subversion-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Nico Kadel-Garcia <>
Subject Re: SVN 1.6.17 dump is growing larger than repository size (approx. more than 10 times)
Date Wed, 25 Nov 2015 12:41:54 GMT
On Wed, Nov 25, 2015 at 4:25 AM, arun prasath <> wrote:
> Hello Team,
> I am creating Subversion 1.6.17 dump for a repository hosted in Linux
> server. SVNSERVE is serving the repository. We are migrating to SVN 1.9.2.
> While creating the dump for repository of size 1.8 GB (revisions 3000+), the
> dump command completes revision 119 and hangs and keep updating the dumpfile
> which grows to 7-8 GBs. dump command is not moving to next revision but kept
> updating the dump file. So, i just closed the putty to stop creating the
> dump.

Which Linux? And are you using the verndor provided version, or a
locally compiled one? Subversion 1.6 was last up to 1.6.23: is there
any reason you can't upgrade that? Ideally with a hotcopy or "rsync"
based copy of the repository to somehwere else, for testing the copy?

> However, I ran the svnadmin verify which completes revision 119 and take
> some time to verify revision 120 and complete the verification successfully
> for all the repository revisions.

Hmm. You might consider skipping the svnadmin dump command, and simply
setting up an svnsync mirror on the Subversion 1.9 server. Even using
"rsync" to bring the old repository over should let you run a dump and
reload on the server with Subversion 1.9, as long as there's not some
other weird corruption with revision 119 or 120.

> Since, there is no error output I have no logs to attach. Please suggest the
> work around for this situation. How can create the dump and migrate to
> target server. I am planning to use rsync from the server. I will post how
> it goes.
> I am expecting the workaround to this situation and let me know if this is a
> known issue in SVN 1.6.17. This is my active repository and created the dump
> without stopping the svnserve program in the production server.

It's not one I've seen, but I don't let my repositories get that big.
And I'd look at revisions 119 and 120 to see if there were erroneous
commits of huge binary files, in which case I'd want to exclude them
from the backup.

View raw message