On Wed, Jun 23, 2010 at 05:01:16PM -0700, Teri Holmes wrote:
> We have a Postgres database that is so huge(images are stored inside the
> database) that any Pg_dump configuration I tried wouldn't finish. After
> taking with the software vendor it was decided that nightly a pg_dump would
> be done excluding the table with the images in it. Then once a week the
> images would be backed up by stopping the server, tar & zipping the whole
> folder where the database resided. The database and the backup lived on
> different volumes but all on the same SAN. The SAN failed one afternoon and
> completely died the next AM. We were left with nothing. The drives were
> collected and sent to a data recovery specialist. I am fairly certain that
> the tar/zip that was created of the database folder is corrupt because
> during that couple hour window between the first drive failure the ultimate
> third drive failure. I checked out the backup and it had some issues.
>
> We purchased and configured a new SAN unit and have volumes created for the
> Postgres Database that duplicate what the old SAN looked like. So when(I
> have to think when and not IF) I get the data back, can I just copy the old
> postgres database folder contents to the new location and start the service
> back up?
>
> I've contacted the software vendor and they said...."We don't have a
> postgres expert on staff, so you are sort of on your own with that"
Provided you can get a valid zip file made while the database really was
stopped, you should be able simply to unzip it and start a server on it.
--
Joshua Tolley / eggyknap
End Point Corporation
http://www.endpoint.com