Repository recovery problem

[Andrew Vit <an...@dihedral.ca>]

I have a repository hosted on Mac OS X 10.3 running svn 1.1.3,  
accessed via svn:// protocol. I believe my problem started after a  
system reboot (we had some storms and power failures here on the west  
coast). After trying to access the repository again, I got the  
following error via SmartSVN:


svn: Berkeley DB error while checkpointing after Berkeley DB  
transaction for filesystem /Users/Shared/Repositories/ontrack/db:  
Invalid Argument
svn: bdb: Ignoring log file: /Users/Shared/Repositories/ontrack/db/ 
log.0000000099: magic number 0, not 40988
svn: bdb: DB_ENV->log_put: 99: Invalid argument
svn: bdb: txn_checkpoint: failed to flush the buffer cache Invalid  
argument


I tried following the instructions in the svnbook for restoring a  
"wedged" repository. When I ran the svnadmin recover command, I got  
this error:


Repository lock acquired.
Please wait; recovering the repository may take some time...
svn: DB_RUNRECOVERY: Fatal error, run database recovery
svn: bdb: Ignoring log file: ontrack/db/log.0000000099: magic number  
0, not 40988
svn: bdb: Invalid log file: log.0000000099: Invalid argument
svn: bdb: First log record not found
svn: bdb: PANIC: Invalid argument


Please let me know if there's anything else I might try before going  
to my backup.

Thanks,
--Andrew Vit

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org

Re: Repository recovery problem

[John Szakmeister <jo...@szakmeister.net>]

----- Andrew Vit <an...@dihedral.ca> wrote:
> I have a repository hosted on Mac OS X 10.3 running svn 1.1.3,  
> accessed via svn:// protocol. I believe my problem started after a  
> system reboot (we had some storms and power failures here on the west 
> 
> coast). After trying to access the repository again, I got the  
> following error via SmartSVN:
> 
> 
> svn: Berkeley DB error while checkpointing after Berkeley DB  
> transaction for filesystem /Users/Shared/Repositories/ontrack/db:  
> Invalid Argument
> svn: bdb: Ignoring log file: /Users/Shared/Repositories/ontrack/db/ 
> log.0000000099: magic number 0, not 40988
> svn: bdb: DB_ENV->log_put: 99: Invalid argument
> svn: bdb: txn_checkpoint: failed to flush the buffer cache Invalid  
> argument
> 
> 
> I tried following the instructions in the svnbook for restoring a  
> "wedged" repository. When I ran the svnadmin recover command, I got  
> this error:
> 
> 
> Repository lock acquired.
> Please wait; recovering the repository may take some time...
> svn: DB_RUNRECOVERY: Fatal error, run database recovery
> svn: bdb: Ignoring log file: ontrack/db/log.0000000099: magic number 
> 
> 0, not 40988
> svn: bdb: Invalid log file: log.0000000099: Invalid argument
> svn: bdb: First log record not found
> svn: bdb: PANIC: Invalid argument
> 
> 
> Please let me know if there's anything else I might try before going 
> 
> to my backup.

You can try doing a "catastrophic recovery" using the BDB's db tools.  I can't remember the command line to drive it, you'll need to look that up in the documentation.

FWIW, BDB and Mac OS X is a very hateful relationship.  I remember there being lots of issues between the two.  You'd do better to migrate to a newer version of svn, and make your repositories FSFS instead.

-John

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org