If it’s a paid app you also have the option to submit it with Trial capabilities. It’s up to you to decide what you offer in the trial version but trial versions can be purchased from within the app so after someone trys out your app (for free) and wants to unlock the Super Secret Obama Spy Ring Level, they can just go to the marketplace from your app (if you built that functionality in) and upgrade to the paid version. However it creates a rift of sorts when it comes to visibility. Some developers go the route of the paid app with a trial version, others decide to submit *two* apps instead of one. One app is the “Free” or “Lite” verions and the other is the paid version. Why go to the hassle of submitting two apps when you can just create a trial version in the same app? Again, visibility. There’s no way to tell Paid apps with Trial versions and ones without (it’s an option, you don’t have to provide trial versions, although I think it’s a good idea). However there is a way to see the Free apps from the Paid ones so some submit the two apps and have the Free version have links to buy the paid one (again through the Marketplace tasks in the API). What we as developers need for visibility is a new filter. Trial. That’s it. It would simply filter on Paid apps that have trial capabilities and surface up those apps just like the free ones. If Microsoft added this filter to the marketplace, it would eliminate the need for people to submit their “Free” and “Lite” versions and make it easier for the developer not to have to maintain two systems. I mean, is it really that hard? Can’t be any more difficult than the XBox Live Filter that’s already there. Location is Everything The last thing on my bucket list is about location. When I launch Zune I’m running in my native location setting, Canada. What’s great is that I navigate to the Travel Tools section where I have one of my apps and behold the splendour that I see: There are my apps in the number 1 and number 4 slot for top selling in that category. I show it to my wife to make up for the sleepless nights writing this stuff and we dance around and celebrate. Then I change my location on my operation system to United States and re-launch Zune. WTF? My flight app has slipped to the 10th spot (I’m only showing 4 across here out of the 7 in Zune) and my border check app that was #1 is now in the 32nd spot! End of celebration. Not only is relevance being looked at here, I value the comments people make on may apps as do most developers. I want to respond to them and show them that I’m listening. The next version of my border app will provide multiple camera angles. However when I’m running in my native Canada location, I only see two reviews. Changing over to United States I see fourteen! While there are tools out there to provide with you a unified view, I shouldn’t have to rely on them. My own Zune desktop software should allow me to see everything. I realize that some developers will submit an app and only target it for some locations and that’s their choice. However I shouldn’t have to jump through hoops to see what apps are ahead of mine, or see people comments and ratings. Another proposal. Either unify the marketplace (i.e. when I’m looking at it show me everything combined) or let me choose a filter. I think the first option might be difficult as you’re trying to average out top selling apps across all markets and have to deal with some apps that have been omitted from some markets. Although I think you could come up with a set of use cases that would handle that, maybe that’s too much work. At the very least, let us developers view the markets in a drop down or something from within the Zune desktop. Having to shut down Zune, change our location, and re-launch Zune to see other perspectives is just too onerous. A Call to Action These are just one mans opinion. Do you agree? Disagree? Feel hungry for a bacon sandwich? Let everyone know via the comments below. Perhaps someone from Microsoft will be reading and take some of these ideas under advisement. Maybe not, but at least let’s get the word out that we really want to see some change. Egypt can do it, why not WP7 developers!

  Read the article

• The Faces in the Crowdsourcing

  - by Applications User Experience

  By Jeff Sauro, Principal Usability Engineer, Oracle Imagine having access to a global workforce of hundreds of thousands of people who can perform tasks or provide feedback on a design quickly and almost immediately. Distributing simple tasks not easily done by computers to the masses is called "crowdsourcing" and until recently was an interesting concept, but due to practical constraints wasn't used often. Enter Amazon.com. For five years, Amazon has hosted a service called Mechanical Turk, which provides an easy interface to the crowds. The service has almost half a million registered, global users performing a quarter of a million human intelligence tasks (HITs). HITs are submitted by individuals and companies in the U.S. and pay from $.01 for simple tasks (such as determining if a picture is offensive) to several dollars (for tasks like transcribing audio). What do we know about the people who toil away in this digital crowd? Can we rely on the work done in this anonymous marketplace? A rendering of the actual Mechanical Turk (from Wikipedia) Knowing who is behind Amazon's Mechanical Turk is fitting, considering the history of the actual Mechanical Turk. In the late 1800's, a mechanical chess-playing machine awed crowds as it beat master chess players in what was thought to be a mechanical miracle. It turned out that the creator, Wolfgang von Kempelen, had a small person (also a chess master) hiding inside the machine operating the arms to provide the illusion of automation. The field of human computer interaction (HCI) is quite familiar with gathering user input and incorporating it into all stages of the design process. It makes sense then that Mechanical Turk was a popular discussion topic at the recent Computer Human Interaction usability conference sponsored by the Association for Computing Machinery in Atlanta. It is already being used as a source for input on Web sites (for example, Feedbackarmy.com) and behavioral research studies. Two papers shed some light on the faces in this crowd. One paper tells us about the shifting demographics from mostly stay-at-home moms to young men in India. The second paper discusses the reliability and quality of work from the workers. Just who exactly would spend time doing tasks for pennies? In "Who are the crowdworkers?" University of California researchers Ross, Silberman, Zaldivar and Tomlinson conducted a survey of Mechanical Turk worker demographics and compared it to a similar survey done two years before. The initial survey reported workers consisting largely of young, well-educated women living in the U.S. with annual household incomes above $40,000. The more recent survey reveals a shift in demographics largely driven by an influx of workers from India. Indian workers went from 5% to over 30% of the crowd, and this block is largely male (two-thirds) with a higher average education than U.S. workers, and 64% report an annual income of less than $10,000 (keeping in mind $1 has a lot more purchasing power in India). This shifting demographic certainly has implications as language and culture can play critical roles in the outcome of HITs. Of course, the demographic data came from paying Turkers $.10 to fill out a survey, so there is some question about both a self-selection bias (characteristics which cause Turks to take this survey may be unrepresentative of the larger population), not to mention whether we can really trust the data we get from the crowd. Crowds can perform tasks or provide feedback on a design quickly and almost immediately for usability testing. (Photo attributed to victoriapeckham Flikr While having immediate access to a global workforce is nice, one major problem with Mechanical Turk is the incentive structure. Individuals and companies that deploy HITs want quality responses for a low price. Workers, on the other hand, want to complete the task and get paid as quickly as possible, so that they can get on to the next task. Since many HITs on Mechanical Turk are surveys, how valid and reliable are these results? How do we know whether workers are just rushing through the multiple-choice responses haphazardly answering? In "Are your participants gaming the system?" researchers at Carnegie Mellon (Downs, Holbrook, Sheng and Cranor) set up an experiment to find out what percentage of their workers were just in it for the money. The authors set up a 30-minute HIT (one of the more lengthy ones for Mechanical Turk) and offered a very high $4 to those who qualified and $.20 to those who did not. As part of the HIT, workers were asked to read an email and respond to two questions that determined whether workers were likely rushing through the HIT and not answering conscientiously. One question was simple and took little effort, while the second question required a bit more work to find the answer. Workers were led to believe other factors than these two questions were the qualifying aspect of the HIT. Of the 2000 participants, roughly 1200 (or 61%) answered both questions correctly. Eighty-eight percent answered the easy question correctly, and 64% answered the difficult question correctly. In other words, about 12% of the crowd were gaming the system, not paying enough attention to the question or making careless errors. Up to about 40% won't put in more than a modest effort to get paid for a HIT. Young men and those that considered themselves in the financial industry tended to be the most likely to try to game the system. There wasn't a breakdown by country, but given the demographic information from the first article, we could infer that many of these young men come from India, which makes language and other cultural differences a factor. These articles raise questions about the role of crowdsourcing as a means for getting quick user input at low cost. While compensating users for their time is nothing new, the incentive structure and anonymity of Mechanical Turk raises some interesting questions. How complex of a task can we ask of the crowd, and how much should these workers be paid? Can we rely on the information we get from these professional users, and if so, how can we best incorporate it into designing more usable products? Traditional usability testing will still play a central role in enterprise software. Crowdsourcing doesn't replace testing; instead, it makes certain parts of gathering user feedback easier. One can turn to the crowd for simple tasks that don't require specialized skills and get a lot of data fast. As more studies are conducted on Mechanical Turk, I suspect we will see crowdsourcing playing an increasing role in human computer interaction and enterprise computing. References: Downs, J. S., Holbrook, M. B., Sheng, S., and Cranor, L. F. 2010. Are your participants gaming the system?: screening mechanical turk workers. In Proceedings of the 28th international Conference on Human Factors in Computing Systems (Atlanta, Georgia, USA, April 10 - 15, 2010). CHI '10. ACM, New York, NY, 2399-2402. Link: http://doi.acm.org/10.1145/1753326.1753688 Ross, J., Irani, L., Silberman, M. S., Zaldivar, A., and Tomlinson, B. 2010. Who are the crowdworkers?: shifting demographics in mechanical turk. In Proceedings of the 28th of the international Conference Extended Abstracts on Human Factors in Computing Systems (Atlanta, Georgia, USA, April 10 - 15, 2010). CHI EA '10. ACM, New York, NY, 2863-2872. Link: http://doi.acm.org/10.1145/1753846.1753873

  Read the article

• Communities - The importance of exchange and discussion

  Communication with your environment is an essential part of everyone's life. And it doesn't matter whether you are actually living in a rural area in the middle of nowhere, within the pulsating heart of a big city, or in my case on a wonderful island in the Indian Ocean. The ability to exchange your thoughts, your experience and your worries with another person helps you to get different points of view and new ideas on how to resolve an issue you might be confronted with. Benefits of community work What happens to be common sense in your daily life, also applies to your work environment. Working in IT, or ICT as it is called in Mauritius, requires a lot of reading and learning. Not only during your lectures at the university but with your colleagues in a project assignment and hopefully with 'unknown' pals in the universe of online communities. At least I can say that I learned quite a lot from other developers code, their responses in various forums, their numerous blog articles, and while attending local user group meetings. When I started to work as a professional software developer (or engineer some may say) years ago I immediately checked the existence of communities on the programming language, the database technology and other vital information on software development in general. Luckily, it wasn't too difficult to find. My employer had a subscription of the monthly magazines and newsletters of a national organisation which also run the biggest forum in that area. Getting in touch with other developers and reading their common problems but also solutions was a huge benefit to my growth. Image courtesy of Michael Kappel (CC BY-NC 2.0) Active participation and regular contribution to this community gave me some nice advantages, too. Within three years I was listed as a conference speaker at the annual developer's conference and provided several sessions on different topics during consecutive years. Back in 2004, I took over the responsibility and management of the monthly meetings of a regional user group, and organised it for more than two years. Furthermore, I was invited to the newly-founded community program of Microsoft Germany (Community Leader/Insider Program - CLIP). My website on Active FoxPro Pages was nominated in the second batch of online communities. Due to my community work and providing advice to others, I had the honour to be awarded as Microsoft Most Valuable Professional (MVP) - Visual Developer for Visual FoxPro in the years 2006 and 2007. It was a great experience to meet with other like-minded people and I'm really grateful for that. Just in case, more details are listed in my Curriculum Vitae. But this all changed when I moved to Mauritius... Cyber island Mauritius? During the first months in Mauritius I was way too busy to think about community activities at all. First of all, there was the new company that had to be set up, the new staff had to be trained and of course the communication work-flows and so on with the project managers back in Germany had to be sorted out, too. Second, I had to get a grip of my private matters like getting the basics for my new household or exploring the neighbourhood, and last but not least I needed a break from the hectic and intensive work prior to my departure. As soon as the sea literally calmed down, I started to have conversations with my colleagues about communities and user groups. Sadly, it turned out that there were none, or at least no one was aware of any at that time. Oh oh, what did I do? Anyway, having this kind of background and very positive experience with off-line and on-line activities I decided for myself that some day I'm going to found a community in Mauritius for all kind of IT/ICT-related fields. The main focus might be on software development but not on a certain technology or methodology. It was clear to me that it should be an open infrastructure and anyone is welcome to join, to experience, to share and to contribute if they would like to. That was the idea at that time... Ok, fast-forward to recent events. At the end of October 2012 I was invited to an event called Open Days organised by Microsoft Indian Ocean Islands together with other local partners and resellers. There I got in touch with local Technical Evangelist Arnaud Meslier and we had a good conversation on communities during the breaks. Eventually, I left a good impression on him, as we are having chats on Facebook or Skype irregularly. Well, seeing that my personal and professional surroundings have been settled and running smooth, having that great exchange and contact with Microsoft IOI (again), and being really eager to re-animate my intentions from 2007, I recently founded a new community: Mauritius Software Craftsmanship Community - #MSCC It took me a while to settle down with the name but it was obvious that the community should not be attached to one single technology, like ie. .NET user group, Oracle developers, or Joomla friends (these are fictitious names). There are several other reasons why I came up with 'Craftsmanship' as the core topic of this community. The expression of 'engineering' didn't feel right with the fields covered. Software development in all kind of facets is a craft, and therefore demands a lot of practice but also guidance from more experienced developers. It also includes the process of designing, modelling and drafting the ideas. Has to deal with various types of tests and test methodologies, and of course should be focused on flexible and agile ways of acting. In order to meet and to excel a customer's request for a solution. Next, I was looking for an easy way to handle the organisation of events and meeting appointments. Using all kind of social media platforms like Google+, LinkedIn, Facebook, Xing, etc. I was never really confident about their features of event handling. More by chance I stumbled upon Meetup.com and in combination with the other entities (G+ Communities, FB Pages or in Groups) I am looking forward to advertise and manage all future activities here: Mauritius Software Craftsmanship Community This is a community for those who care and are proud of what they do. For those developers, regardless how experienced they are, who want to improve and master their craft. This is a community for those who believe that being average is just not good enough. I know, there are not many 'craftsmen' yet but it's a start... Let's see how it looks like by the end of the year. There are free smartphone apps for Android and iOS from Meetup.com that allow you to keep track of meetings and to stay informed on latest updates. And last but not least, there will be a Trello workspace to collect and share ideas and provide downloads of slides, etc. Sharing is caring! As mentioned, the #MSCC is present in various social media networks in order to cover as many people as possible here in Mauritius. Following is an overview of the current networks: Twitter - Latest updates and quickies Google+ - Community channel Facebook - Community Page LinkedIn - Community Group Trello - Collaboration workspace to share and develop ideas Hopefully, this covers the majority of computer-related people in Mauritius. Please spread the word about the #MSCC between your colleagues, your friends and other interested 'geeks'. Your future looks bright Running and participating in a user group or any kind of community usually provides quite a number of advantages for anyone. On the one side it is very joyful for me to organise appointments and get in touch with people that might be interested to present a little demo of their projects or their recent problems they had to tackle down, and on the other side there are lots of companies that have various support programs or sponsorships especially tailored for user groups. At the moment, I already have a couple of gimmicks that I would like to hand out in small contests or raffles during one of the upcoming meetings, and as said, companies provide all kind of goodies, books free of charge, or sometimes even licenses for communities. Meeting other software developers or IT guys also opens up your point of view on the local market and there might be interesting projects or job offers available, too. A community like the Mauritius Software Craftsmanship Community is great for freelancers, self-employed, students and of course employees. Meetings will be organised on a regular basis, and I'm open to all kind of suggestions from you. Please leave a comment here in blog or join the conversations in the above mentioned social networks. Let's get this community up and running, my fellow Mauritians!

  Read the article

• NUMA-aware placement of communication variables

  - by Dave

  For classic NUMA-aware programming I'm typically most concerned about simple cold, capacity and compulsory misses and whether we can satisfy the miss by locally connected memory or whether we have to pull the line from its home node over the coherent interconnect -- we'd like to minimize channel contention and conserve interconnect bandwidth. That is, for this style of programming we're quite aware of where memory is homed relative to the threads that will be accessing it. Ideally, a page is collocated on the node with the thread that's expected to most frequently access the page, as simple misses on the page can be satisfied without resorting to transferring the line over the interconnect. The default "first touch" NUMA page placement policy tends to work reasonable well in this regard. When a virtual page is first accessed, the operating system will attempt to provision and map that virtual page to a physical page allocated from the node where the accessing thread is running. It's worth noting that the node-level memory interleaving granularity is usually a multiple of the page size, so we can say that a given page P resides on some node N. That is, the memory underlying a page resides on just one node. But when thinking about accesses to heavily-written communication variables we normally consider what caches the lines underlying such variables might be resident in, and in what states. We want to minimize coherence misses and cache probe activity and interconnect traffic in general. I don't usually give much thought to the location of the home NUMA node underlying such highly shared variables. On a SPARC T5440, for instance, which consists of 4 T2+ processors connected by a central coherence hub, the home node and placement of heavily accessed communication variables has very little impact on performance. The variables are frequently accessed so likely in M-state in some cache, and the location of the home node is of little consequence because a requester can use cache-to-cache transfers to get the line. Or at least that's what I thought. Recently, though, I was exploring a simple shared memory point-to-point communication model where a client writes a request into a request mailbox and then busy-waits on a response variable. It's a simple example of delegation based on message passing. The server polls the request mailbox, and having fetched a new request value, performs some operation and then writes a reply value into the response variable. As noted above, on a T5440 performance is insensitive to the placement of the communication variables -- the request and response mailbox words. But on a Sun/Oracle X4800 I noticed that was not the case and that NUMA placement of the communication variables was actually quite important. For background an X4800 system consists of 8 Intel X7560 Xeons . Each package (socket) has 8 cores with 2 contexts per core, so the system is 8x8x2. Each package is also a NUMA node and has locally attached memory. Every package has 3 point-to-point QPI links for cache coherence, and the system is configured with a twisted ladder "mobius" topology. The cache coherence fabric is glueless -- there's not central arbiter or coherence hub. The maximum distance between any two nodes is just 2 hops over the QPI links. For any given node, 3 other nodes are 1 hop distant and the remaining 4 nodes are 2 hops distant. Using a single request (client) thread and a single response (server) thread, a benchmark harness explored all permutations of NUMA placement for the two threads and the two communication variables, measuring the average round-trip-time and throughput rate between the client and server. In this benchmark the server simply acts as a simple transponder, writing the request value plus 1 back into the reply field, so there's no particular computation phase and we're only measuring communication overheads. In addition to varying the placement of communication variables over pairs of nodes, we also explored variations where both variables were placed on one page (and thus on one node) -- either on the same cache line or different cache lines -- while varying the node where the variables reside along with the placement of the threads. The key observation was that if the client and server threads were on different nodes, then the best placement of variables was to have the request variable (written by the client and read by the server) reside on the same node as the client thread, and to place the response variable (written by the server and read by the client) on the same node as the server. That is, if you have a variable that's to be written by one thread and read by another, it should be homed with the writer thread. For our simple client-server model that means using split request and response communication variables with unidirectional message flow on a given page. This can yield up to twice the throughput of less favorable placement strategies. Our X4800 uses the QPI 1.0 protocol with source-based snooping. Briefly, when node A needs to probe a cache line it fires off snoop requests to all the nodes in the system. Those recipients then forward their response not to the original requester, but to the home node H of the cache line. H waits for and collects the responses, adjudicates and resolves conflicts and ensures memory-model ordering, and then sends a definitive reply back to the original requester A. If some node B needed to transfer the line to A, it will do so by cache-to-cache transfer and let H know about the disposition of the cache line. A needs to wait for the authoritative response from H. So if a thread on node A wants to write a value to be read by a thread on node B, the latency is dependent on the distances between A, B, and H. We observe the best performance when the written-to variable is co-homed with the writer A. That is, we want H and A to be the same node, as the writer doesn't need the home to respond over the QPI link, as the writer and the home reside on the very same node. With architecturally informed placement of communication variables we eliminate at least one QPI hop from the critical path. Newer Intel processors use the QPI 1.1 coherence protocol with home-based snooping. As noted above, under source-snooping a requester broadcasts snoop requests to all nodes. Those nodes send their response to the home node of the location, which provides memory ordering, reconciles conflicts, etc., and then posts a definitive reply to the requester. In home-based snooping the snoop probe goes directly to the home node and are not broadcast. The home node can consult snoop filters -- if present -- and send out requests to retrieve the line if necessary. The 3rd party owner of the line, if any, can respond either to the home or the original requester (or even to both) according to the protocol policies. There are myriad variations that have been implemented, and unfortunately vendor terminology doesn't always agree between vendors or with the academic taxonomy papers. The key is that home-snooping enables the use of a snoop filter to reduce interconnect traffic. And while home-snooping might have a longer critical path (latency) than source-based snooping, it also may require fewer messages and less overall bandwidth. It'll be interesting to reprise these experiments on a platform with home-based snooping. While collecting data I also noticed that there are placement concerns even in the seemingly trivial case when both threads and both variables reside on a single node. Internally, the cores on each X7560 package are connected by an internal ring. (Actually there are multiple contra-rotating rings). And the last-level on-chip cache (LLC) is partitioned in banks or slices, which with each slice being associated with a core on the ring topology. A hardware hash function associates each physical address with a specific home bank. Thus we face distance and topology concerns even for intra-package communications, although the latencies are not nearly the magnitude we see inter-package. I've not seen such communication distance artifacts on the T2+, where the cache banks are connected to the cores via a high-speed crossbar instead of a ring -- communication latencies seem more regular.

  Read the article

• How to create a virtual network with Azure Connect

  - by Herve Roggero

  If you are trying to establish a virtual network between machines located in disparate networks, you can either use VPN, Virtual Network or Azure Connect. If you want to establish a connection between machines located in Windows Azure, you should consider using the Virtual Network service. If you want to establish a connection between local machines and Virtual Machines in Windows Azure, you may be able to use your existing VPN device (assuming you have one), as long as the device is supported by Microsoft. If the VPN device you are using isn’t supported, or if you are trying to create a virtual network between machines from disparate networks (such as machines located in another cloud provider), you can use Azure Connect. This blog post explains how Azure Connect can help you create virtual networks between multiple servers in the cloud, various servers in different cloud environments, and on-premise. Note: Azure Connect is currently in Technical Preview. About Azure Connect Let’s do a quick review of Azure Connect. This technology implements an IPSec tunnel from machines to to a relay service located in the Microsoft cloud (Azure). So in essence, Azure Connect doesn’t provide a point-to-point connection between machines; the network communication is tunneled through the relay service. The relay service in turn offers a mechanism to enforce basic communication rules that you define through Groups. We will review this later. You could network two or more VMs in the Azure cloud (although you should consider using a Virtual Network if you go this route), or servers in the Azure cloud and other machines in the Amazon cloud for example, or even two or more on-premise servers located in different locations for which a direct network connection is not an option. You can place any number of machines in your topology. Azure Connect gives you great flexibility on how you want to build your virtual network across various environments. So Azure Connect makes sense when you want to: Connect machines located in different cloud providers Connect on-premise machines running in different locations Connect Azure VMs with on-premise (if you do not have a VPN device, or if your device is not supported) Connect Azure Roles (Worker Roles, Web Roles) with on-premise servers or in other cloud providers The diagram below shows you a high level network topology that involves machines in the Windows Azure cloud, other cloud providers and on-premise. You should note that the only required component in this diagram is the Relay itself. The other machines are optional (although your network is useful only if you have two or more machines involved). Relay agents are currently available in three geographic areas: US, Europe and Asia. You can change which region you want to use in the Windows Azure management portal. High Level Network Topology With Azure Connect Azure Connect Agent Azure Connect establishes a virtual network and creates virtual adapters on your machines; these virtual adapters communicate through the Relay using IPSec. This is achieved by installing an agent (the Azure Connect Agent) on all the machines you want in your network topology. However, you do not need to install the agent on Worker Roles and Web Roles; that’s because the agent is already installed for you. Any other machine, including Virtual Machines in Windows Azure, needs the agent installed.  To install the agent, simply go to your Windows Azure portal (http://windows.azure.com) and click on Networks on the bottom left panel. You will see a list of subscriptions under Connect. If you select a subscription, you will be able to click on the Install Local Endpoint icon on top. Clicking on this icon will begin the download and installation process for the agent. Activating Roles for Azure Connect As previously mentioned, you do not need to install the Azure Connect Agent on Worker Roles and Web Roles because it is already loaded. However, you do need to activate them if you want the roles to participate in your network topology. To do this, you will need to click on the Get Activation Token icon. The activation token must then be copied and placed in the configuration file of your roles. For more information on how to perform this step, visit MSDN at http://msdn.microsoft.com/en-us/library/windowsazure/gg432964.aspx. Firewall Rules Note that specific firewall rules must exist to allow the agent to communicate through the Relay. You will need to allow TCP 443 and ICMPv6. For additional information, please visit MSDN at http://msdn.microsoft.com/en-us/library/windowsazure/gg433061.aspx. CA Certificates You can optionally require agents to sign their activation request with the Relay using a trusted certificate issued by a Certificate Authority (CA). Click on Activation Options to learn more. Groups To create your network topology you must first create a group. A group represents a logical container of endpoints (or machines) that can communicate through the Relay. You can create multiple groups allowing you to manage network communication differently. For example you could create a DEVELOPMENT group and a PRODUCTION group. To add an endpoint you must first install an agent that will create a virtual adapter on the machine on which it is installed (as discussed in the previous section). Once you have created a group and installed the agents, the machines will appear in the Windows Azure management portal and you can start assigning machines to groups. The next figure shows you that I created a group called LocalGroup and assigned two machines (both on-premise) to that group. Groups and Computers in Azure Connect As I mentioned previously you can allow these machines to establish a network connection. To do this, you must enable the Interconnected option in the group. The following diagram shows you the definition of the group. In this topology I chose to include local machines only, but I could also add worker roles and web roles in the Azure Roles section (you must first activate your roles, as discussed previously). You could also add other Groups, allowing you to manage inter-group communication. Defining a Group in Azure Connect Testing the Connection Now that my agents have been installed on my two machines, the group defined and the Interconnected option checked, I can test the connection between my machines. The next screenshot shows you that I sent a PING request to DEVLAP02 from DEVDSK02. The PING request was successful. Note however that the time is in the hundreds of milliseconds on average. That is to be expected because the machines are connecting through the Relay located in the cloud. Going through the Relay introduces an extra hop in the communication chain, so if your systems rely on high performance, you may want to conduct some basic performance tests. Sending a PING Request Through The Relay Conclusion As you can see, creating a network topology between machines using the Azure Connect service is simple. It took me less than five minutes to create the above configuration, including the time it took to install the Azure Connect agents on the two machines. The flexibility of Azure Connect allows you to create a virtual network between disparate environments, as long as your operating systems are supported by the agent. For more information on Azure Connect, visit the MSDN website at http://msdn.microsoft.com/en-us/library/windowsazure/gg432997.aspx. About Herve Roggero Herve Roggero, Windows Azure MVP, is the founder of Blue Syntax Consulting, a company specialized in cloud computing products and services. Herve's experience includes software development, architecture, database administration and senior management with both global corporations and startup companies. Herve holds multiple certifications, including an MCDBA, MCSE, MCSD. He also holds a Master's degree in Business Administration from Indiana University. Herve is the co-author of "PRO SQL Azure" from Apress and runs the Azure Florida Association (on LinkedIn: http://www.linkedin.com/groups?gid=4177626). For more information on Blue Syntax Consulting, visit www.bluesyntax.net. Special Thanks I would like thank those that helped me figure out how Azure Connect works: Marcel Meijer - http://blogs.msmvps.com/marcelmeijer/ Michael Wood - Http://www.mvwood.com Glenn Block - http://www.codebetter.com/glennblock Yves Goeleven - http://cloudshaper.wordpress.com/ Sandrino Di Mattia - http://fabriccontroller.net/ Mike Martin - http://techmike2kx.wordpress.com

  Read the article

< Previous Page | 72 73 74 75 76 77 78 79 80 81 82  | Next Page >