Path: news.f.de.plusline.net!news-fra1.dfn.de!newsfeed.hanau.net!news.tiscali.de!newsfeed.freenet.de!newspeer1.nwr.nac.net!border2.nntp.dca.giganews.com!nntp.giganews.com!out03b.usenetserver.com!news.usenetserver.com!in02.usenetserver.com!news.usenetserver.com!postnews.google.com!q19g2000prn.googlegroups.com!not-for-mail From: Jimmy Newsgroups: comp.databases.oracle.server Subject: Re: oracle threads increasing, even after closing sessions Date: Thu, 14 Jun 2007 13:56:14 -0000 Organization: http://groups.google.com Lines: 146 Message-ID: <1181829374.476036.222990@q19g2000prn.googlegroups.com> References: <1181760622.819384.66390@q19g2000prn.googlegroups.com> <1181761749.935135.117340@z28g2000prd.googlegroups.com> <1181807312.515867.91910@x35g2000prf.googlegroups.com> <1181809993.937836.68900@d30g2000prg.googlegroups.com> <1181818466.520119.67400@q19g2000prn.googlegroups.com> <1181828546.213798.42190@o11g2000prd.googlegroups.com> NNTP-Posting-Host: 57.67.161.6 Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" X-Trace: posting.google.com 1181829375 9205 127.0.0.1 (14 Jun 2007 13:56:15 GMT) X-Complaints-To: groups-abuse@google.com NNTP-Posting-Date: Thu, 14 Jun 2007 13:56:15 +0000 (UTC) In-Reply-To: <1181828546.213798.42190@o11g2000prd.googlegroups.com> User-Agent: G2/1.0 X-HTTP-UserAgent: Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; SV1; .NET CLR 1.1.4322; .NET CLR 2.0.50727; .NET CLR 3.0.04506.30),gzip(gfe),gzip(gfe) Complaints-To: groups-abuse@google.com Injection-Info: q19g2000prn.googlegroups.com; posting-host=57.67.161.6; posting-account=v16DpQ0AAAAdxjlueYjLd2tp3vOU4ydX Xref: news.f.de.plusline.net comp.databases.oracle.server:199301 On 14 jun, 15:42, EscVector wrote: > On Jun 14, 6:54 am, Cristian Cudizio > wrote: > > > > > > > On Jun 14, 10:33 am, Jimmy wrote: > > > > On 14 jun, 09:48, Jimmy wrote: > > > > > On 13 jun, 21:09, "fitzjarr...@cox.net" wrote: > > > > > > On Jun 13, 1:50 pm, Jimmy wrote: > > > > > > > Dear, > > > > > > We have a problem with 2 Oracle Server 10.2.0.0 on Windows 2003 > > > > > > Server. > > > > > > The number of threads in the oracle.exe process keeps increasing, even > > > > > > after the user session is finished. > > > > > > We currently have +- 35 session open, and already 136 threads ! > > > > > > We already changed the PROCESSES ini parameter to 300, because we > > > > > > received ORA-00020 error. > > > > > > It seems that the threads are not properly cleaned up, after the > > > > > > session is finished. > > > > > > Is PMON not supposed to clean this up ? > > > > > > > Any ideas ? > > > > > > Hmmm, 10.2.0.0 .... I thought 10.2.0.2 was the base release... > > > > > > You have checked metalink for this issue, I presume. And you do > > > > > realise that 10gR2 is certified for Windows 2003 server ONLY at the > > > > > 10.2.0.2 and 10.2.0.3 patch levels. > > > > > > It appears you need to patch your installation. > > > > > > David Fitzjarrell > > > > > Dear, > > > > sorry litle mistake, the oracle version we're using is 10.2.0.3 .- Tekst uit oorspronkelijk bericht niet weergeven - > > > > > - Tekst uit oorspronkelijk bericht weergeven - > > > > After reboot of the server, still the same problem :-( > > > It runs fine for a couple of hours, but after that, threads are > > > increasing, > > > session seem to disappear properly, but the thread remains . > > > mmmh, i suspect it's slightly different from my problem. Threads on > > windows are process on Oracle, if you count > > a lot of process on v$process it seems as if clients are closing > > sessions but non disconnecting, as suggest > > Thomas Kyte on his last book you can see that by connecting with > > sqlplus, then launching "disconnect" command, > > the session is being closed, but the shadow process (so the windows > > thread) remains. > > Then remains my suspect that Oracle is unable to cleanup such > > processes if clients close uncleanly. Maybe that > > activatinc DCD (Dead Connection Detection) may help, but it is subject > > to TCP timeout that on windows defaults to 2 hours. > > > Bye > > > Cristian Cudizio > > >http://oracledb.wordpress.comhttp://cristiancudizio.wordpress.com > > NO CONNECTIONS > C:\work\admin\sysinternals>pslist |find "oracle" > oracle 2036 8 25 << Threads 541 655484 > 0:00:07.687 1:23:36.435 > > Launched 10 connections via sqlplus and stared closing the cmd > windows via the x. > > C:\work\admin\sysinternals>pslist |find "oracle" > oracle 2036 8 35 << Threads 700 669164 > 0:00:07.828 1:30:10.060 > > C:\work\admin\sysinternals>pslist |find "oracle" > oracle 2036 8 32 << Threads 649 665336 > 0:00:07.843 1:31:20.795 > > C:\work\admin\sysinternals>pslist |find "oracle" > oracle 2036 8 30 << Threads 623 662332 > 0:00:07.843 1:35:36.998 > > C:\work\admin\sysinternals>pslist |find "oracle" > oracle 2036 8 29 << Threads 608 661024 > 0:00:07.843 1:35:43.513 > > C:\work\admin\sysinternals>pslist |find "oracle" > oracle 2036 8 28 << Threads 593 659656 > 0:00:07.843 1:35:59.341 > > C:\work\admin\sysinternals>pslist |find "oracle" > oracle 2036 8 27 << Threads 578 658356 > 0:00:07.859 1:36:15.795 > > Here's where it get interesting.... > C:\work\admin\sysinternals>pslist |find "oracle" > oracle 2036 8 26 563 657064 0:00:07.937 > 1:42:09.795 > > Last cmd window open with sqlplus connected. > Issued Disconnect and checked threads with cmd window still open but > sqlplus closed. > > C:\work\admin\sysinternals>pslist |find "oracle" > oracle 2036 8 26 << Threads Remain 563 657064 > 0:00:07.953 1:43:07.138 > > Double check.... > C:\work\admin\sysinternals>pslist |find "oracle" > oracle 2036 8 26 563 657064 0:00:07.953 > 1:43:24.888 > > Closed cmd window and viola: > C:\work\admin\sysinternals>pslist |find "oracle" > oracle 2036 8 25 << Tread is gone.... 546 655484 > 0:00:07.968 1:43:31.060 > > Are you launching the connections via cmd prompt from batch file? > Are users using sqlplus and not exiting the window? > > check to make sure the cmd prompts are not really the issue. > If they are hanging around after disconnect, the threads will increase > and never go away.... > > Anytime a cmd prompt is not exited, it starts to kill the machine, > slowly, but surely. > > Running batch this way is just plain bad. But this might not be your > issue.- Tekst uit oorspronkelijk bericht niet weergeven - > > - Tekst uit oorspronkelijk bericht weergeven - Hi, in our case the number of threads is not decreasing, if the clients close their program, the number of threads stays the same