[ZODB-Dev] FileStorage backup from snapshot-capable block-device?

[Morten W. Petersen]

--Sw7tCqrGA+HQ0/zt
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Mon, Feb 18, 2002 at 03:14:09PM -0500, Chris McDonough wrote:
> > If changes are made to the Data.fs file (i.e. changes are committed)
> > while the Data.fs file is 'slowly' being copied, could the copied
> > Data.fs get corrupted?  (This is on a Linux box).
>=20
> It's almost nondeterministic as far as I know.  rsync evidently has probl=
ems
> with the source file changing.  Whether other programs like your backup
> program have problems is unknown.  You should run the "fstest.py" program
> that Jeremy created lately after you back up your Data.fs to make sure the
> copy is not corrupted.  That will also probably give you your answer.

OK, I've talked to a bit of an OS guru here, and he says files can be
write-locked on Linux, but that the database process wouldn't be able to
write to the file.

So, to make sure that the backup of the Data.fs file isn't corrupted,
one could either copy the file and test it using "fstest.py" or stop
the database process and then copy the file.

-Morten

--Sw7tCqrGA+HQ0/zt
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE8cXLaWAD3rkjzkAoRAv+lAJ4sxYf15QuAXBULt8AQeeunuRIbuACfUdfw
a6dNwuySTdiqnjjNV174MDM=
=n94m
-----END PGP SIGNATURE-----

--Sw7tCqrGA+HQ0/zt--