Search Results

Search found 16809 results on 673 pages for 'nathan long'.

Page 15/673 | < Previous Page | 11 12 13 14 15 16 17 18 19 20 21 22  | Next Page >

  • Long variable names

    - by RaouL
    Lets say i have a variable that contains the number of search engine names in a file, what would you name it? number_of_seach_engine_names search_engine_name_count num_search_engines engines engine_names other name? The first name describes what the variable contains precisely, but isn't it too long?, any advice for choosing variable names? especially how to shorten a name that is too long or what kind of abbreviations to use?

    Read the article

  • Cannot convert object, recieved from ajax call, into a long

    - by Matt
    I'm using Asp.Net-Mvc, I have this method in my controller: [AcceptVerbs(HttpVerbs.Post)] public ActionResult LinkAccount(string site, object id) { return this.Json(id); } Here's the ajax method that calls it: $.post("/Account/LinkAccount", { site: "Facebook", id: FB.Facebook.apiClient.get_session().uid }, function(result) { alert(result); }, "json" ); returning this.Json(id); makes the alert work... it alerts 7128383 (something similar to that). but if I change this.Json(id) to this.Json(Conver.ToInt64(id)); the alert does not fire... Any idea of why I can't convert an object received from an object to a long? I already know changing the LinkAccount method to accept a long instead works just fine. It's just I need it as an object because some other sites I'm linking up have strings for id's rather than longs. UPDATE: I tried running the code on localhost so I could set a breakpoint. First I changed the line return this.Json(Convert.ToInt64(id)); to long idAsLong = Convert.ToInt64(id));. Here's what the debugger is telling me: When I hover over id it says: "id | {string[1]}" and when I press the plus button is shows: "[0] | '7128383'" When I hover over idAsLong, it says: "idAsLong | 0" Why isn't it converting it properly?

    Read the article

  • NSURLConnection receives data even if no data was thrown back

    - by Anna Fortuna
    Let me explain my situation. Currently, I am experimenting long-polling using NSURLConnection. I found this and I decided to try it. What I do is send a request to the server with a timeout interval of 300 secs. (or 5 mins.) Here is a code snippet: NSURL *url = [NSURL URLWithString:urlString]; NSURLRequest *request = [NSURLRequest requestWithURL:url cachePolicy:NSURLCacheStorageAllowedInMemoryOnly timeoutInterval:300]; NSData *data = [NSURLConnection sendSynchronousRequest:request returningResponse:&resp error:&err]; Now I want to test if the connection will "hold" the request if no data was thrown back from the server, so what I did was this: if (data != nil) [self performSelectorOnMainThread:@selector(dataReceived:) withObject:data waitUntilDone:YES]; And the function dataReceived: looks like this: - (void)dataReceived:(NSData *)data { NSLog(@"DATA RECEIVED!"); NSString *string = [NSString stringWithUTF8String:[data bytes]]; NSLog(@"THE DATA: %@", string); } Server-side, I created a function that will return a data once it fits the arguments and returns none if nothing fits. Here is a snippet of the PHP function: function retrieveMessages($vardata) { if (!empty($vardata)) { $result = check_data($vardata) //check_data is the function which returns 1 if $vardata //fits the arguments, and 0 if it fails to fit if ($result == 1) { $jsonArray = array('Data' => $vardata); echo json_encode($jsonArray); } } } As you can see, the function will only return data if the $result is equal to 1. However, even if the function returns nothing, NSURLConnection will still perform the function dataReceived: meaning the NSURLConnection still receives data, albeit an empty one. So can anyone help me here? How will I perform long-polling using NSURLConnection? Basically, I want to maintain the connection as long as no data is returned. So how will I do it? NOTE: I am new to PHP, so if my code is wrong, please point it out so I can correct it.

    Read the article

  • Short names versus long names in Windows

    - by normski
    I have some code which gets the short name from a file path, using GetShortNameW(), and then later retrieves the long name view GetLongNameA(). The original file is of the form "C:/ProgramData/My Folder/File.ext" However, following conversion to short, then back to long, the filename becomes "C:/Program Files/My Folder/Filename.ext". The short name is of the form "C:/PROGRA~2/MY_FOL~1/FIL~1.EXT" The short name is being incorrectly resolved. The code compiles using VS 2005 on Windows 7 (I cannot upgrade the project to VS2008) Does anybody have any idea why this might be happening? DWORD pathLengthNeeded = ::GetShortPathNameW(aRef->GetFilePath().c_str(), NULL, 0); if(pathLengthNeeded != 0) { WCHAR* shortPath = new WCHAR[pathLengthNeeded]; DWORD newPathNameLength = ::GetShortPathNameW(aRef->GetFilePath().c_str(), shortPath, pathLengthNeeded); if(newPathNameLength != 0) { UI_STRING unicodePath(shortPath); std::string asciiPath = StringFromUserString(unicodePath); pathLengthNeeded = ::GetLongPathNameA(asciiPath.c_str(),NULL, 0); if(pathLengthNeeded != 0) {// convert it back to a long path if possible. For goodness sake can't we use Unicode throughout?F char* longPath = new char[pathLengthNeeded]; DWORD newPathNameLength = ::GetLongPathNameA(asciiPath.c_str(), longPath, pathLengthNeeded); if(newPathNameLength != 0) { std::string longPathString(longPath, newPathNameLength); asciiPath = longPathString; } delete [] longPath; } SetFullPathName(asciiPath); } delete [] shortPath; }

    Read the article

  • reshaping a data frame into long format in R

    - by user1773115
    I'm struggling with a reshape in R. I have 2 types of error (err and rel_err) that have been calculated for 3 different models. This gives me a total of 6 error variables (i.e. err_1, err_2, err_3, rel_err_1, rel_err_2, and rel_err_3). For each of these types of error I have 3 different types of predivtive validity tests (ie random holdouts, backcast, forecast). I would like to make my data set long so I keep the 4 types of test long while also making the two error measurements long. So in the end I will have one variable called err and one called rel_err as well as an id variable for what model the error corresponds to (1,2,or 3) Here is my data right now: iter err_1 rel_err_1 err_2 rel_err_2 err_3 rel_err_3 test_type 1 -0.09385732 -0.2235443 -0.1216982 -0.2898543 -0.1058366 -0.2520759 random 1 0.16141630 0.8575728 0.1418732 0.7537442 0.1584816 0.8419816 back 1 0.16376930 0.8700738 0.1431505 0.7605302 0.1596502 0.8481901 front 1 0.14345986 0.6765194 0.1213689 0.5723444 0.1374676 0.6482615 random 1 0.15890059 0.7435382 0.1589823 0.7439204 0.1608709 0.7527580 back 1 0.14412360 0.6743928 0.1442039 0.6747684 0.1463520 0.6848202 front and here is what I would like it to look like: iter model err rel_err test_type 1 1 -0.09385732 (#'s) random 1 2 -0.1216982 (#'s) random 1 3 -0.1216982 (#'s) random and on... I've tried playing around with the syntax but can't quite figure out what to put for the time.varying argument Thanks very much for any help you can offer.

    Read the article

  • Short file names versus long file names in Windows

    - by normski
    I have some code which gets the short name from a file path, using GetShortNameW(), and then later retrieves the long name view GetLongNameA(). The original file is of the form "C:/ProgramData/My Folder/File.ext" However, following conversion to short, then back to long, the filename becomes "C:/Program Files/My Folder/Filename.ext". The short name is of the form "C:/PROGRA~2/MY_FOL~1/FIL~1.EXT" The short name is being incorrectly resolved. The code compiles using VS 2005 on Windows 7 (I cannot upgrade the project to VS2008) Does anybody have any idea why this might be happening? DWORD pathLengthNeeded = ::GetShortPathNameW(aRef->GetFilePath().c_str(), NULL, 0); if(pathLengthNeeded != 0) { WCHAR* shortPath = new WCHAR[pathLengthNeeded]; DWORD newPathNameLength = ::GetShortPathNameW(aRef->GetFilePath().c_str(), shortPath, pathLengthNeeded); if(newPathNameLength != 0) { UI_STRING unicodePath(shortPath); std::string asciiPath = StringFromUserString(unicodePath); pathLengthNeeded = ::GetLongPathNameA(asciiPath.c_str(),NULL, 0); if(pathLengthNeeded != 0) {// convert it back to a long path if possible. For goodness sake can't we use Unicode throughout?F char* longPath = new char[pathLengthNeeded]; DWORD newPathNameLength = ::GetLongPathNameA(asciiPath.c_str(), longPath, pathLengthNeeded); if(newPathNameLength != 0) { std::string longPathString(longPath, newPathNameLength); asciiPath = longPathString; } delete [] longPath; } SetFullPathName(asciiPath); } delete [] shortPath; }

    Read the article

  • Is long long in C++ known to be very nasty in terms of precision?

    - by Kevin
    The Given Problem: Given a theater with n rows, m seats, and a list of seats that are reserved. Given these values, determine how many ways two friends can sit together in the same row. So, if the theater was a size of 2x3 and the very first seat in the first row was reserved, there would be 3 different seatings that these two guys can take. The Problem That I'm Dealing With The function itself is supposed to return the number of seatings that there are based on these constraints. The return value is a long long. I've gone through my code many many times...and I'm pretty sure that it's right. All I'm doing is incrementing this one value. However, ALL of the values that my function return differ from the actual solution by 1 or 2. Any ideas? And if you think that it's just something wrong with my code, please tell me. I don't mind being called an idiot just as long as I learn something.

    Read the article

  • wpf window refresh works at first, then stops

    - by mcl
    I've got a routine that grabs a list of all images in a directory, then runs an MD5 digest on all of them. Since this takes a while to do, I pop up a window with a progress bar. The progress bar is updated by a lambda that I pass in to the long-running routine. The first problem was that the progress window was never updated (which is normal in WPF I guess). Since WPF lacks a Refresh() command I fixed this with a call to Dispatcher.Invoke(). Now the progress bar is updated for a while, then the window stops being updated. The long-running work does eventually finish and the windows go back to normal. I have already tried a BackgroundWorker and quickly became frustrated by a threading issue related to an event triggered by the long-running process. So if that's really the best solution and I just need to learn the paradigm better, please say so. But I'd be really much happier with the approach I've got here, except that it stops updating after a bit (for example, in a folder with 1000 files, it might update for 50-100 files, then "hang"). The UI does not need to be responsive during this activity, except to report on progress. Anyway, here's the code. First the progress window itself: public partial class ProgressWindow : Window { public ProgressWindow(string title, string supertext, string subtext) { InitializeComponent(); this.Title = title; this.SuperText.Text = supertext; this.SubText.Text = subtext; } internal void UpdateProgress(int count, int total) { this.ProgressBar.Maximum = Convert.ToDouble(total); this.ProgressBar.Value = Convert.ToDouble(count); this.SubText.Text = String.Format("{0} of {1} finished", count, total); this.Dispatcher.Invoke(DispatcherPriority.Render, EmptyDelegate); } private static Action EmptyDelegate = delegate() { }; } <Window x:Class="Pixort.ProgressWindow" xmlns="http://schemas.microsoft.com/winfx/2006/xaml/presentation" xmlns:x="http://schemas.microsoft.com/winfx/2006/xaml" Title="Pixort Progress" Height="128" Width="256" WindowStartupLocation="CenterOwner" WindowStyle="SingleBorderWindow" ResizeMode="NoResize"> <DockPanel> <TextBlock DockPanel.Dock="Top" x:Name="SuperText" TextAlignment="Left" Padding="6"></TextBlock> <TextBlock DockPanel.Dock="Bottom" x:Name="SubText" TextAlignment="Right" Padding="6"></TextBlock> <ProgressBar x:Name="ProgressBar" Height="24" Margin="6"/> </DockPanel> </Window> The long running method (in Gallery.cs): public void ImportFolder(string folderPath, Action<int, int> progressUpdate) { string[] files = this.FileIO.GetFiles(folderPath); for (int i = 0; i < files.Length; i++) { // do stuff with the file if (null != progressUpdate) { progressUpdate.Invoke(i + 1, files.Length); } } } Which is called thusly: ProgressWindow progress = new ProgressWindow("Import Folder Progress", String.Format("Importing {0}", folder), String.Empty); progress.Show(); this.Gallery.ImportFolder(folder, ((c, t) => progress.UpdateProgress(c, t))); progress.Close();

    Read the article

  • Sending Adobe PDF attachments from Adobe Reader (in Outlook 2003) takes too long

    - by White Island
    I have a customer who is using Outlook 2003 (Microsoft Online Services) and Adobe reader 9+. When they send a PDF from Adobe reader to Outlook (via the Send as attachment to e-mail feature in Adobe), it freezes for 30 seconds to 5 minutes before the new e-mail pops up with the PDF attachment. I'm pretty sure the issue is on the Outlook side of things, as I've tried Adobe reader 8 and Foxit Reader with the same results (Windows XP/7 doesn't seem to make a difference, either). I tried Outlook in safe mode on the first (Win7) machine I was working on, and the e-mail attachment worked a lot faster, but when I tried to replicate the results on another machine, one wouldn't go into safe mode, the other didn't seem to show a difference. In an effort to fix the problem in Outlook normal mode, I tried disabling all add-ins, Com add-in (Office Communicator is the only one), reading pane, Word 2003 as e-mail editor... but none of these seemed to address the issue. Does anyone have any other ideas? I need to get this resolved as soon as possible, and it doesn't seem practical to make them run in safe mode. :P

    Read the article

  • Mail server responds with Error 500 line too long

    - by mawimawi
    I am occasionally receiving error messages from mail servers with the above message. It seems that at least one line in the e-mail has more than 999 characters and therefore the e-mail is bounced. Is this "by design"? in an RFC? Or some weird pseudo-spam-filter? Or just a bad mailserver? I googled around a bit but did not find a competent answer. Hopefully one of you guys can enlighten me.

    Read the article

  • MS Office Communicator: Long delays in setting up audio connection when starting a call

    - by geofftnz
    I am using Microsoft Office Communicator with a USB headset as my work phone. OCS is connected to our PABX so we can take and make calls to regular, non-OCS phones. When making an external call to a cellphone, it can take up to 5-10 seconds for audio to start flowing. eg: Work Phone Cellphone - dial cellphone (ringing) (ringing) answer cellphone (hearing nothing) speak "1" . speak "2" . speak "3" . ... . speak "14" hear "15" speak "15" hear "16" speak "16" Has anyone experienced this kind of thing with an OCS setup? Any pointers?

    Read the article

  • Sending Adobe PDF attachments from Adobe Reader (in Outlook 2003) takes too long

    - by White Island
    I have a customer who is using Outlook 2003 (Microsoft Online Services) and Adobe reader 9+. When they send a PDF from Adobe reader to Outlook (via the Send as attachment to e-mail feature in Adobe), it freezes for 30 seconds to 5 minutes before the new e-mail pops up with the PDF attachment. I'm pretty sure the issue is on the Outlook side of things, as I've tried Adobe reader 8 and Foxit Reader with the same results (Windows XP/7 doesn't seem to make a difference, either). I tried Outlook in safe mode on the first (Win7) machine I was working on, and the e-mail attachment worked a lot faster, but when I tried to replicate the results on another machine, one wouldn't go into safe mode, the other didn't seem to show a difference. In an effort to fix the problem in Outlook normal mode, I tried disabling all add-ins, Com add-in (Office Communicator is the only one), reading pane, Word 2003 as e-mail editor... but none of these seemed to address the issue. Does anyone have any other ideas? I need to get this resolved as soon as possible, and it doesn't seem practical to make them run in safe mode. :P

    Read the article

  • SVN checkout/export too long to download

    - by user41671
    Hi, My checkout/export session in svn is kinda weird. The file is just a 300KB in size but the downloading keeps going and it reaches a megabytes in size. The file is in RPM format. I don't know if the file is corrupt or the SVN has a bug. I tried to download the file using web browser and seems the downloading works fine. What probably is the main problem is here?

    Read the article

  • SVN checkout/export too long to download

    - by sasayins
    Hi, My checkout/export session in svn is kinda weird. The file is just a 300KB in size but the downloading keeps going and it reaches a megabytes in size. The file is in RPM format. I don't know if the file is corrupt or the SVN has a bug. I tried to download the file using web browser and seems the downloading works fine. What probably is the main problem is here?

    Read the article

  • How long do Lithium Ion batteries normally last?

    - by Zifre
    On the laptop I have, I've had to buy a new Li-Ion batter roughly every year. I do use this computer quite a lot, but I'm wondering if this is normal. Right now, my battery is completely dead (it lasts for about 0.1 seconds), so I plan on buying a new one soon. Is there anything you can do to prevent Li-Ion batteries from going dead so quickly?

    Read the article

  • Troubleshooting iptables and configuring it to drop the priority of long-term connections

    - by intuited
    I'm somewhat familiar with the general concepts of iptables, and would like to learn it in more detail. I'm hoping that my learning experience can also be useful. The situation: I'm running dd-wrt on my router. Despite its purported QoS skills, I'm still seeing connection latency shoot up hugely whenever there's an ongoing http connection, eg some large download. Under such conditions, it can take 10 seconds or more to load a basic webpage; sometimes the connections are dropped entirely. I've tried adjusting the parameters, dropping the allotted bandwidth for up and download to well under my limit, but nothing seems to work. dd-wrt is configured to use HTB as the QoS algorithm; HFSC, although presented as an option, seems to cause the router to crash, and is rumoured to not actually work on any linux system. I'd like to be able to troubleshoot this issue and hopefully improve the settings that dd-wrt is using, but I'm finding the learning curve a bit overwhelming. For starters I am not sure what HTB actually specifies: is this a set of iptables commands, or do some of those commands specify how HTB is to be used? I would like it to prioritize based on protocol the way that it already supposed to, and in addition I'd like to have it drop the priority of connections which have a high total byte count, say over 400KB. Also tips on utilities that can be run under dd-wrt to get more info on what's going on in there are appreciated. I've tried to get iftop to work but there were issues running curses. I'm leaning towards replacing dd-wrt with openwrt; comments on this strategy are also welcome. I suspect that I would be well advised to get a second router as a standin before trying that. It may be worth noting that my total bandwidth is pretty limited (256Kbit/s).

    Read the article

  • Long Gigabit Ethernet Run

    - by Timothy R. Butler
    I am trying to get an Gig-E network between two buildings that are approximately 260 ft. away. While some TRENDnet switches failed to be able to connect to each other over Cat 6 at that distance, two Netgear 5-port Gig-E switches do so just fine. However, it still fails after I put in place APC PNET1GB ethernet surge protectors at each end before the line connects to the respective switches. So I find myself wondering if I simply need to find a better surge protector that doesn't degrade the signal as much (if so, what kind would you recommend?) or if I should give up on copper and use fiber between the buildings. If I opt to go the latter route, I could really use some pointers. It looks like LC connectors are the most common, but I keep running into some others as well. A media converter on each end seems like the simplest solution, but perhaps a Gig-E switch with an SFP port would make more sense? Given a very limited budget, sticking with my existing copper seems best, but if it is bound to be a headache, a 100 meter fiber cable is something I think I can swing cost wise.

    Read the article

  • Shut Down took way too long because of "Background Programs"

    - by Christopher Chipps
    I tried shutting my desktop PC (with Windows7) down but after several attempts (like 4 or 5) at Start -- Shut Down, the GUI was still there and it was not shutting down. I didn't think there were any programs running when I pressed Shut Down, so I went into the taskbar (Ctrl + Alt + Del) to check out the processes. Once I did that, a screen appeared with a message stating that there are "background programs" still running and it gave me an option to "Force Shut Down" which I pressed and it shut down normally. Does anyone know why this would happen?

    Read the article

  • Internet Explorer 9 takes a long time to load websites

    - by Steve
    IE9 on Windows 7 would load Google okay, and we could search on Google okay, but almost any other website would take an inordinate amount of time to load. The loading sprite would sit there indefinitely. I uninstalled IE9 to roll back to IE8, and the same issue occurs. We've reset IE settings back to defaults, and there are no add-ons causing this. Firefox loads websites fine on this computer. Could it be an IE-specific virus/trojan? IE was displaying an incorrect/hijacked home page.

    Read the article

  • MySQL taking a long time to start

    - by Dscoduc
    I'm running Windows Server 2008 with MySQL installed and every time I reboot the server the MySQL Service doesn't start right away. A look into the Windows Eventlog shows that the MySQL Service was hung at startup. Looking at the Services.msc console shows the service state at Starting... Eventually, like 10 minutes, the MySQL Service actually finishes the startup process and the database becomes available for my Wordpress server... I looked at the MySQL .err files and didn't find anything that would indicate a delay in the statup process... Can anyone suggest a way to determine what is causing the delay, and more importantly, how to prevent the delay in the MySQL Startup? UPDATE: Here is the .err log contents from the shutdown to the startup complete. Notice the startup begins at 10:30:00 and the MySQL isn't ready for connections until 10:47:14, a full 17 minutes later: 100322 10:27:06 [Note] C:\Program Files\MySQL\MySQL Server 5.1\bin\mysqld: Normal shutdown 100322 10:27:06 [Note] Event Scheduler: Purging the queue. 0 events 100322 10:27:06 InnoDB: Starting shutdown... 100322 10:27:08 InnoDB: Shutdown completed; log sequence number 4 3854351346 100322 10:27:08 [Warning] Forcing shutdown of 1 plugins 100322 10:27:08 [Note] C:\Program Files\MySQL\MySQL Server 5.1\bin\mysqld: Shutdown complete 100322 10:30:00 [Note] Plugin 'FEDERATED' is disabled. 100322 10:30:01 InnoDB: Started; log sequence number 4 3854351346 100322 10:47:14 [Note] Event Scheduler: Loaded 0 events 100322 10:47:14 [Note] C:\Program Files\MySQL\MySQL Server 5.1\bin\mysqld: ready for connections. UPDATE 2: MySQL is configured as a service (part of the install process, nothing I did) and executes the following syntax (as it appears in the registry): "C:\Program Files\MySQL\MySQL Server 5.1\bin\mysqld" --defaults-file="C:\Program Files\MySQL\MySQL Server 5.1\my.ini" MySQL

    Read the article

  • Mysql table Crashed during Optimize, table repair taking too long

    - by hellohellosharp
    One of my vital tables was being optimzied and MySQL crashed in the middle of it. It then said the table was corrupt. I am running repair but it is taking over an hour, I need this table up ASAP (I will even truncate it if necessary). Please help me get this solved. The table has about 54 million rows. This is a MyISAM. Any additional information needed, just ask. Here are the contents of my my.cnf: [mysqld] max_connections = 850 max_user_connections = 850 query_cache_size = 128M skip-external-locking key_buffer_size = 64M max_allowed_packet = 8M table_open_cache = 256 sort_buffer_size = 4M net_buffer_length = 16K read_buffer_size = 1M read_rnd_buffer_size = 1M myisam_sort_buffer_size = 32M innodb_file_per_table tmp_table_size = 100M max_heap_table_size = 64M thread_cache_size = 8 wait_timeout=25 interactive_timeout=25 table_cache=600 innodb_buffer_pool_size = 4G innodb_thread_concurrency = 8 innodb_flush_method = O_DIRECT # This setting allows the use of asynchronous I/O in InnoDB. # The following files track usage of this resource: # - /proc/sys/fs/aio-max-nr # - /proc/sys/fs/aio-nr # Default limit is 65536, of which a single instance of mysql uses 2661 out of the box innodb_use_native_aio = 1

    Read the article

  • Unable to restore from Shadow Copy due to long filename

    - by Spongeboy
    We have shadow copy enabled on our Windows SBS 2008 server. Attempting to restore a file from shadow copy gave the following error- The source file name(s) are larger than is supported by the file system. Try moving to a location which has a shorter path name, or try renaming to shorter name(s) before attempting this operation. The filename has 67 characters, and it's shadow copy path is 170 characters. These seem to be under the NTFS limits (260?). We tried- Copying to the shortest path possible (C:) Copying to the shortest path possible on both a client computer and the server itself Is it possible to rename files in a shadow copy, before doing the copy? Any idea why the error is appearing despite the filename size appearing to be within limits?

    Read the article

  • Squid closing the connection on long HTTP GET requests

    - by Rhys
    Hello, When running a database query on a specific external site we use, Squid seems to cut off the connection after a consistent period of time (just over a minute). The query is submitted through a standard web form is that uses GET to query their database. Firefox 3 just displays a blank page. Internet Explorer throws a 'Page Cannot Be Displayed' error (tested in v6 and v8). When we perform the same query on the same machine, but bypass the Squid proxy, it works fine. The query takes about two and a half minutes to complete. There are a few timeout settings in Squid, but I honestly don't know what one to be looking at. Any possible solutions would be much appreciated. Cheers

    Read the article

  • Long connection times from PHP to MySQL on EC2

    - by Erik Giberti
    I'm having an intermittent issue connecting to a database slave with InnoDB. Intermittently I get connections taking longer than 2 seconds. These servers are hosted on Amazon's EC2. The app server is PHP 5.2/Apache running on Ubuntu. The DB slave is running Percona's XtraDB 5.1 on Ubuntu 9.10. It's using an EBS Raid array for the data storage. We already use skip name resolve and bind to address 0.0.0.0. This is a stub of the PHP code that's failing $tmp = mysqli_init(); $start_time = microtime(true); $tmp-options(MYSQLI_OPT_CONNECT_TIMEOUT, 2); $tmp-real_connect($DB_SERVERS[$server]['server'], $DB_SERVERS[$server]['username'], $DB_SERVERS[$server]['password'], $DB_SERVERS[$server]['schema'], $DB_SERVERS[$server]['port']); if(mysqli_connect_errno()){ $timer = microtime(true) - $start_time; mail($errors_to,'DB connection error',$timer); } There's more than 300Mb available on the DB server for new connections and the server is nowhere near the max allowed (60 of 1,200). Loading on both servers is < 2 on 4 core m1.xlarge instances. Some highlights from the mysql config max_connections = 1200 thread_stack = 512K thread_cache_size = 1024 thread_concurrency = 16 innodb-file-per-table innodb_additional_mem_pool_size = 16M innodb_buffer_pool_size = 13G Any help on tracing the source of the slowdown is appreciated. [EDIT] I have been updating the sysctl values for the network but they don't seem to be fixing the problem. I made the following adjustments on both the database and application servers. net.ipv4.tcp_window_scaling = 1 net.ipv4.tcp_sack = 0 net.ipv4.tcp_timestamps = 0 net.ipv4.tcp_fin_timeout = 20 net.ipv4.tcp_keepalive_time = 180 net.ipv4.tcp_max_syn_backlog = 1280 net.ipv4.tcp_synack_retries = 1 net.core.rmem_max = 16777216 net.core.wmem_max = 16777216 net.ipv4.tcp_rmem = 4096 87380 16777216 net.ipv4.tcp_wmem = 4096 87380 16777216 [EDIT] Per jaimieb's suggestion, I added some tracing and captured the following data using time. This server handles about 51 queries/second at this the time of day. The connection error was raised once (at 13:06:36) during the 3 minute window outlined below. Since there was 1 failure and roughly 9,200 successful connections, I think this isn't going to produce anything meaningful in terms of reporting. Script: date /root/database_server.txt (time mysql -h database_Server -D schema_name -u appuser -p apppassword -e '') /dev/null 2 /root/database_server.txt Results: === Application Server 1 === Mon Feb 22 13:05:01 EST 2010 real 0m0.008s user 0m0.001s sys 0m0.000s Mon Feb 22 13:06:01 EST 2010 real 0m0.007s user 0m0.002s sys 0m0.000s Mon Feb 22 13:07:01 EST 2010 real 0m0.008s user 0m0.000s sys 0m0.001s === Application Server 2 === Mon Feb 22 13:05:01 EST 2010 real 0m0.009s user 0m0.000s sys 0m0.002s Mon Feb 22 13:06:01 EST 2010 real 0m0.009s user 0m0.001s sys 0m0.003s Mon Feb 22 13:07:01 EST 2010 real 0m0.008s user 0m0.000s sys 0m0.001s === Database Server === Mon Feb 22 13:05:01 EST 2010 real 0m0.016s user 0m0.000s sys 0m0.010s Mon Feb 22 13:06:01 EST 2010 real 0m0.006s user 0m0.010s sys 0m0.000s Mon Feb 22 13:07:01 EST 2010 real 0m0.016s user 0m0.000s sys 0m0.010s [EDIT] Per a suggestion received on a LinkedIn question, I tried setting the back_log value higher. We had been running the default value (50) and increased it to 150. We also raised the kernel value /proc/sys/net/core/somaxconn (maximum socket connections) to 256 on both the application and database server from the default 128. We did see some elevation in processor utilization as a result but still received connection timeouts.

    Read the article

< Previous Page | 11 12 13 14 15 16 17 18 19 20 21 22  | Next Page >