FW: Another Subversion repository gone bad [#8143]

FYI,

Sander

-----Original Message-----
From: cmpilato@collab.net [mailto:cmpilato@collab.net]
Sent: Thursday, July 10, 2003 7:52 PM
To: Michael Ubell
Cc: Sleepycat support; bostic@sleepycat.com; striker@apache.org;
jcollins@asgardsrealm.net
Subject: Re: Another Subversion repository gone bad [#8143]

First of all, Michael, I wanna thank you for patiently working through
this issue.

Michael Ubell <ubell@mindspring.com> writes:

> Is it possible that two distinct environments were operating at the
> same time? That is was the file being accessed by two processes
> using different files for the environment? This could happen if one
> process is running and another starts up specifying
> DB_RECOVER|DB_CREATE. This removes the environment files, but on
> *inx systems the files still exisit and the first process could
> continue on. This could also happen if someone removed the __db.*
> files from the environment directory.

If I'm understanding you correctly, it is dangerous to, say, run
'db_recover -ve' while an environment is in use by another process.

There is a not-to-little chance that this type of behavior has occured
on these environments. I certainly have been guilty of forgetting to,
say, shut down a hung DB-accessing process before running recovery.

> While I am not sure it could account for all this, you can run into
> trouble if your environment files are on an NFS file system. Could
> this be a possibility?

I am not aware of this being the case -- Jamin, was this the situation
in your scenario?

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Fri Aug 1 20:07:31 2003