Search Results

Search found 10429 results on 418 pages for 'high resolution'.

Page 40/418 | < Previous Page | 36 37 38 39 40 41 42 43 44 45 46 47  | Next Page >

  • High availability virtual machines

    - by Jeremy
    I've been reading a lot about high availability virtualization, either via Hyper-V or VMWare. In that context, essentially high availabliity means that the VM is hosted by a closter of physical servers (nodes), so if one of the physical servers goes down, the VM can still be served by other physical servers. So far so good, the physical cluster and the VM itself are highly available. However if the service being provided, let's say SQL server, MSDTC, or any other service, are actually being provided by the VM image and the virtualized operating system. So I imagine that there is still a point of failure at the virtual layer that isn't accounted for. Something could happen within the virtual machine itself that the physican cluster can not account for, correct? In that instance the physican failover cluster (Hyper-V) or VMWare host, can not fail over, because the issue is not with one of the servers in the physical cluster - failing over a physical node would not do any good. Does this necessitate building a virtual failover cluster on top of the physical one, or is this not necessary? Alternatively, I suppose you could skip the phsyical clustering, and just cluster at the virtual layer (Child based failover clustering), because that should still survive a physical failure. See image below showing parent based (left), child based (right) and a combination (center). Is parent based as far as you need to go, or is child based more appropriate?

    Read the article

  • What cause high CPU usage on the server during file upload

    - by bosiang
    When I try to upload a huge file size (approx 2GB), the server cpu usage goes really high. What should I do to fix this? I just use standard html form and php, for file upload. I'm sorry if I post on the wrong forum. Please point me to the right direction here is the result of "top" command during uploading 4 files (18mb, 38mb, 60mb, 33mb) 1904 apache 20 0 33504 5740 1952 R 28.3 0.2 0:02.19 httpd 1905 apache 20 0 33504 5740 1952 R 28.3 0.2 0:01.99 httpd 1903 apache 20 0 33232 6968 3060 R 28.0 0.2 0:01.98 httpd 1910 apache 20 0 33240 6020 2248 S 11.5 0.2 0:02.85 httpd 2133 root 20 0 2656 1124 896 R 1.6 0.0 0:00.71 top 1 root 20 0 2864 1404 1188 S 0.0 0.0 0:03.99 init the code for chunking, although eventhough I don't use this code (just simple file upload), it still cause that high cpu usage function sendRequest() { //clean the screen //bars.innerHTML = ''; var file = document.getElementById('fileToUpload'); for(var i = 0; i < file.files.length; i++) { var blob = file.files[i]; var originalFileName = blob.name; var filePart = 0 const BYTES_PER_CHUNK = 100 * 1024 * 1024; // 10MB chunk sizes. var realFileSize = blob.size; var start = 0; var end = BYTES_PER_CHUNK; totalChunks = Math.ceil(realFileSize / BYTES_PER_CHUNK); alert(realFileSize); while( start < realFileSize ) { if (blob.webkitSlice) { //for Google Chrome var chunk = blob.webkitSlice(start, end); } else if (blob.mozSlice) { //for Mozilla Firefox var chunk = blob.mozSlice(start, end); } uploadFile(chunk, originalFileName, filePart, totalChunks, i); filePart++; start = end; end = start + BYTES_PER_CHUNK; } } }

    Read the article

  • Need help in support multiple resolution screen on android

    - by michael
    Hi, In my android application, I would like to support multiple screens. So I have my layout xml files in res/layout (the layout are the same across different screen resolution). And I place my high-resolution asserts in res/drawable-hdpi In my layout xml, I have <?xml version="1.0" encoding="utf-8"?> <TableLayout xmlns:android="http://schemas.android.com/apk/res/android" android:id="@+id/table" android:background="@drawable/bkg"> And I have put bkg.png in res/drawable-hdpi And I have started my emulator with WVGA-800 as avd. But my application crashes: E/AndroidRuntime( 347): Caused by: android.content.res.Resources$NotFoundException: Resource is not a Drawable (color or path): TypedValue{t=0x1/d=0x7f020023 a=-1 r=0x7f020023} E/AndroidRuntime( 347): at android.content.res.Resources.loadDrawable(Resources.java:1677) E/AndroidRuntime( 347): at android.content.res.TypedArray.getDrawable(TypedArray.java:548) E/AndroidRuntime( 347): at android.view.View.<init>(View.java:1850) E/AndroidRuntime( 347): at android.view.View.<init>(View.java:1799) E/AndroidRuntime( 347): at android.view.ViewGroup.<init>(ViewGroup.java:284) E/AndroidRuntime( 347): at android.widget.LinearLayout.<init>(LinearLayout.java:92) E/AndroidRuntime( 347): ... 42 more Does anyone know how to fix my problem? Thank you.

    Read the article

  • Exchange 2010 DAG + VMWare HA = no support?

    - by Dan
    We currently have an Exchange 2003 clustered environment (two machine cluster) that we're looking to upgrade to 2010. We recently purchased a VMWare virtualization environment (three Dell R710's with an EMC NS-120 serving up NFS datastores - iSCSI is available) that we wish to use for this new environment. I'm seeing that Microsoft does not support Exchange 2010 DAGs with a virtualization high availability solution (see links below). I would like to utilize the DAG to ensure the data stays available if one host goes down, and HA to ensure that if the physical host goes down, the VM will come back up on the other available host. Does anybody know why MS does not support this? VMWare HA will only restart the VM if it is hung/down - I don't see any difference between this and restarting the physical box if someone pulled the power... Will we only run into issues with support if it has something to do with HA/DAG failover or will they see we have HA and tell us to put it on a physical box even if it has nothing to do with HA? If we disable HA for these VM's will that satisfy them on a support case? Has anybody set up an Exchange 2010 DAG on VMware with HA enabled? Will they have any issues with using an NFS datastore? We have much greater flexibility on the EMC with NFS vs iSCSI, so I would prefer to continue utilizing that. Thanks for any input! http://www.vmwareinfo.com/2010/01/verifying-microsoft-exchange-2010.html Take a look at the second image under "Not Supported" http://technet.microsoft.com/en-us/library/aa996719.aspx "Microsoft doesn't support combining Exchange high availability solutions (database availability groups (DAGs)) with hypervisor-based clustering, high availability, or migration solutions. DAGs are supported in hardware virtualization environments provided that the virtualization environment doesn't employ clustered root servers."

    Read the article

  • Oracle Application Server Performance Monitoring and Tuning (CPU load high)

    - by Berkay
    Oracle Application Server Performance Monitoring and Tuning (CPU load high) i have just hired by a company and my boss give me a performance issue to solve as soon as possible. I don't have any experience with the Java EE before at the server side. Let me begin what i learned about the system and still couldn't find the solution: We have an Oracle Application Server (10.1.) and Oracle Database server (9.2.), the software guys wrote a kind of big J2EE project (X project) using specifically JSF 1.2 with Ajax which is only used in this project. They actively use PL/SQL in their code. So, we started the application server (Solaris machine), everything seems OK. users start using the app starting Monday from different locations (app 200 have user accounts,i just checked and see that the connection pool is set right, the session are active only 15 minutes). After sometime (2 days) CPU utilization gets high,%60, at night it is still same nothing changed (the online user amount is nearly 1 or 2 at this time), even it starts using the CPU allocated for other applications on the same server because they freed If we don't restart the server, the utilization becomes %90 following 2 days, application is so slow that end users starts calling. The main problem is software engineers say that code is clear, and the System and DBA managers say that we have the correct configuration,the other applications seems OK why this problem happens only for X application. I start copying the DB to a test platform and upgrade it to the latest version, also did in same with the application server (Weblogic) if there is a bug or not. i only tested by myself only one user and weblogic admin panel i can track the threads and dump them. i noticed that there are some threads showing as a hogging. when i checked the manuals and control the trace i see that it directs me the line number where PL/SQL code is called from a .java file. The software eng. says that yes we have really complex PL/SQL codes but what's the relation with Application server? this is the problem of DB server, i guess they're right... I know the question has many holes, i'd like to give more in detail but i appreciate the way you guide me. Thanks in advance ... Edit: The server both in CPU and Memory enough to run more complex applications

    Read the article

  • iptables DNS resolution

    - by Favolas
    I have a virtual machine with Fedora 19 acting as a router. This machine as an interface (p8p1) with the IP 172.16.1.254 that is connected to another machine (IP 172.16.1.1) that's simulating the external network. I've installed snort 2.9.2.2, applied the snortsam-2.9.2.2.diff.gz patch and installed snortsam 2.70 on the routermachine In snort.conf besides altering some RULE_PATH I believe I've only added the following line to the file. output alert_fwsam: 127.0.0.1:898/password After doing this two comands: ifconfig p8p1 promisc /usr/local/snort/bin/snort -v -i p8p1 If I ping from the external network to the router IP, I can see the info about the pings. One of the rules that I have is icmp-info.rules that as this single line: alert icmp $EXTERNAL_NET any -> $HOME_NET any (msg:"ICMP-INFO Echo Reply"; icode:0; itype:0; classtype:misc-activity; sid:408; rev:6;fwsam: src, 5 minutes;) snortsam.conf as this data: defaultkey password accept localhost keyinterval 30 minutes dontblock 192.168.1.1 # rede local rollbackhosts 50 rollbackthreshold 20 / 30 secs rollbacksleeptime 1 minute logfile /var/log/snort/snortsam.log loglevel 3 daemon nothreads # linha importante para gerar os bloqueios via iptables iptables p8p1 LOG bindip 127.0.0.1 Now I run this command: /usr/local/snort/bin/snort -u snort -i p8p1 -c /etc/snort/snort.conf -l /var/log/snort -Dq Terminal gives this message: Spawning daemon child... My daemon child 2080 lives... Daemon parent exiting (0) and when I runsnortsam in terminal i got this: SnortSam, v 2.70. Copyright (c) 2001-2009 Frank Knobbe . All rights reserved. Plugin 'fwsam': v 2.5, by Frank Knobbe Plugin 'fwexec': v 2.7, by Frank Knobbe Plugin 'pix': v 2.9, by Frank Knobbe Plugin 'ciscoacl': v 2.12, by Ali Basel <[email protected]> Plugin 'cisconullroute': v 2.5, by Frank Knobbe Plugin 'cisconullroute2': v 2.2, by Wouter de Jong <[email protected]> Plugin 'netscreen': v 2.10, by Frank Knobbe Plugin 'ipchains': v 2.8, by Hector A. Paterno <[email protected]> Plugin 'iptables': v 2.9, by Fabrizio Tivano <[email protected]>, Luis Marichal <[email protected]> Plugin 'ebtables': v 2.4, by Bruno Scatolin <[email protected]> Plugin 'watchguard': v 2.7, by Thomas Maier <[email protected]> Plugin 'email': v 2.12, by Frank Knobbe Plugin 'email-blocks-only': v 2.12, by Frank Knobbe Plugin 'snmpinterfacedown': v 2.3, by Ali BASEL <[email protected]> Plugin 'forward': v 2.8, by Frank Knobbe Parsing config file /etc/snortsam.conf... Linking plugin 'iptables'... Checking for existing state file "/var/db/snortsam.state". Found. Reading state file. Starting to listen for Snort alerts. and snortsam.log as an entry like this 2013/10/25, 10:15:17, -, 1, snortsam, Starting to listen for Snort alerts. Now, from the external machine I do ping 172.16.1.254 and it starts showing the info and an alert file is created in /var/log/snort/ that as the info about the PINGS. Something like: [**] [1:408:6] ICMP-INFO Echo Reply [**] [Classification: Misc activity] [Priority: 3] 10/25-10:35:16.061319 172.16.1.254 -> 172.16.1.1 ICMP TTL:64 TOS:0x0 ID:38720 IpLen:20 DgmLen:84 Type:0 Code:0 ID:1389 Seq:1 ECHO REPLY Also, if I run instead /usr/local/snort/bin/snort snort -v -i p8p1 i got this message: Running in packet dump mode --== Initializing Snort ==-- Initializing Output Plugins! Snort BPF option: snort pcap DAQ configured to passive. The DAQ version does not support reload. Acquiring network traffic from "p8p1". ERROR: Can't set DAQ BPF filter to 'snort' (pcap_daq_set_filter: pcap_compile: syntax error)! Fatal Error, Quitting.. So, this are my questions: Shouldn't snortsam block the PING? Is that DAQ error causing the problem? If so, How can I solve it?

    Read the article

  • Can't set 1280x1024 on Ubuntu with Nvidia Geforce 8400 GS

    - by torbengb
    Ubuntu doesn't offer 1280x1024 but that's what my monitor has -- I want to set that resolution but need help doing so. Ubuntu offers 1360x768, 1152x864, 1024x768 when using the Nvidia driver version 180. This is Ubuntu 9.04 installed as "Wubi" through Windows, I believe it's the amd64 version. Google gave me this precise explanation which is complete latin to me; I'm a Windows guy and not familiar with Terminal. These SU questions also don't provide answers, hence this new question. I hope you can help, and I will provide additional info if required.

    Read the article

  • Investigate high load on RHEL

    - by Adam Matan
    One of my RHEL 5 was showing high load (~4-5) on uptime. The load increased, and when it reached 6 (±), the server froze and needed restart. top-wise, the server had no significant CPU or memory issues and sar showed no increase in iowait. Therefore, the thrashing must have been related to other factors. Any ideas how to investigate this? In particular, how do I know that which processes are waiting in the queue?

    Read the article

  • Exclude IPs from FQDN Resolution

    - by Jon Rauschenberger
    Is there a way to EXCLUDE IPs on an interface from FQDN resulution? Here's my use case; I have a domain joined Windows Server 2008 R2 machine that hosts multiple Web sites. We ahve DNS A records for the various sites that resolve to different IP addresses. I've added those IPs to the single NIC in the machine and the sites work just fine. The problem is that the FQDN of the machine now resolves any of the IPs on the machine. I only want the primary IP associated with the FQDN on the internal DNS servers. Is there ANY way to accomplish this? jon

    Read the article

  • DNS Resolution doesn't work after uninstalling Cisco VPN & Deterministic Network Enhancer in Win 7

    - by Craig M
    I just upgraded my home PC to Windows 7 Ultimate 32 bit. After trying various methods to get the Cisco VPN client to work, I gave up and decided to just run it in XP mode. The last steps I tried were in this article ( http://social.technet.microsoft.com/Forums/en-US/w7itproappcompat/thread/d880dfe5-7f44-4955-8620-2a9355d8ea8b/ ) After that, I uninstalled the Cisco client and rebooted. I uninstalled the Deterministic Network Enhancer and rebooted again. Both uninstalled successfully, but now I'm not able to resolve any DNS. The only way I can resolve DNS is to reinstall the DNE, reboot, and uninstall the DNE. Then I am able to resolve DNS lookups until I reboot again. Once it's rebooted, no more DNS. Any ideas? Edit: I completely forgot I'd asked this question until harrymc posted his answer. I've since found out that to fix this problem, I need to disable my Local Area Connection and re-enable it. Once I do that I have no trouble making network connections until the next time I reboot at which point I repeat the process. It's annoying, but manageable since I reboot very infrequently.

    Read the article

  • DNS resolution over DHCP

    - by Eric
    I have a m0n0wall router a VMWare workstation running ubuntu a windows 7 workstation running the VM The ubuntu hostname is "renraku" From the windows machine I can't resolve dns automatically for this host. For example, when I ping renraku Ping request could not find host renraku. Please check the name and try again. However nslookup seems to work nslookup renraku Server: m0n0wall.local Address: 192.168.123.254 Name: renraku Address: 192.168.123.248 I don't get how to have ping to work with hostnames. The main goal behind this is to have my web server to work with hostnames instead of ip addresses EDIT : ping 192.168.123.248 works

    Read the article

  • Amazon EC2 High-Memory Extra Large Instance

    - by Simpanoz
    I am new to Mongodb and EC2. If I use following single MongoDb server : High-Memory Extra Large Instance 17.1 GiB memory, 6.5 ECU (2 virtual cores with 3.25 EC2 Compute Units each), 420 GB of local instance storage, 64-bit platform As a layman, if we quantify I/O, data in MB/sec. How much I/O transactions mongodb server can handle easily, without being burnt out. Consider default settings of EC2 server with Ubuntu and MongoDb version available in AWS marketplace. Thanks in advance

    Read the article

  • DNS resolution Windows 7 & browsing to locally hosted web site

    - by Aidan Whitehall
    We host two Intranet sites, http://intranet/ and http://sales.intranet/, both on the same server on the LAN. Local DNS (a Windows 2003 Server) was updated and both hostnames are configured to be CNAMEs that point to the FQDN name of the server on which they're hosted. On the LAN, Windows XP Professional clients can browse to both sites. However, Windows 7 Professional clients can browse to the main Intranet site, but not the Sales Intranet (neither using Firefox 3 nor Internet Explorer 8). Using nslookup on the command line on the Windows 7 boxes, intranet and sales.intranet both correctly resolve as CNAMEs of the server hosting them, and that in turn correctly resolves to the host's IP address. So the Q is... can anyone think why this might be, or what test to try next? Thank you for any suggestions!

    Read the article

  • Apache server configuration name resolution (virtual host naming + security)

    - by Homunculus Reticulli
    I have just setup a minimal (hopefully secure? - comments welcome) apache website using the following configuration file: <VirtualHost *:80> ServerName foobar.com ServerAlias www.foobar.com ServerAdmin [email protected] DocumentRoot /path/to/websites/foobar/web DirectoryIndex index.php # CustomLog with format nickname LogFormat "%h %l %u %t \"%r\" %>s %b" common CustomLog "|/usr/bin/cronolog /var/log/apache2/%Y%m.foobar.access.log" common LogLevel notice ErrorLog "|/usr/bin/cronolog /var/log/apache2/%Y%m.foobar.errors.log" <Directory /> AllowOverride None Order Deny,Allow Deny from all </Directory> <Directory /path/to/websites/> Options -Indexes FollowSymLinks MultiViews AllowOverride None Order allow,deny allow from all </Directory> </VirtualHost> I am able to access the website by using www.foobar.com, however when I type foobar.com, I get the error 'Server not found' - why is this? My second question concerns the security implications of the directive: <Directory /path/to/websites/> Options -Indexes FollowSymLinks MultiViews AllowOverride None Order allow,deny allow from all </Directory> in the configuration above. What exactly is it doing, and is it necessary?. From my (admitedly limited) understanding of Apache configuration files, this means that anyone will be able to access (write to?) the /path/to/websites/ folder. Is my understanding correct? - and if yes, how is this not a security risk?

    Read the article

  • Low CPU usage but High fan speed

    - by dhasu
    The problem I am facing is the CPU fan rotates with high speed even though the CPU usage is very low, less than 10%. I even cleaned the Dust off the Fan, I replaced the Thermal Material on the Processor. Installed an Extra Fan. Nothing seems to Work. Please Help.

    Read the article

  • Mixed network, Linux-to-Linux hostname resolution issues

    - by James
    At work we have an WinSBS domain at the heart of our network, which is all Windows PCs. The domain controller is acting as a DNS for these computers. I have recently added some personal use Linux machines to the network, without joining them to the domain. I have set up Samba with "wins server" pointing to the domain controller, which lets the Windows boxes resolve the Linux hostnames just fine. I also have resolvconf set up with the domain controller as a nameserver and the local domain as a searched domain, which lets the Linux boxes resolve the Windows hostnames just fine. However, the Linux boxes will not resolve other Linux hostnames at all. Given that I don't have control over the DNS server (I am not the network admin) and that at least one of the Linux boxes is not an always-on machine and is likely to change its LAN IP frequently (via DHCP), what service am I missing to make their hostnames visible to each other?

    Read the article

  • Sony Bravia KDL-32L5000 PC resolution slightly off.

    - by user18818
    Hi, I have a PC running two Nvidia 8500 GTs in SLI mode and I am trying to use my TV in dual mode. When I switch the TV to PC the screen is nearly centered with a slight offset. All resolutions are effected from 800x600 all the way up to the TVs native 1360x768. I have tried with SLI on and off and have PhysX turned off as well as I thought that might have an effect. I am running Windows XP 64-bit SP2 DirectX ver 9.0c Nvidia driver version 181.22 Any other information please let me know. Thanks in advance.

    Read the article

  • Best way to upscale a video?

    - by Josh
    If I have a video file at 320x240 resolution which I want to re-encode (because I don't like the encoding it's in now) and I also want to play it at double size (640x480), will I get higher quality if I scale it up to 640x480 when I convert it to a new format, verses keeping it at 320x240 in the new format and playing it at double size? This probably depends on the program used to convert, and if so, please let me know any program which might increase the quality. Here's my thinking. If I play a 320x240 file at double size, the system has to scale up each frame in real time, whereas if I scale up while recompressing the system may be able to use a more intensive algorythm like Bicubic interpolation . However I am not sure if this is true or not.

    Read the article

  • Why does text look so Horrible in my HD monitor???

    - by Laura
    I just bought a 1080p 22" Samsung HD monitor (connected via HDMI) and the picture and video quality is great but the text quality is absolutely horrible. Even as I type now all the text in this text box as well as in the browser toolbar and start menu, etc looks weird - like it all has a white outline around it that makes it jagged and hard to read. It hurts my eyes just to look at it. I am running my PC in the suggested native resolution of 1920 by 1080, so what's the problem? Is this one of the unavoidable downsides of using a HD monitor? Is there a solution to the problem?

    Read the article

  • Erratic DNS name resolution

    - by alex
    Hi all, We have a client we host a web for (blog.foobar.es). We do not manage foobar.es's DNS setup, we just told them to point blog.foobar.es to our web server's IP. We have noticed that sometimes we cannot browse to blog.foobar.es, but we can browse to other sites on that server. Troubleshooting a bit using host(1) yields something funny: $ host blog.foobar.es 8.8.8.8 Using domain server: Name: 8.8.8.8 Address: 8.8.8.8#53 Aliases: Host blog.foobar.es not found: 3(NXDOMAIN) , being 8.8.8.8 one of Google's public DNS servers. However, sometimes the same server resolves the name correctly (!). Another funny thing, is that our ISP's DNS servers sometimes say: $ host blog.foobar.es 80.58.61.250 Using domain server: Name: 80.58.61.250 Address: 80.58.61.250#53 Aliases: blog.foobar.es has address x.x.x.x Host blog.foobar.es not found: 3(NXDOMAIN) Which I don't really understand. I've dug around using dig(1), and have noticed they've set up a SOA record for foobar.es: $ dig foobar.es ; <<>> DiG 9.7.0-P1 <<>> foobar.es ;; global options: +cmd ;; Got answer: ;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 59824 ;; flags: qr rd ra; QUERY: 1, ANSWER: 0, AUTHORITY: 1, ADDITIONAL: 0 ;; QUESTION SECTION: ;foobar.es. IN A ;; AUTHORITY SECTION: foobar.es. 86400 IN SOA dns1.provider.es. root.dns1.provider.es. 2011030301 86400 7200 2592000 172800 ;; Query time: 78 msec ;; SERVER: 80.58.61.250#53(80.58.61.250) ;; WHEN: Thu Mar 3 16:16:19 2011 ;; MSG SIZE rcvd: 78 ... which I'm completely unfamiliar with. Ideas? We can't really do much as we do not control DNS, but we'd like to point our clients in the right direction...

    Read the article

  • Strange scaling when duplicating monitors with another screen

    - by Aerione
    I can't get two monitors to scale application resolutions the same way. My main monitor works normally. My second monitor however, which is set to duplicate its image onto a TV I have in my room, renders the applications in a far lower resolution than the 1080p I've set it to. Also, the mouse pointer on the second monitor is enormous, it looks 2-3 times bigger than the one on the main monitor. I've checked the "Let me choose one scaling level for all my displays", to no avail. Here are some comparison pictures. Metro on the main monitor: Metro on monitor 2 (set to 1080p and to duplicate on a TV): This issue only seem to arise when I duplicate the monitor onto the TV. Does anyone have any idea of how to solve this?

    Read the article

  • Is there a way to get a larger desktop than the screen?

    - by Cajuntechie
    Is there any freeware out there that will allow me to run the desktop that is larger than the resolution of the screen? My situation: Win XP, Dell Mini 10, iirc- Intel GMA950 I need to use a netbook for one specific web site (with firefox) that doesn't render properly if the screen isn't at least 1024x768. It is a netbook, so the res is 1024x600. Is there anything out there that will allow me to use a 1024x768 desktop on this smaller screen WITHOUT shrinking the desktop to fit. I want to pan around. Thanks! Cajuntechie

    Read the article

  • Wireless traffic stops when downloading large files at high speed: packets lost (Linksys WRT120N router)

    - by Torious
    The problem Note: First I'd like to understand WHY this is happening. Ofcourse, a solution would be nice too. :) When downloading a large file over HTTP at high-speeds, my wireless traffic basically stops: I can't open webpages and the download itself pauses. It pauses pretty much immediately after starting it; sometimes at 800 KB, sometimes at a few MB. After some time, the download (and other traffic) resumes, but the problem keeps reoccurring during the same download. The problem does not occur when using a wired connection through the same router (Linskys WRT120N). Also note that the connection is not dropped when this happens. It's just that the traffic stops and I can't browse to web pages, etc. (SYN packets are sent but nothing is received, etc.) Inspection with Wireshark shows that the following happens: Server sends data packets which are acknowledged by client Server sends a packet, but SEQ indicates some packets were lost (6 packets in one occurrence). Server sends a few more packets and client acknowledges these using "selective acknowledgement" Server stops sending data for a while (since the lost packets were not acknowledged or the router stops forwarding them?) Eventually, server does a "retransmission" and traffic resumes as normal. This all seems normal behavior to me when packet loss occurs. It's the consistent packet loss throughout a large, high-speed download that puzzles me. What might cause this? My own idea is the following: My internet is pretty fast (100 mbps), so when starting a large-file download, the router buffers the incoming data (since wireless introduces some slight delay / lower speed, in part due to other networks), but the buffer overflows and the router drops packets to regulate traffic (and because it has no choice). But how could that happen? Doesn't the TCP window size limit the amount of data that can go unacknowledged? So how can the router's buffer overflow if there can only be like 64 KB waiting to be acknowledged? Note: I've disabled TCP window scaling and dynamic window size through netsh options, in an attempt to fix this, but it doesn't seem to matter. Also, Wireshark shows a pattern of the server sending 2 packets (of 1514 bytes) and the client sending an ACK, so does that rule out a possible buffer overflow? And a few more subsequent packets are received... I'm at a loss here. Thanks for any insights. Things that are (probably) NOT the cause / I have experimented with The browser Various TCP options in Windows 7 (netsh etc.) Router settings such as MTU, beacon interval, UPnP, ...

    Read the article

< Previous Page | 36 37 38 39 40 41 42 43 44 45 46 47  | Next Page >