Search Results

Search found 20569 results on 823 pages for 'long press'.

Page 450/823 | < Previous Page | 446 447 448 449 450 451 452 453 454 455 456 457  | Next Page >

  • The Correct Usage of DLLs with a DirectX Game?

    - by smoth190
    I'm using DirectX 10 (in C++) to make a game engine, and a test driver program on top of it. Now that I've written many messy rough drafts of an engine, I want to make the final (or sorta final) clean version. I choose to follow how I've seen other engines do it, and that's to have all the core nasty messy crap in a DLL, and then you can create games with just a few functions (well, not really :D). However, I'm unsure of what nasty messy crap to put in that DLL. I don't know about speed restrictions with DLLs. What I've done is put my winproc in the DLL, and have a class that takes the messages, and sends them through to the program using the DLL. Then that program does what it needs to do, and calls a rendering functions back in the DLL that renders everything. Only problem is it gets very low FPS (2, to be exact...). I've looked through everything, and I don't know if the way I'm using DLLs in causing this, or its something different. Whether it's the DLLs or not, I still want to know how to use a DLL correctly with a game engine. I like being neat, I hate having to see all those long names of DirectX classes. I use typedef a lot.

    Read the article

  • Internet Timeouts with TP-Link TL-WN821N v2 wireless usb stick

    - by user1622959
    A short time after accessing the internet, the browser/download times out. Before the timeout, the internet works OK briefly; afterwards, the wireless is still connected with a strong signal, but every internet access results in a timeout. When I leave the PC for a while, the internet is back just to timeout again as soon as I start using it. The same happens when I reconnect to the router. Also, when I surf the internet, it takes a couple of minutes until the timeout, but when I download something, it times out in a matter of seconds. The wireless adapter works just fine in Windows and internet via ethernet cable works just fine in Ubuntu. Does anyone have the same problem or knows a solution. I use Ubuntu 12.10 x64. The problem occurs since I installed ubuntu (which was a few days ago). Here some stuff that might be usefull: serus@serus-Ubuntu-PC:~$ lsusb Bus 002 Device 002: ID 0cf3:1002 Atheros Communications, Inc. TP-Link TL-WN821N v2 802.11n [Atheros AR9170] serus@serus-Ubuntu-PC:~$ lsmod Module Size Used by carl9170 82083 0 serus@serus-Ubuntu-PC:~$ modinfo carl9170 filename: /lib/modules/3.5.0-21- generic/kernel/drivers/net/wireless/ath/carl9170/carl9170.ko alias: arusb_lnx alias: ar9170usb firmware: carl9170-1.fw description: Atheros AR9170 802.11n USB wireless serus@serus-Ubuntu-PC:~$ iwconfig wlan0 IEEE 802.11bgn ESSID:"virginmedia0137463" Mode:Managed Frequency:2.462 GHz Access Point: A0:21:B7:F8:29:B6 Bit Rate=240 Mb/s Tx-Power=20 dBm Retry long limit:7 RTS thr:off Fragment thr:off Power Management:off Link Quality=66/70 Signal level=-44 dBm Rx invalid nwid:0 Rx invalid crypt:0 Rx invalid frag:0 Tx excessive retries:1399 Invalid misc:18 Missed beacon:0 serus@serus-Ubuntu-PC:~$ sudo lshw -C network *-network description: Wireless interface physical id: 1 bus info: usb@2:2 logical name: wlan0 serial: 00:27:19:bb:00:19 capabilities: ethernet physical wireless configuration: broadcast=yes driver=carl9170 driverversion=3.5.0-21-generic firmware=1.9.4 ip=192.168.0.6 link=yes multicast=yes wireless=IEEE 802.11bgn

    Read the article

  • Tumblr custom domain not redirecting properly

    - by Manic
    I decided to host my blog at Tumblr, using their custom domain setup (http://blog.smokingfishgames.com/ instead of http://smokingfishgames.tumblr.com). However, it's been 72 hours and I'm still getting spotty redirection. It works some of the time--I go and see the page and blog, and it's all fine. However, it occasionally just stops working and redirects back to my web host, which is a directory with nothing but a single file called BUGGER.html (which I stuck in to make sure that it was my web host and not some Tumblr empty directory). Clearing the Chrome DNS cache makes the problem go away--for a while. After a few minutes, or an hour, or however long, I'll start seeing BUGGER.html again. I clear the cache, and poof, the blog shows up. The thing that's curious to me is that when I clear the cache and get BUGGER.html again (which happens occasionally), I can look at my Chrome DNS cache and see assets.tumblr.com UNSPECIFIED blog.smokingfishgames.com UNSPECIFIED www.tumblr.com UNSPECIFIED IP addresses and expiration times omitted for brevity's sake--if they're important I'm sure I can replicate the issue. This implies, to me anyway, that my browser is reaching Tumblr but getting bounced back to my web host. Any reason why this would be happening, or is this a normal symptom of DNS propagation? If it is a problem, should I be bothering Tumblr or my host with it, or is this something I can fix myself?

    Read the article

  • Exporting a VM running as a Domain Controller

    - by AndyM
    There is a Microsoft KB article that talks about best practices for running a DC on a VM. One of the notes is "Do not use the Export feature on a virtual machine that is running a domain controller". In order to export a VM, the VM needs to be turned off. If you turn off a DC VM, export it, import it on another server and then power the VM back on, there should be no risk as long as the original (exported) VM is never powered back on. Can someone explain to me why this isn't supported? After doing some searching on Google, the sites I found just regurgitated this line from the KB article, but didn't offer any explanation as to why this isn't supported.

    Read the article

  • Standards & compliances for secure web application development?

    - by MarkusK
    I am working with developers right now that write code the way they want and when i tell them to do it other way they respond that its just matter of preference how to do it and they have their way and i have mine. I am not talking about the formatting of code, but rather of way site is organized in classes and the way the utilize them. and the way they create functions and process forms etc. Their coding does not match my standards, but again they argue that its matter of preference and as long as goal achieved the can be different way's to do it. I agree but their way is proven to have bugs and we spend a lot of time going back and forth with them to fix all problems security or functionality, yet they still write same code no matter how many times i asked them to stop doing certain things. Now i am ready to dismiss them but friend of mine told me that he has same exact problem with freelance developers he work with. So i don't want to trade one bad apple for another. Question is is there some world wide (or at least europe and usa) accepted standard or compliance on how write secure web based applications. What application architecture should be for maintainable application. Is there are some general standard that can be used for any language ruby php or java govern security and functionality and quality of code? Or at least for PHP and MySQL i use for my website. So i can make them follow this strict standard and stop making excuses.

    Read the article

  • Why do LCD screens need a backlight? (seriously)

    - by Tudor
    Bare with me on this, I'm well aware that most LCDs have backlights. But what if, they DIDN'T? I spend long hours in front of a screen and I hate the concept of staring into a lightbulb all the time and my eyes getting sore. I need to get a new lcd (some sort of IPS that doesn't cost a fortune) soon, but no matter what I'll get, I'll still be looking at a damn light. How come you can read on a kindle w/o a backlight but there aren't any monitors like that? I mean, ideally, a screen should be like a book or any other object in real life. Just for curiosity's sake, does anyone know of any new monitor concepts? Also, http://www.indiaonrent.com/view/l/laptop-transparent-monitor.html hah!

    Read the article

  • Please Help - PHP Form, when no text is entered [migrated]

    - by Joe Turner
    I'm creating a mobile landing page and I have also created a form that allows me to create more, by duplicating a folder that's host to a template file. The script then takes you to a page where you input the company details one by one and press submit. Then the page is created. My problem is, when a field is left out (YouTube for instance), the button is created and is blank. I would like there to be a default text for when there is no text. I've tried a few things and have been struggling to make this work for DAYS! <?php $company = $_POST["company"]; $phone = $_POST["phone"]; $colour = $_POST["colour"]; $email = $_POST["email"]; $website = $_POST["website"]; $video = $_POST["video"]; ?> <div id="contact-area"> <form method="post" action="generate.php"><br> <input type="text" name="company" placeholder="Company Name" /><br> <input type="text" name="slogan" placeholder="Slogan" /><br> <input class="color {required:false}" name="colour" placeholder="Company Colour"><br> <input type="text" name="phone" placeholder="Phone Number" /><br> <input type="text" name="email" placeholder="Email Address" /><br> <input type="text" name="website" placeholder="Full Website - Include http://" /><br> <input type="text" name="video" placeholder="Video URL" /><br> <input type="submit" value="Generate QuickLinks" style="background:url(images/submit.png) repeat-x; color:#FFF"/> </form> That's the form. It takes the variables and post's them to the file below. <?php $File = "includes/details.php"; $Handle = fopen($File, 'w'); ?> <?php $File = "includes/details.php"; $Handle = fopen($File, 'w'); $Data = "<div id='logo'> <h1 style='color:#$_POST[colour]'>$_POST[company]</h1> <h2>$_POST[slogan]</h2> </div> <ul data-role='listview' data-inset='true' data-theme='b'> <li style='background-color:#$_POST[colour]'><a href='tel:$_POST[phone]'>Phone Us</a></li> <li style='background-color:#$_POST[colour]'><a href='mailto:$_POST[email]'>Email Us</a></li> <li style='background-color:#$_POST[colour]'><a href='$_POST[website]'>View Full Website</a></li> <li style='background-color:#$_POST[colour]'><a href='$_POST[video]'>Watch Us</a></li> </ul> \n"; fwrite($Handle, $Data); fclose($Handle); ?> and there is what the form turns into. I need there to be a default link put in incase the field is left blank, witch it is sometimes. Thanks in advance guys.

    Read the article

  • Redundant APC UPS units, single server set up

    - by Sholom
    Hi All We have a very simple set up, looking for a very simple (reliable) solution: Setup: One Dell box with redundant power supplies running Windows 2003, plugged into two separate APC SmartUPS 1500 units (USB, no smart cards) on two separate circuits. Solution required: IF (UPS1 = Low) AND (UPS2 = Low) THEN: Shutdown gracefully ELSE: DO NOTHING!! APCUPSD only allowes for one instance (and therefore one UPS) in a windows environment. PowerChute can't do this without using APC Smart Cards which means utilizing our switch, but the switch does not have redundant power supplies, so it will only live for as long as one of the two UPS units. And no, i don't have the budget to buy two smart cards pluse a switch with redundancy ;) Thanks

    Read the article

  • A Quarter Century of SPARC

    - by kemer
    You might have missed an interesting milestone: the 25th anniversary of SPARC. Twenty-five years! Almost 40% of my life: humbling, maybe a little scary. When I joined Sun Microsystems in 1988, SPARC was just starting to shake things up. The next year we introduced the SPARCstation 1, which had basically triple the performance of our Motrolla-based Sun–3 systems. Not too long after that, our competition began a campaign of “SPARC is dead.” We really distressed them with our success, in spite of our small size. “It won’t last.” “It can’t last!” So they told themselves. For a stroll down memory lane take a look at this page. I remember the sales meeting we had in Atlanta to internally announce the SPARCstation 1. Sun hadn’t really hit the big times, yet. Our much bigger competitors viewed us as an ill-mannered pest, certain of our demise. And, why wouldn’t they be certain: other startups more our size, such as Apollo (remember them?), Silicon Graphics (they fought the good fight!), and the incredibly cool Symbolics are memories. Wait! There was also a BIG company, DEC, who scoffed at us: they are history, too. In fact, we really upset them with what was supposed to be an internal-only video production that was a take-off on Bruce Lee movies, in which we battled the evil Doctor DEC – complete with computer mice (or is that “mouses”?) wielded like nun chucks with the new SPARCstation 1 somehow in the middle of everything. The memory is vivid, but the details hazy. After all, that was almost a quarter century ago. So, here’s to Oracle’s SPARC: still going strong after all these years. – Kemer

    Read the article

  • Deactivating website in ISPConfig shows another site

    - by Mattias
    A long time ago, one of our clients setup a subdomain pointing to our ip-adress. We added a website (SitesWebsiteAdd new website) that points to one of our servers. The project is now closed and the client wants us to remove the content. When we deactivate (by unclicking active) this site it automatically defaults to another website we have in our list (!?). So, because the client is still pointing to our ip, when entering project.client.com another client project is showing up by default. How is this possible? Any suggestions? I can ofcourse give you guys more details when you tell me what details you need. Thanks

    Read the article

  • Weird system freeze. Nothing works keyboard/mouse/reset button - Ubuntu 12.04 64bits [closed]

    - by Simon
    I have fresh PC: i5 3570K with Intel 4000 onboard ASrock Z77 Extreme4-m 8GG RAM - Adata 1333Mhz... 1TB Seagate drive 7200rpm I've also fresh systems - Ubuntu/Win7 and today there was something strange. Ubuntu twice just freezed. Everything stopped, even keyboard and mouse wasn't responding. Even RESET button didn't work. Right now memtest is running, but I'd like to know, where else can i look for cause. Can it be software fault if even reset isn't working? Only long reset pressing rebooted PC... I'm a bit confused. Or should I test components - CPU, motherboard, disk. Which logs in Ubuntu should I check to diagnose cause? EHm I had few adventures with this PC already. Shipped motherboard was broken (ASrock Z68 Extreme3) and had old bios, so I had to contact with reseller, replace it and at the end decided on Z77, but everything took 3 weeks, so I have bad feelings... Edit: Both freezes were during editing something in gedit (it can be coincidence) and after few updates today - when memtest is end I'll check what was updated

    Read the article

  • Why is Ubuntu One slow to sync in 11.10, either backup or any sub-folder contents?

    - by pst007x
    I have been trying to sync my documents folder of 1.4GB, it still hasn't worked and it has been syncing for a month. The top level syncs, files and folders in the Document folders, but contents of sub-folders just hang. (Gave up and stopped syncing this folder) However,I have tried using the backup facility in 11.10, to backup to Ubuntu One.... I upgraded my HDD space in Ubuntu One. It has been going now for 24hours-ish and only backed up what looks like a couple of percent. (By the way what an excellent idea to backup to Ubuntu One, if only we could get it to actually work! :-o) The odd thing is I can sync to drop box within hours, rather than months. This is bad, and has been an issue since Ubuntu One's release. I have reported this problem and there were promises in later releases this would be fixed, but it hasn't. Canonical cannot help either... I posted on several blogs, a lot of people have the same problem but no fixes. So do I use dropbox or another service, until it is sorted, as Ubuntu does not seem to see this as an issue, I think a fix will be a long time in coming. (However,I love the potential of Ubuntu One and the integration with the OS) Yes my internet speeds are fine, etc... :-) No firewall (sudo ufw status: STATUS: INACTIVE), No Proxy, etc NB: I have raised this as a separate question to others posted here, because my question relates to Ubuntu 11.10, though I have commented elsewhere for help. Plus my question also relates to deja-dup backup to Ubuntu One. Thanks

    Read the article

  • See network computer description in Explorer

    - by John Gardeniers
    When viewing the list of network computers in Windows Explorer on a Win 7 machine there appears to be no way of readily seeing the description field, which makes that pretty well useless. I believe this problem started with Vista but I'm seeing it for the first time because I was fortunate enough to skip over Vista. We're currently rolling out Win 7 on new boxes and although I'm changing the naming system to something more sensible, the old boxes have names such as FWS01, FWS02, etc. Managing a network of computers with such names is made more difficult when you can't readily see who's machine each one is. Googling the problem brought up a workaround involving creating a shortcut on an XP machine and copying that to the Win 7 box but that's a long way from ideal. Is there some kind of hack that can be applied to machines, the operator of which needs to be able to see the description field?

    Read the article

  • telnet - is there a maximum line limit?

    - by benc
    I am working on several servers that use HTTP for transport of commands. What I have encountered is that some of the commands I am trying to issue by hand are very long GETs, several lines, and that when I telnet from my Mac to my Solaris system, I cannot seem to cut and paste the line successfully. I get a couple bounching sounds (which I assume is a control-g - bell) and then it never pastes everything. From trying to break it up into smaller pieces, I am getting the impression that TELNET, or my bundled telnet client or server has a maximum line length that I had never bumped into. I did some googling and superusering, but did not find anything definitive.

    Read the article

  • Two graphical entities, smooth blending between them (e.g. asphalt and grass)

    - by Gabriel Conrad
    Supposedly in a scenario there are, among other things, a tarmac strip and a meadow. The tarmac has an asphalt texture and its model is a triangle strip long that might bifurcate at some point into other tinier strips, and suppose that the meadow is covered with grass. What can be done to make the two graphical entities seem less cut out from a photo and just pasted one on top of the other at the edges? To better understand the problem, picture a strip of asphalt and a plane covered with grass. The grass texture should also "enter" the tarmac strip a little bit at the edges (i.e. feathering effect). My ideas involve two approaches: put two textures on the tarmac entity, but that involves a serious restriction in how the strip is modeled and its texture coordinates are mapped or try and apply a post-processing filter that mimics a bloom effect where "grass" is used instead of light. This could be a terrible failure to achieve correct results. So, is there a better or at least a more obvious way that's widely used in the game dev industry?

    Read the article

  • IIS FTP service - download timeouts and restarts getting the data twice

    - by accel229
    We have an IIS FTP site on a Windows Server 2003 x64 machine. Application Layer Gateway service is disabled (so http://support.microsoft.com/kb/931130 does not apply). Windows Firewall service is disabled as well. Connection timeout for the FTP site (there is only one) is set to 1,200 seconds = 20 minutes. An external client can connect to the site, list directory contents and download small files. When a client attempts to download a large file (eg, if the download continues for 3 minutes, which is still under 20 minutes, but relatively long), the server sends all data, then the connection times out, the client issues REST / RETR commands attempting to restart the download since after the last byte (which I believe should succeed and receive exactly 0 bytes), and the server behaves as if the client tried to restart after byte 0, that is, it sends the entire file all over. Any ideas on how to fix this?

    Read the article

  • Web application design with distributed servers

    - by Bonn
    I want to build a web application/server with this structure: main-server sub-server transaction-server (create, update, delete) view-server (view, search) authentication-server documents-server reporting-server library-server e-learning-server The main-server acts as host server for sub-server. I can add many sub-servers and connect it to main-server (via plug-play interface maybe), then it can begin querying data from another sub-servers (which has been connected to the main-server). The sub-servers can be anywhere as long as connected to internet. The main-server can manage all sub-servers which are connected to it (query data, setting permission between sub-servers, etc). The purpose is simple, the web application will be huge as the company grows, so I want to distribute it into small connected plug-able servers. My question is, does the structure above already have a standardized method? or are there any different views? what are the technologies needed? I need a lot of researches before the execution plan begin. thanks a lot.

    Read the article

  • Blocking non-virtual host access in Apache?

    - by cmbrnt
    I'm running an apache-server, with a bunch of virtual hosts for about seven domain names. Now I'd like to disallow access for clients who try to access my server using only its IP-adress. So: When someone accesses my website through www.domain.com, they reach the site hosted in /var/www/domain.com/public_html/. When someone enters the ip-address of the server they reach a 403 Forbidden-message. The problem with this is that they are theoretically able to reach my other sites through bruteforce, when getting http://11.22.33.44/domain.com/public_html/. I rather want them to reach a 403 Forbidden all the time, as long as they don't access my server by a valid domain name. How do I solve this problem?

    Read the article

  • Do I need to retain Sharepoint usage analysis log files

    - by dunxd
    Our Sharepoint installation currently has 30Gb of Usage Analysis Log file - these date back about six months. I have configured Sharepoint to do Usage Analysis Processing every night, so I am wondering whether I need to keep these files for so long. Sharepoint doesn't seem to clean up these files automatically - I think six months ago I had to clear out logs due to disk space issues. So my question is, do I need to retain these files in order to get decent usage analysis reports, or can I delete them as soon as the usage analysis processing has completed?

    Read the article

  • Send less Server Data with "AFK"

    - by Oliver Schöning
    I am working on a 2D (Realtime) MultiPlayer Game. With Construct2 and a Socket.IO JavaScript Server. Right now the code does not include the Array for each Player. var io = require("socket.io").listen(80); var x = 10; io.sockets.on("connection", function (socket) { socket.on("message", function(data) { x = x+1; }); }); setInterval(function() { io.sockets.emit("message", 'Pos,' + x); },100); I noticed a very annoying problem with my server today. It sends my X Coordinates every 100 milliseconds. The Problem was, that when I went into another Browser Tab, the Browser stopped the Game from running. And when I went back, I think the Game had to run through all the packages. Because my Offline Debugging Button still worked immediately and the Online Button only responded after some seconds. So then I changed my Code so that it would only send out an update when it received a player Input: var io = require("socket.io").listen(80); var x = 10; io.sockets.on("connection", function (socket) { socket.on("message", function(data) { x = x+1; io.sockets.emit("message", 'Pos,' + x); }); }); And it Updated Immediately, even when I had been inactive on the Browser Tab for a long time. Confirming my suspicion that it had to get through all the data. Confirm Please! It would be insane to only send information on Client Input in a Real Time Game. But how would I write a AFK function? I would think it is easier to run a AFK Boolean Loop on the Server. Here is what I need help for: playerArray[Me] if ( "Not Given any Input for X amount of Seconds" ) { "Don't send Data" } else { "Send Data" }

    Read the article

  • Platform game collisions with Block

    - by Sri Harsha Chilakapati
    I am trying to create a platform game and doing wrong collision detection with the blocks. Here's my code // Variables GTimer jump = new GTimer(1000); boolean onground = true; // The update method public void update(long elapsedTime){ MapView.follow(this); // Add the gravity if (!onground && !jump.active){ setVelocityY(4); } // Jumping if (isPressed(VK_SPACE) && onground){ jump.start(); setVelocityY(-4); onground = false; } if (jump.action(elapsedTime)){ // jump expired jump.stop(); } // Horizontal movement setVelocityX(0); if (isPressed(VK_LEFT)){ setVelocityX(-4); } if (isPressed(VK_RIGHT)){ setVelocityX(4); } } // The collision method public void collision(GObject other){ if (other instanceof Block){ // Determine the horizontal distance between centers float h_dist = Math.abs((other.getX() + other.getWidth()/2) - (getX() + getWidth()/2)); // Now the vertical distance float v_dist = Math.abs((other.getY() + other.getHeight()/2) - (getY() + getHeight()/2)); // If h_dist > v_dist horizontal collision else vertical collision if (h_dist > v_dist){ // Are we moving right? if (getX()<other.getX()){ setX(other.getX()-getWidth()); } // Are we moving left? else if (getX()>other.getX()){ setX(other.getX()+other.getWidth()); } } else { // Are we moving up? if (jump.active){ jump.stop(); } // We are moving down else { setY(other.getY()-getHeight()); setVelocityY(0); onground = true; } } } } The problem is that the object jumps well but does not fall when moved out of platform. Here's an image describing the problem. I know I'm not checking underneath the object but I don't know how. The map is a list of objects and should I have to iterate over all the objects??? Thanks

    Read the article

  • Ubuntu: Alternative task-switcher to what is provided in compiz

    - by user22106
    I'm running Ubuntu 11.04 and really enjoying it. The task switcher default in Ubuntu shows a list of small icons for each running app much like the windows task switcher. It's fast and effective. I enabled Compiz and noticed that the task switcher changed to one with small thumbnails showing a smaller version of each window. Nice, but it slows the task manager down to about 0.5 seconds to display, which for me is far too long. Have googled the forums and it's a very common problem with ubuntu/compiz. I want to run compiz as it provides so many other features, but the slow task switcher is just not going to run with me. So was wondering if there is some way of NOT letting compiz override the task switcher, or, if there is some other task switcher that I can install which will not be slow like the compiz one. Any help would be appreciated.

    Read the article

  • Single hardware unit to protect web servers and implement smart publishing

    - by Maxim V. Pavlov
    Thus far we've been using the combination of Forefront TMG 2010 as an edge firewall + intrusion prevention system + web site publishing mechanism in the data center to work with a few web server machines. Since we develop on ASP.NET, we are IIS and in general - Microsoft crowd. Since TMG is being deprecated, we need to come up with a hardware alternative to protect and serve our data center web cloud. Could you please advise a hardware or virtual appliance solution that can provide routing, flood prevention and smart web-site publishing (one IP - many web sites based on domain name filter) all in one. Even if it is hard to configure, as long as it covers all these features, we will invest to learn and replace TMG eventually.

    Read the article

  • Slow http traffic between VMWare guest and host.

    - by toluju
    I have a web application running as an http server inside the VMWare guest OS, and I'm trying to access the content from the host OS. The guest is running Ubuntu, and the host is running Windows XP. The problem is, when I try to access the application from a browser in the host OS, the content takes a very long time to load (up to a minute for a single page). A browser in the guest OS can access the application with no problems. I've tried using both NAT and bridged networking, but the results are the same. The Windows firewall is turned off. The connection itself appears fine, as ping requests from guest to host as well as host to guest complete without errors or delays. Both guest and host can access the external Internet connection without a problem. I'm using VMWare Player. Any ideas?

    Read the article

  • How can a single disk in a hardware SATA RAID-10 array bring the entire array to a screeching halt?

    - by Stu Thompson
    Prelude: I'm a code-monkey that's increasingly taken on SysAdmin duties for my small company. My code is our product, and increasingly we provide the same app as SaaS. About 18 months ago I moved our servers from a premium hosting centric vendor to a barebones rack pusher in a tier IV data center. (Literally across the street.) This ment doing much more ourselves--things like networking, storage and monitoring. As part the big move, to replace our leased direct attached storage from the hosting company, I built a 9TB two-node NAS based on SuperMicro chassises, 3ware RAID cards, Ubuntu 10.04, two dozen SATA disks, DRBD and . It's all lovingly documented in three blog posts: Building up & testing a new 9TB SATA RAID10 NFSv4 NAS: Part I, Part II and Part III. We also setup a Cacit monitoring system. Recently we've been adding more and more data points, like SMART values. I could not have done all this without the awesome boffins at ServerFault. It's been a fun and educational experience. My boss is happy (we saved bucket loads of $$$), our customers are happy (storage costs are down), I'm happy (fun, fun, fun). Until yesterday. Outage & Recovery: Some time after lunch we started getting reports of sluggish performance from our application, an on-demand streaming media CMS. About the same time our Cacti monitoring system sent a blizzard of emails. One of the more telling alerts was a graph of iostat await. Performance became so degraded that Pingdom began sending "server down" notifications. The overall load was moderate, there was not traffic spike. After logging onto the application servers, NFS clients of the NAS, I confirmed that just about everything was experiencing highly intermittent and insanely long IO wait times. And once I hopped onto the primary NAS node itself, the same delays were evident when trying to navigate the problem array's file system. Time to fail over, that went well. Within 20 minuts everything was confirmed to be back up and running perfectly. Post-Mortem: After any and all system failures I perform a post-mortem to determine the cause of the failure. First thing I did was ssh back into the box and start reviewing logs. It was offline, completely. Time for a trip to the data center. Hardware reset, backup an and running. In /var/syslog I found this scary looking entry: Nov 15 06:49:44 umbilo smartd[2827]: Device: /dev/twa0 [3ware_disk_00], 6 Currently unreadable (pending) sectors Nov 15 06:49:44 umbilo smartd[2827]: Device: /dev/twa0 [3ware_disk_07], SMART Prefailure Attribute: 1 Raw_Read_Error_Rate changed from 171 to 170 Nov 15 06:49:45 umbilo smartd[2827]: Device: /dev/twa0 [3ware_disk_10], 16 Currently unreadable (pending) sectors Nov 15 06:49:45 umbilo smartd[2827]: Device: /dev/twa0 [3ware_disk_10], 4 Offline uncorrectable sectors Nov 15 06:49:45 umbilo smartd[2827]: Num Test_Description Status Remaining LifeTime(hours) LBA_of_first_error Nov 15 06:49:45 umbilo smartd[2827]: # 1 Short offline Completed: read failure 90% 6576 3421766910 Nov 15 06:49:45 umbilo smartd[2827]: # 2 Short offline Completed: read failure 90% 6087 3421766910 Nov 15 06:49:45 umbilo smartd[2827]: # 3 Short offline Completed: read failure 10% 5901 656821791 Nov 15 06:49:45 umbilo smartd[2827]: # 4 Short offline Completed: read failure 90% 5818 651637856 Nov 15 06:49:45 umbilo smartd[2827]: So I went to check the Cacti graphs for the disks in the array. Here we see that, yes, disk 7 is slipping away just like syslog says it is. But we also see that disk 8's SMART Read Erros are fluctuating. There are no messages about disk 8 in syslog. More interesting is that the fluctuating values for disk 8 directly correlate to the high IO wait times! My interpretation is that: Disk 8 is experiencing an odd hardware fault that results in intermittent long operation times. Somehow this fault condition on the disk is locking up the entire array Maybe there is a more accurate or correct description, but the net result has been that the one disk is impacting the performance of the whole array. The Question(s) How can a single disk in a hardware SATA RAID-10 array bring the entire array to a screeching halt? Am I being naïve to think that the RAID card should have dealt with this? How can I prevent a single misbehaving disk from impacting the entire array? Am I missing something?

    Read the article

< Previous Page | 446 447 448 449 450 451 452 453 454 455 456 457  | Next Page >