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

Re: Suddenly corrupted Subversion repository.

From: John Szakmeister <john_at_szakmeister.net>
Date: 2004-05-29 00:38:28 CEST

> This is memory on our server.
> [root@vcs root]# free -m
> total used free shared buffers
> cached
> Mem: 1008 992 15 0 128
> 558
> -/+ buffers/cache: 305 702
> Swap: 1027 0 1027
> During the problem, I was monitoring the server. So all memory and
> 100%CPU were used by apache or svnadmin.
> May be the problem is located in the size of our repository?
> [root@vcs root]# du -sh /var/Repository/repos
> 2.9G /var/Repository/repos

Well, I tried a few things in hoping of creating a similar problem. I wrote a
script yesterday that created a repository with 23,000 revs and was about
2.4GB in size and averaging 100K a commit. All ran fine on my machine with
both 'svnadmin recover' and 'svnadmin verify'.

I find it disturbing that you can't finish verifying the repository. I
suppose that means you are also unable to dump it as well?

Is there some way I could have a copy of the repository? I'm really
interested in what is causing svnadmin to use so much memory. Watching the
verify process on my computer the memory consumption was at a constant level.

Also, can you make a copy and attempt to access the repository using file://
access? I'm curious if this presents the same problem that it does over
http:// access. If you get the same behavior, then it'll be easier to track
down the problem using that method with debugging tools.


To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Received on Sat May 29 00:38:59 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.