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

Hello Stan,

We actually used this function on trucluster.  However, its main use was =
_after_ detection of a failed node and even then to revalidate the lock =
value blocks and resources assuming you are using lock value blocks.
=20
I have thinking about this whole problem and wondering what everyone's =
definition of 'persistent' lock is?  Is it expected to have every =
resource and its information replicated across the entire cluster or =
just have each node keep track of its own resource locks?

-Don


> -----Original Message-----
> From: [email protected]=20
> [mailto:[email protected]] On Behalf=20
> Of Stanley Wang
> Sent: Thursday, April 29, 2004 6:34 AM
> To: [email protected]; OpenDLM Dev Mail List
> Subject: Re: [ogfs-dev]RE: [Opendlm-devel] ODLM/OGFS Recovery
>=20
>=20
> [snip]
>=20
> >I did a little googling on DLM persistent locks and found this man=20
> >page.  It looks like it is for trucluster.  It describes persistent=20
> >locks and how recovery was done.  This one is similar to what I=20
> >remember (which makes since since persistent locks were=20
> implemented for=20
> >a large database software company and many vendors implemented what=20
> >they asked for :) )
> >
> >http://h30097.www3.hp.com/docs/cluster_doc/cluster_16/MAN/MAN
> 3/0024____
> >.HTM
> >
> >I am a bit surprised that openDLM's persistent locks do not=20
> provide the=20
> >same kind of functionality.  The whole point was to give an=20
> application=20
> >a chance to do a recovery before granting more locks without=20
> having to=20
> >keep the locks open.
> >
> >Are there any openDLM folks who can explain how these ORPHAN can be=20
> >used in a distributed application if it cannot handle node death?  I=20
> >thought node death was one of the big reasons for clusters=20
> and DLMs...
> >
> >Depending on the order of lock recovery seems to be asking=20
> for trouble.
> > =20
> >
> Sorry for late responding :)
>=20
> I finally get the url you mentioned from Google's cache, thanks a lot=20
> for your help :)
> And I'm trying to find how they store the orphan/persistent=20
> locks, but=20
> failed to get that. Do you have any ideas about that? AFAIK, it's the=20
> major barrier for current OpenDLM to support=20
> orphan/persistent lock even=20
> on node failure cases. (For worest case, the failed node holds one=20
> orphan lock, and it also acted as both directory/master for=20
> this lock.=20
> To avoid this happening, we need keep one copy for each=20
> orphan lock on=20
> each node, is that right?)
>=20
> Thanks again,
> Stan
>=20
> --=20
> Opinions expressed are those of the author and do not=20
> represent Intel Corporation "gpg --recv-keys --keyserver=20
> wwwkeys.pgp.net E1390A7F" {E1390A7F:3AD1 1B0C 2019 E183 0CFF =20
> 55E8 369A 8B75 E139 0A7F}
>=20
>=20
>=20
> -------------------------------------------------------
> This SF.Net email is sponsored by: Oracle 10g
> Get certified on the hottest thing ever to hit the market...=20
> Oracle 10g.=20
> Take an Oracle 10g class now, and we'll give you the exam=20
> FREE. http://ads.osdn.com/?ad_id=3D3149&alloc_id=3D8166&op=3Dclick
> _______________________________________________
> Opendlm-devel mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/opendlm-devel
>=20
>=20


-------------------------------------------------------
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