[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Stone-dead repository

From: Stuart Robertson <dogmatix_at_absolutesys.com>
Date: 2004-02-20 10:36:22 CET

Yesterday one of our developers tried to check in 400MB of product
documentation (Word docs) into a test repository, and half-way through the
commit failed. When the commit was re-attempted, the result was an HTTP
timeout.

 

When I was asked to look into the problem, I initially thought that Apache
had died for some reason. We have two three SVN repositories along with the
test repository, all hosted from the save server machine. When I checked,
the other repositories were working fine. Apache was running without any
problems.

 

Next I logged onto the svn server, thinking that perhaps there was a problem
with the test repository itself. After shutting down Apache, I tried
'svnlook info /svnrepos/repositories/TestRepos', but the svnlook process
simply hung.

 

I tried running 'svnadmin lstxns /svnrepos/repositories/TestRepos', but it
too hung. Next I tried 'svnadmin recover /svnrepos/repositories/TestRepos'
which stated that it might take a long while. Half an hour later I tried
Ctrl+Break, to no avail. The only way to stop the 'svnadmin recover' was
'killall -9 svnadmin' (killall svnadmin had no effect).

 

Eventually I restored the backup from the night before, effectively losing
one day's commits to the test repository.

 

This has resulted in some perception issues. We have internally been
running a pilot project on SVN for about 9 months now, without any problems.
Just recently we have begun the process of migrating two additional projects
from VSS to SVN. Again, so far without problems. Our test repository is
just that, a test repository that all staff members in the company have
commit-rights to, primarily used to allow staff members to become familiar
with the copy-modify-merge model used by SVN.

 

As a result of today's little adventure I have two questions:

 

1) Does anyone know what might have caused the repository to die (or
become corrupted?)?

2) What strategies can I adopt to enable a commit-by-commit backup
(incremental preferably) of each repository so that, in the event of
something going wrong, I can (hopefully fairly quickly) restore the
repository to a working state and lost at most the last commit.

 

I've seen some discussion on the users mailing list about problems
committing large binary files (400MB+ if I remember correctly). The largest
Word document committed was around 35MB. Are there any known issues with
committing binary files into an SVN repository.

 

Our server configuration is as follows:

 

- RedHat Linux 9.3

- Subversion-0.37.0-1

- subversion-tools-0.37.0-1

- apr-0.9.5-0.2

- neon-0.24.4-1

- httpd-2.0.48-3

- 256MB RAM

 

Interestingly enough, when we retried the documentation commit after
restoring the backup, the commit succeeded, so the problem doesn't appear to
be a consistent one.

 

Any pointers/recommendations for a good incremental backup strategy will be
most appreciated.

 

Regards,

Stuart.

 

DISCLAIMER: This information is intended only for the person or entity to
which it is addressed and may contain private, confidential, proprietary
and/or privileged material and may be subject to confidentiality agreements.
Any review, retransmission, dissemination, or any other use of or taking of
any action in reliance upon this information, by persons or entities other
than the intended recipient, is prohibited. If you received this in error,
please contact the sender and delete the material from all storage media.

Absolute Systems (Pty) Ltd
Tel: +27 (0)11 784 0078
Fax: +27 (0)11 784 0148
Web: <http://www.absolutesys.com/> www.absolutesys.com

 
Received on Fri Feb 20 10:37:03 2004

This is an archived mail posted to the Subversion Users mailing list.

This site is subject to the Apache Privacy Policy and the Apache Public Forum Archive Policy.