[Pacemaker] problem about move node from one cluster to another
Dan Frincu
dfrincu at streamwide.ro
Thu Oct 14 12:02:52 UTC 2010
Hi,
Just go to http://clusterlabs.org/rpm/ and search for the RPMs for your
OS version. If you can shut the nodes down to perform the upgrade, it
would be a lot easier, if it has to be done without downtime, that is
also possible, but requires a careful planning of what to change and
when so that you don't mess up anything and is closely related to how
the cluster is set up.
How I've done it was as follows:
- 2 node setup, primary - secondary.
- went on secondary node, put it in standby
- shutdown openais
- removed all cluster related RPMs (openais, pacemaker, heartbeat, etc.)
but not the drbd RPMs
- installed the new RPMs, used corosync instead of openais, configured
corosync's /etc/corosync/corosync.conf based on the settings of the old
openais (actually did a meld between the old openais.conf and the new
corosync.conf and saved a unified corosync.conf)
- removed everything from /var/lib/heartbeat/crm/
- started corosync on secondary node
- waited for node joing, then issued an standby on the primary,
everything migrated to the secondary
- went on the primary, removed the RPMs, removed
/var/lib/heartbeat/crm/* configured corosync, started it
Et voila, it works. BTW, awesome job done to maintain compatibility
between versions, the upgrade was truly seamless.
Regards,
Dan
jiaju liu wrote:
>
>
> Hi
>
> Thank you for your help. I want to upgrade my openais. Do I need
> to restall linux and download openais of the latest version? or
> any other simple way?Thanks:-).
>
>
> Hi,
>
> Depending on the openais version (please mention it)
> ?
> Hi Thank you for your reply my openais version is openais-0.80.5-15.1
> pacemaker version is pacemaker-1.0.5-4.1.
> I use restart but it does not work.?I found?it could not stop
> ?
> ?
> this behavior could
> happen, I've seen it as well, on openais-0.8.0. What I've done to
> fix it
> was to restart the openais process via /etc/init.d/openais
> restart. And
> then it worked, however, this was one of the reasons I updated the
> packages to the latest versions of corosync, pacemaker, etc. The
> tricky
> part was doing the migration procedure for upgrading production
> servers
> without service downtime, but that's another story.
>
> Regards,
>
> Dan
>
> jiaju liu wrote:
> >
> >
> >
> >?????
>
>
>
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL:
> <http://oss.clusterlabs.org/pipermail/pacemaker/attachments/20101013/2f55034d/attachment-0001.htm>
>
> ------------------------------
>
> Message: 4
> Date: Wed, 13 Oct 2010 14:37:51 +0300
> From: Dan Frincu <dfrincu at streamwide.ro
> <http://cn.mc157.mail.yahoo.com/mc/compose?to=dfrincu@streamwide.ro>>
> To: The Pacemaker cluster resource manager
> <pacemaker at oss.clusterlabs.org
> <http://cn.mc157.mail.yahoo.com/mc/compose?to=pacemaker@oss.clusterlabs.org>>
> Subject: Re: [Pacemaker] problem about move node from one cluster to
> another cluster
> Message-ID: <4CB59A0F.80601 at streamwide.ro
> <http://cn.mc157.mail.yahoo.com/mc/compose?to=4CB59A0F.80601@streamwide.ro>>
> Content-Type: text/plain; charset="iso-8859-1"; Format="flowed"
>
> Hi,
>
> Yes, it sometime needs to be killed manually because the process
> hangs
> and the restart operation never seems to end. Yet another reason
> to upgrade.
>
> All,
>
> Question: given the fact that this type of software usually gets
> installed on a platform once and then usually goes into service
> for many
> years, on servers where downtime should be kept to a minimum (gee,
> that's why you use a cluster :)), how does this fit the release
> schedule?
>
> I mean, there are plenty of users out there with question related to
> Heartbeat 2, openais-0.8.0, and so on and so forth, some environments
> cannot be changed lightly, others, not at all, so what is the
> response
> to "this feature doesn't work on that version of software?",
> upgrade? If
> so, at what interval (keeping in mind that you probably want the
> stable
> packages on your system)?
>
> I'm asking this because when I started working with openais, the
> latest
> version available was 0.8.0 on some SUSE repos that aren't available
> anymore.
>
> Regards,
>
> Dan
>
> jiaju liu wrote:
> > Hi,
> >
> > Depending on the openais version (please mention it)
> >
> > Hi
> > Thank you for your reply my openais version is openais-0.80.5-15.1
> > pacemaker version is pacemaker-1.0.5-4.1.
> > I use restart but it does not work. I found it could not stop
> >
> >
> > this behavior could
> > happen, I've seen it as well, on openais-0.8.0. What I've done
> to fix it
> > was to restart the openais process via /etc/init.d/openais
> restart. And
> > then it worked, however, this was one of the reasons I updated the
> > packages to the latest versions of corosync, pacemaker, etc. The
> tricky
> > part was doing the migration procedure for upgrading production
> servers
> > without service downtime, but that's another story.
> >
> >
> >
>
> --
> Dan FRINCU
> Systems Engineer
> CCNA, RHCE
> Streamwide Romania
>
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL:
> <http://oss.clusterlabs.org/pipermail/pacemaker/attachments/20101013/9573d51d/attachment-0001.htm>
>
> ------------------------------
>
> _______________________________________________
> Pacemaker mailing list
> Pacemaker at oss.clusterlabs.org
> <http://cn.mc157.mail.yahoo.com/mc/compose?to=Pacemaker@oss.clusterlabs.org>
> http://oss.clusterlabs.org/mailman/listinfo/pacemaker
>
>
> End of Pacemaker Digest, Vol 35, Issue 46
> *****************************************
>
>
>
> ------------------------------------------------------------------------
>
> _______________________________________________
> Pacemaker mailing list: Pacemaker at oss.clusterlabs.org
> http://oss.clusterlabs.org/mailman/listinfo/pacemaker
>
> Project Home: http://www.clusterlabs.org
> Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
> Bugs: http://developerbugs.linux-foundation.org/enter_bug.cgi?product=Pacemaker
>
--
Dan FRINCU
Systems Engineer
CCNA, RHCE
Streamwide Romania
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.clusterlabs.org/pipermail/pacemaker/attachments/20101014/f7d8b020/attachment-0002.htm>
More information about the Pacemaker
mailing list