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

Re: Subversion Histroy question

From: Ryan Schmidt <subversion-2010a_at_ryandesign.com>
Date: Wed, 6 Jan 2010 08:44:33 -0600

On Jan 6, 2010, at 06:28, Andreas Hoegger wrote:

> We have been using SVN for years. Our repositories raised in size over the years. To get our repositories easier to maintain/backup we are looking for a possibility to get rid of unused old revisions.
> E.g. Out of a repository of 10'000 revisions the interest is in the last 1'000 revisions. Is there a way to remove revision 1-8'999 without losing the correct revision number 10'0000.
> My idea was to dump revision 9'000 to 10'000 and load it into a new repository. The load unfortunately starts with revision 1 and ends with the head revision 1'000 what ended up in a mess since several users checked out the revision 10'000 before the dump and are not anymore able to commit against the new repository's head revision 1'000.
> Is there a suitable solution for this issue?

You could commit 9000 empty revisions to the new empty repository before you load the dumpfile. But either way the new repository is not the same as the old one, and therefore has a new UUID, and therefore all clients will need to check out new working copies (and will not be able to commit anymore from their existing working copies) regardless of whether you renumber revisions or not.

Deleting old revisions will also not necessarily save you space. It may in fact make your repository larger. The standard advice is not to attempt this procedure. See also the various prior requests for an "svnadmin obliterate" function which may have more information.

> Is there a bug request concerning this problem?

I don't think it's considered a bug.
Received on 2010-01-06 15:45:14 CET

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.