RE: RE: [Opendlm-devel] ODLM/OGFS Recovery

--=-NyMw9k/rzG7z4XQUiEzG
Content-Type: text/plain
Content-Transfer-Encoding: quoted-printable

On Mon, 2004-05-03 at 16:20, Cahill, Ben M wrote:
> =20
> > -----Original Message-----
> > From: [email protected]=20
> > [mailto:[email protected]] On Behalf=20
> > Of Stanley Wang
> > Sent: Saturday, May 01, 2004 6:34 AM
> > To: OpenDLM-DEV; John T Devine
> > Cc: Zickus II, Don; OpenGFS-DEV
> > Subject: RE: [ogfs-dev]RE: [Opendlm-devel] ODLM/OGFS Recovery
> >=20
> > Hi John,
> >=20
> > Thanks for your great comments!
> >=20
> > On Fri, 2004-04-30 at 15:04, Devine, John T wrote:
> > > Don, locks granted due to the owner of the blocking lock=20
> > releasing them
> > > due to node failure MUST be granted with an invalidated=20
> > lock value block
> > > if the blocking mode was PW or stronger.   If this is not so, then
> > > unless dead man locks had a mechanism guaranteeing they are=20
> > granted 1st
> > > (they don't), any OpenDLM user could wind up with a=20
> > consistency problem
> > > as resources could be granted before node failure detected.
> >=20
> > It's exactly my concern for OpenGFS.
> >=20
>=20
> I think this is handled during ODLM's lock recovery process, which
> invalidates LVBs that might have been altered by a dead node.  See
> clmr_clean() and clmr_lkvlb().

Yes, ODLM will invalidate LVBs that might have been altered by a dead
node. But my concern is :=20

If there is only one PW or EX lock request in the wait queue of this
resource, we won't get DLM_VALNOTVALID (even the LVB of this resouce is
invalid) when the queued request is granted.

It means that we might miss some locks held by dead node. Please read:
rc_barrier6->wake_all_locks->convert_locks->grant_convert->valueblock

Any comments?

Best Regards,
Stan=20

> -- Ben --
>=20
> Opinions are mine, not Intel's
>=20
>=20
> -------------------------------------------------------
> This SF.Net email is sponsored by: Oracle 10g
> Get certified on the hottest thing ever to hit the market... Oracle 10g.=20
> Take an Oracle 10g class now, and we'll give you the exam FREE.
> http://ads.osdn.com/?ad_id149&alloc_id66&op=3Dclick
> _______________________________________________
> Opendlm-devel mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/opendlm-devel
--=20
Opinions expressed are those of the author and do not represent Intel
Corporation
=20
"gpg --recv-keys --keyserver wwwkeys.pgp.net E1390A7F"
{E1390A7F:3AD1 1B0C 2019 E183 0CFF  55E8 369A 8B75 E139 0A7F}

--=-NyMw9k/rzG7z4XQUiEzG
Content-Type: application/pgp-signature; name=signature.asc
Content-Description: This is a digitally signed message part

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (GNU/Linux)

iD8DBQBAl2uGNpqLdeE5Cn8RAqa1AKCEEjnywnj3snogjRreBtRbOMw5agCfWdtl
0dvwvha1ImoleBZJ5Vlg+8s=
=v/MZ
-----END PGP SIGNATURE-----

--=-NyMw9k/rzG7z4XQUiEzG--



-------------------------------------------------------
This SF.Net email is sponsored by: Oracle 10g
Get certified on the hottest thing ever to hit the market... Oracle 10g. 
Take an Oracle 10g class now, and we'll give you the exam FREE.
http://ads.osdn.com/?ad_id=3149&alloc_id=8166&op=click