mvdzwaan Posted March 23, 2011 Report Share Posted March 23, 2011 We're running the latest build on a 1Gb server (which doubles as an IIS server) with SQL Server backend on another machine. Even if we are not using the plasticSCM server it's memory usage grows. After one day it is up to 600Mb, 0% cpu usage (IIS frees its used memory so the w3wp processes are back to 30Mb each). If I stop the Plastic Server service the cpu usage goes to 25% (4 cores), and although net stop and/or the service manager claims it has stopped the service, plasticd is still running using 25% cpu and 600Mb cpu usage. I have to kill this process. When I try to connect to the server once it reaches 500Mb+ (ofcourse without stopping the process) from a workstation it is not responding (or slower than the timeout). After ending plasticd process and restarting it, everything is working fine again. Known issue ? Any hints on where to look for a solution ? Link to comment Share on other sites More sharing options...
psantosl Posted March 23, 2011 Report Share Posted March 23, 2011 It is not normal to reach that big amount of mem unless you're handling huge repos and lots of concurrent users. The only thing I can think of is some database transaction being locked or something (and that's why it "dies hard" ), otherwise there's no reason for the server to stop immediately. Is there any big query you're running against the server from a batch or something? You can take a look at loader.log.txt and check what's going on there, maybe some errors are being dumped there. Otherwise, it looks like a case for support. Link to comment Share on other sites More sharing options...
mvdzwaan Posted March 23, 2011 Author Report Share Posted March 23, 2011 We have a number of small repositories. We're not using the server at the moment (work is being done on some website projects which are not in the repository) loader.log.txt only contains a few lines per day 2011-03-17 11:54:30,560 ActivitySchedulerThread EVERY_USER_OR_GROUP c6308b5b-9413-471a-9f47-xxxxxxx Server:WBIIS INFO UpdatePerf - CmDataReader- Execute 125 ms. [sELECT * FROM seid WHERE sactive != 'T' AND sgroup = 'F'] 2011-03-17 11:54:30,732 ActivitySchedulerThread EVERY_USER_OR_GROUP c6308b5b-9413-471a-9f47-xxxxxxx Server:WBIIS INFO UpdatePerf - CmDataReader- Execute 109 ms. [sELECT * FROM seid WHERE sactive != 'T' AND sgroup = 'F'] 2011-03-17 11:54:30,966 ActivitySchedulerThread EVERY_USER_OR_GROUP c6308b5b-9413-471a-9f47-xxxxxxx Server:WBIIS INFO UpdatePerf - CmDataReader- Execute 156 ms. [sELECT * FROM seid WHERE sactive != 'T' AND sgroup = 'F'] 2011-03-17 11:54:31,107 ActivitySchedulerThread EVERY_USER_OR_GROUP c6308b5b-9413-471a-9f47-xxxxxxx Server:WBIIS INFO UpdatePerf - CmDataReader- Execute 141 ms. [sELECT * FROM seid WHERE sactive != 'T' AND sgroup = 'F'] 2011-03-17 11:54:31,404 ActivitySchedulerThread EVERY_USER_OR_GROUP c6308b5b-9413-471a-9f47-xxxxxxx Server:WBIIS INFO UpdatePerf - CmDataReader- Execute 109 ms. [sELECT * FROM seid WHERE sactive != 'T' AND sgroup = 'F'] 2011-03-17 20:00:46,828 9172 EVERY_USER_OR_GROUP 3a0ac3ad-7b91-47ec-a989-xxxxxxxx Server:WBIIS INFO Transaction - Transaction timeout -> 120000ms 2011-03-18 10:12:12,149 ActivitySchedulerThread EVERY_USER_OR_GROUP 5188f502-9ba0-498b-b88cxxxxxxxx Server:WBIIS INFO UpdatePerf - CmDataReader- Execute 125 ms. [sELECT * FROM seid WHERE sactive != 'T' AND sgroup = 'F'] 2011-03-18 15:46:38,321 11864 EVERY_USER_OR_GROUP fe343756-1b36-4816-b99c-xxxxxxxxx Server:WBIIS INFO Transaction - Transaction timeout -> 120000ms 2011-03-18 16:22:45,308 ActivitySchedulerThread EVERY_USER_OR_GROUP 378d0f64-4c6d-4397-a432-xxxxxxxxxx Server:WBIIS INFO UpdatePerf - CmDataReader- Execute 125 ms. [sELECT * FROM seid WHERE sactive = 'T' AND sgroup = 'F'] This does contain a timeout. I'll run a profiler on the sql databases Link to comment Share on other sites More sharing options...
miller Posted March 23, 2011 Report Share Posted March 23, 2011 Hi mvd, You can increase the log dump level here is how to do it, you are interrested in the server section: http://www.plasticscm.com/infocenter/technical-articles/kb-enabling-logging-for-plastic-scm-part-i.aspx Miller Link to comment Share on other sites More sharing options...
mvdzwaan Posted March 24, 2011 Author Report Share Posted March 24, 2011 It seems the only activity the plastic scm server is performing (while idle) is enumerating all LDAP objects every 5 minutes. Every 5 minutes the plasticd.exe process also grows by 2Mbyte, this would account for ~24Mbyte every hour and after one day the memory usage I'm experiencing. Possible memory leak in the LDAP provider ? Link to comment Share on other sites More sharing options...
mvdzwaan Posted March 24, 2011 Author Report Share Posted March 24, 2011 Just switched the server to the AD provider and the memory growth does not occur any more. Is it safe to just switch the provider ? cm chkdb shows no errors, but the log shows Error qn - Error The Security ID structure is invalid Link to comment Share on other sites More sharing options...
miller Posted March 24, 2011 Report Share Posted March 24, 2011 Thanks mvd, we will look into it! Miller Link to comment Share on other sites More sharing options...
Guest Posted March 29, 2011 Report Share Posted March 29, 2011 Hi, mvd: The error "The Security ID structure is invalid" seems to be a Windows error, may be a configuration issue on your Windows Server, as I have just googled. Is it the 2003 version or the 2008 version? Regarding the memory leak when using LDAP authentication mode, my concern is that the library that we use as LDAP provider could be the problem. I'll do some testing and return back to you. Thanks for your patience, Luis Link to comment Share on other sites More sharing options...
mvdzwaan Posted March 29, 2011 Author Report Share Posted March 29, 2011 Hi' date=' mvd:The error "The Security ID structure is invalid" seems to be a Windows error, may be a configuration issue on your Windows Server, as I have just googled. Is it the 2003 version or the 2008 version? Regarding the memory leak when using LDAP authentication mode, my concern is that the library that we use as LDAP provider could be the problem. I'll do some testing and return back to you. Thanks for your patience, Luis[/quote'] Windows2008r2 web edition. I did not notice any adverse effects from this error message, but had included it just to be complete. Currently we're running with AD authentication and plasticd stays at ~40Mb in process manager. Link to comment Share on other sites More sharing options...
Guest Posted March 29, 2011 Report Share Posted March 29, 2011 Hi, mvd: I haven't been able to reproduce the case; could you please tell us which LDAP software do you use, please? We continue trying to figure out what the problem could be. Thanks for your patience, Luis Link to comment Share on other sites More sharing options...
mvdzwaan Posted March 29, 2011 Author Report Share Posted March 29, 2011 Hi' date=' mvd:I haven't been able to reproduce the case; could you please tell us which LDAP software do you use, please? We continue trying to figure out what the problem could be. Thanks for your patience, Luis[/quote'] A Win2008r2 Standard 64bit server with AD. Link to comment Share on other sites More sharing options...
Guest Posted March 30, 2011 Report Share Posted March 30, 2011 Ok, so you connected to the AD server via LDAP, isn't it? That was causing the memory problem, so then you changed to native AD mode configuration, right? Link to comment Share on other sites More sharing options...
mvdzwaan Posted March 30, 2011 Author Report Share Posted March 30, 2011 Ok' date=' so you connected to the AD server via LDAP, isn't it? That was causing the memory problem, so then you changed to native AD mode configuration, right?[/quote']Correct, I was using LDAP + AD as not all of our clients are domain members, but now it seems this is possible with pure AD as well. Link to comment Share on other sites More sharing options...
Guest Posted March 30, 2011 Report Share Posted March 30, 2011 Well, I guess that the problem is related to the "The Security ID structure is invalid" that you get. It's possible that in native AD mode the exception is retrieved and all works fine and smoothly but in LDAP mode the provider cannot handle that and does not release memory. I'm studying more about the error and skimming the Novell LDAP provider code, to figure out if this is the case. It would be a good idea to try to solve the security ID error in the AD server in the meantime and then change to LDAP configuration afterwards to check if the memory leak is solved. Link to comment Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.