[Samba] ubuntu, ocfs2 with cman and ctdb

[charles]

>
> From: Jeremy Allison <jra at samba.org>
> To: charles
> Date: Mon, 20 Jun 2011 17:27:03 -0700
> Subject: Re: [Samba] ubuntu, ocfs2 with cman and ctdb
> On Mon, Jun 20, 2011 at 03:18:02PM -0600, charles wrote:
> > hi guys,
> >
> > we're evaluating the available clustering options to get ctdb up and
> running
> > for a highly available file server.
> > we've set up both gluster and ocfs2 both on seperate 2 node setups.
> > ocfs2 seems to provide better throughput and iops to samba clients than
> does
> > gluster and that is comparing a single node server to a ctdb clustered 2
> > node server.
> > problem with ocfs2 is that i've been unable to configure it to utilize
> > cman's stack  to provide proper locking for ctdb. gfs2 is up next.
> >
> > does anyone have any pointers/tutorials/document for getting ocfs2 set up
> > with cman on ubunutu?
>
> Jim Mcdonough and his team @ SuSE has done most of the work
> looking at Samba/CTDB with ocfs2.
>
> Jim, any comments ?
>
> Jeremy.
>
> ---------- Forwarded message ----------
> From: "Daniel Müller"
> To: "'Jeremy Allison'" <jra at samba.org>, "'charles'"
> Date: Tue, 21 Jun 2011 08:35:29 +0200
> Subject: Re: [Samba] ubuntu, ocfs2 with cman and ctdb
> Gfs2 on centOS is working for me with ctdb. Gluster 3.2.1 is working with
> some tricks on Centos without using ctdb on a PDC BDC situation.
> Greetings
> Daniel
>
> -----------------------------------------------
> EDV Daniel Müller
>
> Leitung EDV
> Tropenklinik Paul-Lechler-Krankenhaus
> Paul-Lechler-Str. 24
> 72076 Tübingen
>
> Tel.: 07071/206-463, Fax: 07071/206-499
> eMail: mueller at tropenklinik.de
> Internet: www.tropenklinik.de
> -----------------------------------------------
>
> ---------- Forwarded message ----------
> From: Jim McDonough <jmcd at samba.org>
> To: charles <charlesaburrell at gmail.com>
> Date: Tue, 21 Jun 2011 13:59:16 -0400
> Subject: Re: [Samba] ubuntu, ocfs2 with cman and ctdb
> On Mon, Jun 20, 2011 at 8:27 PM, Jeremy Allison <jra at samba.org> wrote:
> > On Mon, Jun 20, 2011 at 03:18:02PM -0600, charles wrote:
> >> hi guys,
> >>
> >> we're evaluating the available clustering options to get ctdb up and
> running
> >> for a highly available file server.
> >> we've set up both gluster and ocfs2 both on seperate 2 node setups.
> >> ocfs2 seems to provide better throughput and iops to samba clients than
> does
> >> gluster and that is comparing a single node server to a ctdb clustered 2
> >> node server.
> >> problem with ocfs2 is that i've been unable to configure it to utilize
> >> cman's stack  to provide proper locking for ctdb. gfs2 is up next.
> >>
> >> does anyone have any pointers/tutorials/document for getting ocfs2 set
> up
> >> with cman on ubunutu?
> >
> > Jim Mcdonough and his team @ SuSE has done most of the work
> > looking at Samba/CTDB with ocfs2.
> >
> > Jim, any comments ?
> I'm not familiar with cman.  On SUSE distributions, ocfs2 uses the
> pacemaker stack, so I can't really comment on cman.  The locking works
> properly with the pacemaker linux-ha stack and ocfs2.
>
> --
> Jim McDonough
> Samba Team
> SUSE labs
> jmcd at samba dot org
> jmcd at themcdonoughs dot org
>
>
> _______________________________________________
> samba mailing list
> samba at lists.samba.org
> https://lists.samba.org/mailman/listinfo/samba
>
> thanks for the feedback, my next steps are to give centos a shot with ocfs2
and gfs2.
i've given gfs2 a try using cman (red hat's cluster manager). performance
seems on par with gluster, ctdb works on both, more or less.

i've noticed though that windows client locking of files doe not work across
nodes.

smbstatus does not show clients connected to both nodes or locked files only
the node from which the command is run.
i've added directive in my smb.conf files with no improvement.
i also have a problem with the server hanging while use crystal diskmark to
test random 4K write tests (the test can't be completed)
i use defaults for gfs2 clustered disk:  mkfs.gfs2 -t smbhaclust:gfs01 -j 2
/dev/drbd1

appreciate any assistance from those with gluster or gfs2 experience.


gluster nodes smb.conf

   passdb backend = tdbsam
   idmap backend = tdb2
   clustering = yes
[bunches]
        read only = no
        comment = cluster disk
        path = /mnt/gluster/samba00
        guest ok = yes


gfs2 nodes smb.conf
    passdb backend = tdbsam
   idmap backend = tdb2
   clustering = yes
   use mmap = yes
   posix locking = no                         #added in attempt to fix
server hanging
   fileid:algorithm = fsname                 #added in attempt to fix
locking
   vfs objects = fileid                          #added in attempt to fix
locking
   private dir = /mnt/repclust-disk/ctdb #added in attempt to fix locking

[bunches]
        read only = no
        comment = cluster disk
        path = /mnt/repclust-disk/samba00
        guest ok = yes

thanks

-- 
Charles
Belmopan, Belize

*clubmotorsports.bz* <http://clubmotorsports.bz/>
*
*
*club motorsports belize
facebook<http://www.facebook.com/pages/Club-Motorsports/135687249774906>
*
*
*
*http://www.cardomain.com/ride/2400106*

"... we just love cars and we love driving them!"
"... Do things like double clutch automatics, traction control, and lane
departure warnings really make cars better? I suppose by some metrics they
do. But for these amenities we trade character; for luxury we surrender
sensory experience; for comfort, we give up romance."