RE: Broken Repository always fail while svn commit

Please find below the errorlog of Apache.

FYI:
1. In the same Server we have hosted multiple repositories. All work fine
except this one which get stuck in Commit alone!
2. Attached gif file provide the error as thrown from tortoise svn
client(Latest) after commit time out.

ERROR LOG:

--> Logs after Apache started:

[Wed Aug 18 12:17:34 2004] [notice] Parent: Created child process 3136
[Wed Aug 18 12:17:34 2004] [notice] Child 3136: Child process is running
[Wed Aug 18 12:17:34 2004] [notice] Child 3136: Acquired the start mutex.
[Wed Aug 18 12:17:34 2004] [notice] Child 3136: Starting 250 worker threads.

--> Logs during Commit operation which subsequently failed:

abnormal program termination
[Wed Aug 18 12:18:33 2004] [error] Parent: child process exited with status
3 -- Aborting.
[Wed Aug 18 12:18:33 2004] [notice] Parent: Created child process 3164
[Wed Aug 18 12:18:33 2004] [notice] Child 3164: Child process is running
[Wed Aug 18 12:18:33 2004] [notice] Child 3164: Acquired the start mutex.
[Wed Aug 18 12:18:33 2004] [notice] Child 3164: Starting 250 worker threads.

Log Snippets from access.log : (Logs from the restart of Apache server till
commit failure)

192.168.1.204 - - [18/Aug/2004:12:18:28 +0800] "OPTIONS
/svn/repo885/sandbox/kas/svnutils HTTP/1.1" 401 502
192.168.1.204 - msmkasim@ften.com [18/Aug/2004:12:18:28 +0800] "OPTIONS
/svn/repo885/sandbox/kas/svnutils HTTP/1.1" 200 191
192.168.1.204 - msmkasim@ften.com [18/Aug/2004:12:18:29 +0800] "MKACTIVITY
/svn/repo885/!svn/act/7523d494-e5d8-4240-9c91-821a1829115a HTTP/1.1" 201 331
192.168.1.204 - msmkasim@ften.com [18/Aug/2004:12:18:29 +0800] "PROPFIND
/svn/repo885/sandbox/kas/svnutils HTTP/1.1" 207 459
192.168.1.204 - msmkasim@ften.com [18/Aug/2004:12:18:29 +0800] "PROPFIND
/svn/repo885/!svn/vcc/default HTTP/1.1" 207 406
192.168.1.204 - msmkasim@ften.com [18/Aug/2004:12:18:30 +0800] "CHECKOUT
/svn/repo885/!svn/bln/885 HTTP/1.1" 201 347
192.168.1.204 - msmkasim@ften.com [18/Aug/2004:12:18:30 +0800] "PROPPATCH
/svn/repo885/!svn/wbl/7523d494-e5d8-4240-9c91-821a1829115a/885 HTTP/1.1" 207
358
192.168.1.204 - msmkasim@ften.com [18/Aug/2004:12:18:31 +0800] "PROPFIND
/svn/repo885/sandbox/kas/svnutils HTTP/1.1" 207 432
192.168.1.204 - msmkasim@ften.com [18/Aug/2004:12:18:32 +0800] "CHECKOUT
/svn/repo885/!svn/ver/850/sandbox/kas/svnutils/PostCommitTriggers.bat
HTTP/1.1" 201 387

Thanks in advance for your help.

-----Original Message-----
From: Ben Collins-Sussman [mailto:sussman@collab.net]
Sent: Tuesday, August 17, 2004 8:59 PM
To: Mohamed Kasim
Cc: svn-dev-list
Subject: Re: Broken Repository always fail while svn commit

On Tue, 2004-08-17 at 03:02, Mohamed Kasim wrote:

> When commit fail, tortoise SVN client kind of hang ( time out after some
> time). After this failure, the database becomes inaccessible and needs a
> SVNADMIN RECOVER procedure to get it back. The broken transaction is
listed
> when checked up with 'svnadmin lstxns'. Now after recovery the database is
> accessible , checkout, updates, check for updates are working fine. But
> Commit always go back to step 1.

So after recovery, updates/checkouts are fine, but *any* commit at all
causes the repository to get 'stuck' again, and need recovery?

What does the apache errorlog say?

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org