You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@subversion.apache.org by Olivier Meunier <om...@eyrolles.com> on 2004/11/30 13:37:01 UTC
Corrupted repository
Hi,
I upgraded subversion from 1.0.2 to 1.0.8 yesterday. Everything was
fine. Then I build subversion again (to have python2.2-subversion).
Something went wrong but I don't know what. Now, the repository is broken.
When I run "recover" :
$ svnadmin recover SVN
Please wait; recovering the repository may take some time...
Recovery completed.
The latest repos revision is 1037.
It seems to be fine but...
I run "dump" (or verify) :
$ svnadmin dump SVN > dump01
* Dumped revision 0.
svn: Berkeley DB error while reading node revision for filesystem SVN/db:
DB_RUNRECOVERY: Fatal error, run database recovery
svn: Berkeley DB error while closing 'nodes' database for filesystem SVN/db:
DB_RUNRECOVERY: Fatal error, run database recovery
I tried to recover with db_recover, it seems to be fine too :
$ db4.2_recover -evch SVN/db/
db_recover: Finding last valid log LSN: file: 230 offset 721703
db_recover: Recovery starting from [230][28]
db_recover: Recovery complete at Tue Nov 30 14:30:33 2004
db_recover: Maximum transaction ID 80101a12 Recovery checkpoint
[230][721703]
But when I run "dump", I got the same error message.
My system is:
Linux 2.6.9 (I tried on a 2.4.26 too)
Subversion 1.0.8
bdb 4.2.52
I don't really know what to do, is there another way to dump the data
(even manualy) ?
Thanks
--
Olivier Meunier
omeunier@eyrolles.com
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Re: Corrupted repository
Posted by Max Bowsher <ma...@ukf.net>.
Olivier Meunier wrote:
> Hi,
>
> I upgraded subversion from 1.0.2 to 1.0.8 yesterday. Everything was
> fine. Then I build subversion again (to have python2.2-subversion).
>
> Something went wrong but I don't know what. Now, the repository is broken.
>
> When I run "recover" :
>
> $ svnadmin recover SVN
> Please wait; recovering the repository may take some time...
>
> Recovery completed.
> The latest repos revision is 1037.
>
> It seems to be fine but...
>
> I run "dump" (or verify) :
>
> $ svnadmin dump SVN > dump01
> * Dumped revision 0.
> svn: Berkeley DB error while reading node revision for filesystem SVN/db:
> DB_RUNRECOVERY: Fatal error, run database recovery
> svn: Berkeley DB error while closing 'nodes' database for filesystem
> SVN/db:
> DB_RUNRECOVERY: Fatal error, run database recovery
>
> I tried to recover with db_recover, it seems to be fine too :
>
> $ db4.2_recover -evch SVN/db/
> db_recover: Finding last valid log LSN: file: 230 offset 721703
> db_recover: Recovery starting from [230][28]
> db_recover: Recovery complete at Tue Nov 30 14:30:33 2004
> db_recover: Maximum transaction ID 80101a12 Recovery checkpoint
> [230][721703]
>
> But when I run "dump", I got the same error message.
>
> My system is:
>
> Linux 2.6.9 (I tried on a 2.4.26 too)
> Subversion 1.0.8
> bdb 4.2.52
>
> I don't really know what to do, is there another way to dump the data
> (even manualy) ?
Would you be able to upgrade to svn 1.1.1? The 1.1.x series gives much more
detailed error messages when BDB problems occur.
Ma.x
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Re: Corrupted repository
Posted by John Szakmeister <jo...@szakmeister.net>.
On Tuesday 30 November 2004 08:37, Olivier Meunier wrote:
> Hi,
>
> I upgraded subversion from 1.0.2 to 1.0.8 yesterday. Everything was
> fine. Then I build subversion again (to have python2.2-subversion).
>
> Something went wrong but I don't know what. Now, the repository is
> broken.
Did you upgrade DB along the way? That seems to be the most common
problem people are running into. :/
[snip]
> I run "dump" (or verify) :
>
> $ svnadmin dump SVN > dump01
> * Dumped revision 0.
> svn: Berkeley DB error while reading node revision for filesystem
> SVN/db: DB_RUNRECOVERY: Fatal error, run database recovery
> svn: Berkeley DB error while closing 'nodes' database for filesystem
> SVN/db: DB_RUNRECOVERY: Fatal error, run database recovery
>
> I tried to recover with db_recover, it seems to be fine too :
>
> $ db4.2_recover -evch SVN/db/
> db_recover: Finding last valid log LSN: file: 230 offset 721703
> db_recover: Recovery starting from [230][28]
> db_recover: Recovery complete at Tue Nov 30 14:30:33 2004
> db_recover: Maximum transaction ID 80101a12 Recovery checkpoint
> [230][721703]
>
> But when I run "dump", I got the same error message.
can you try doing an ldd /path/to/svn or ldd /path/to/mod_dav_svn if you
using http:// access? In general we see problems with "Invalid
Arguments", "Invalid Environments", etc. when people mix version. I'm
not sure what would cause you to fail to read the nodes table other than
a permissions issue... which it should've informed you of.
> My system is:
>
> Linux 2.6.9 (I tried on a 2.4.26 too)
> Subversion 1.0.8
> bdb 4.2.52
>
> I don't really know what to do, is there another way to dump the data
> (even manualy) ?
You can do it manually by using the db4.2_dump command. You will of
course need use db4.2_load to restore the table elsewhere. Just make
sure to get all of the relevant tables in the db directory (nodes,
representations, revisions, strings, transactions, and uuids).
-John
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org