On Oct 27, 2004, at 7:07 AM, Lukas Cerman wrote:
> Hi all!
>
> We are unable to recover our repository. The repository is inaccesible
> since one user's commit has crashed. We tried use svnadmin recover --
> as described in manual -- but we get allways this response:
>
> raid:/tmp# svnadmin recover /tmp/all
> Please wait; recovering the repository may take some time...
> svn: DB_RUNRECOVERY: Fatal error, run database recovery
Try using the db_recover utility in catastrophic mode (-c):
$ db_recover -vech /tmp/all/db/
>
> We shifted from CVS to SVN one month ago and were very pleasured. But
> we are concerned on data consistency and this make us affraid. What
> should we do to recover our repository and to prevent it happens
> again?
See the FAQ:
http://subversion.tigris.org/project_faq.html#stuck-bdb-repos
>
> Any suggestion?
>
> We have version:
> svnadmin, version 1.0.8 (r11084)
> compiled Sep 27 2004, 01:09:34
>
Upgrade to 1.1. 'svnadmin recover' gives better error messages in 1.1,
and also, if BDB continues to create problems for you, you can try
switching to an FSFS repository instead.
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Received on Wed Oct 27 16:15:14 2004