On 2003-12-12 11:20:38 +0100, Jan Hendrik wrote:
----8<----
> Not having tested this so far here I have an idea that may serve as
> at least indirect proof for a successful execution of deltify:
>
> Assuming that svnadmin deltify issues error codes - developers
> requested here! else trial&error should help, too - you may
> intercept these and only in case of success execute another
> command: svnadmin list-unused-dblogs REPOS_PATH | xargs rm.
>
> So if logfiles keep accumulating this would hint to deltafication not
> being executed. Can't say if mailer.py could be utilized for control
> of success, too. When some day repositories may become more
> stable here and I find some leisure time I might get into this. So far
> I leave logfiles untouched here, cannot forget that my first repos
> corruptions just happened after removing unused logs.
>
> Best regards
>
> Jan Hendrik
I can vouch for the start syntax working. I've been using it for a
couple of weeks to do both incremental dumps and deltification for our
repository. The dumps are definitely getting created just fine, and I
haven't noticed excessive database growth, so I assume deltification is
working. The start command always seems to return a zero error code, so
you can't really check the result of the svnadmin call within the
post-commit script. I suppose you couold write a wrapper batch file
that post-commit calls that will log the error code somewhere.
I've attached my post-commit script if it will help.
-Dominic
- application/pgp-signature attachment: stored
Received on Fri Dec 12 22:00:00 2003