[Ocfs2-users] lvm2 not cluster aware - okay, so how should Istripe my LUNs?

[Alexei_Roudnev]

It can be correct, but EVMS (which supports clustering) != lvm2 .

But I believe, that adding too many new components into existing Linux (such
as RHEL4 + GFS + EVMS + heartBeat + GLM)
means a very high possibility to catch a bug (and so have very low
reliability). Remember, that you need monthes to test such things - even if
yiu have a very good test team, you can't model all failures and real life
situations until you run it somewhere for a long.

Novell tested (show on last LinuxWorld) 4 node system with

  SLES10
    heartbeat2
    evms
    ocfsv2 using heartbeat2

and it looks reasonable minimum (except evms, which is too complicated vs
lvm2). If they (Novell) really test such thing (testing performance,
reliability, failure recovery, disk resizing and so on), it can be a very
good thing.

For now, I don't see any reliable solution except

(Stable linux, such as SLES9 S3 + updates, or possible - RHEL4 + last
updates)
  ASM (on raw devices with 1 partition)
  OCFSv2 only for complimentary FS (arch + backups) and only if you don't
need 100% reliability.
                If you need better reliability, use NFS.
  heartbeat (if required, and better not with Oracle RAC because they will
fence each other in some cases).

All other combinations looks suspicious and untested.



----- Original Message ----- 
From: "Fabio Corazza" <fabio at newbay.com>
To: "Brian Long" <brilong at cisco.com>
Cc: <ocfs2-users at oss.oracle.com>; "Humpherys Philip D Contr 532 CMMXS/CL"
<Phil.Humpherys at HILL.af.mil>
Sent: Friday, October 06, 2006 4:29 AM
Subject: Re: [Ocfs2-users] lvm2 not cluster aware - okay, so how should
Istripe my LUNs?


> Brian Long wrote:
> > I don't believe anything included in RHEL is supported across clusters
> > for software RAID.  This means LVM and MD are out of the picture.
>
> This is incorrect, I'm running EVMS 2.5.5 with Heartbeat 2.0.7 under
> RHEL4.4 and it works pretty well. I've created filters under EVMS to
> manage just the emcpower* devices, created a cluster container, and
> created LVM2/EVMS volumes on top of it. In this way the EVMS devices are
> accessible from both the nodes.
>
> > Oracle only supports RAW devices, OCFS2 filesystems (one filesystem per
> > SAN LUN) or ASM.  They also support Red Hat GFS, but only if you use
> > external locking (GLM).  For a 4-node RAC cluster, you'd need 7 nodes (3
> > dedicated to locking).  This is because Oracle insists that none of the
> > RAC nodes participate in the GLM locking.  I believe Oracle may be
> > certifying DLM, the newer lock manager in GFS, but I'm not sure.
>
> Some people is using OCFS2 for general purposes, so any implementation
> that uses RAC's components is out of the picture.
>
> > Either way, GFS is an added cost on top of RHEL, but it includes RHCS
> > and RHCS has cluster-aware LVM (I believe).
>
> If possible, I suggest to stay away from RHCS.
>
>
> -- 
> Fabio Corazza - Engineering
> NewBay Software, Ltd.
> Wilson House, Fenian Street, Dublin 2, Ireland
> Phone: +353 1 634 5490 - e-mail: fabio at newbay.com
>
> _______________________________________________
> Ocfs2-users mailing list
> Ocfs2-users at oss.oracle.com
> http://oss.oracle.com/mailman/listinfo/ocfs2-users
>