Login | Register For Free | Help
Search for: (Advanced)

Mailing List Archive: Linux-HA: Users

N+N Cluster

 

 

Linux-HA users RSS feed   Index | Next | Previous | View Threaded


colin.hch at gmail

Oct 2, 2009, 4:05 AM

Post #1 of 7 (1008 views)
Permalink
N+N Cluster

Hallo,

is there any experience on using Heartbeat with a N+N cluster, i.e. a
cluster with more than 2 nodes split over two fire-areas?

As far as I can gather from the documentation, this setup has the
problem that if one fire-area is offline, then the other fire-area
will either stop working immediately, or as soon as one of its
computers has a problem (I read somewhere in the documentation that >
50% are needed in a group of nodes to carry on, however in a two-node
setup that would not work...?) which somehow defeats the intention of
a fire-area to improve availability...

Thanks, Colin
_______________________________________________
Linux-HA mailing list
Linux-HA [at] lists
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems


misch at multinet

Oct 2, 2009, 4:14 AM

Post #2 of 7 (934 views)
Permalink
Re: N+N Cluster [In reply to]

Am Freitag, 2. Oktober 2009 13:05:42 schrieb Colin:
> Hallo,
>
> is there any experience on using Heartbeat with a N+N cluster, i.e. a
> cluster with more than 2 nodes split over two fire-areas?
>
> As far as I can gather from the documentation, this setup has the
> problem that if one fire-area is offline, then the other fire-area
> will either stop working immediately, or as soon as one of its
> computers has a problem (I read somewhere in the documentation that >
> 50% are needed in a group of nodes to carry on, however in a two-node
> setup that would not work...?) which somehow defeats the intention of
> a fire-area to improve availability...
>
> Thanks, Colin

It depends on your setup. heartbeat doesn't care about quorum if there are
only two nodes in the cluster. If you have more nodes you can disable this
feature and and/or use stonith.

Anyway: I would advise you not to use heartbeat any more, but to upgrade to
pacemaker.

--
Dr. Michael Schwartzkopff
MultiNET Services GmbH
Addresse: Bretonischer Ring 7; 85630 Grasbrunn; Germany
Tel: +49 - 89 - 45 69 11 0
Fax: +49 - 89 - 45 69 11 21
mob: +49 - 174 - 343 28 75

mail: misch [at] multinet
web: www.multinet.de

Sitz der Gesellschaft: 85630 Grasbrunn
Registergericht: Amtsgericht München HRB 114375
Geschäftsführer: Günter Jurgeneit, Hubert Martens

---

PGP Fingerprint: F919 3919 FF12 ED5A 2801 DEA6 AA77 57A4 EDD8 979B
Skype: misch42
_______________________________________________
Linux-HA mailing list
Linux-HA [at] lists
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems


colin.hch at gmail

Oct 2, 2009, 5:00 AM

Post #3 of 7 (939 views)
Permalink
Re: N+N Cluster [In reply to]

On Fri, Oct 2, 2009 at 1:14 PM, Michael Schwartzkopff <misch [at] multinet> wrote:
> Am Freitag, 2. Oktober 2009 13:05:42 schrieb Colin:
>>
>> is there any experience on using Heartbeat with a N+N cluster, i.e. a
>> cluster with more than 2 nodes split over two fire-areas?
>>
>> As far as I can gather from the documentation, this setup has the
>> problem that if one fire-area is offline, then the other fire-area
>> will either stop working immediately, or as soon as one of its
>> computers has a problem (I read somewhere in the documentation that >
>> 50% are needed in a group of nodes to carry on, however in a two-node
>> setup that would not work...?) which somehow defeats the intention of
>> a fire-area to improve availability...
>>
> It depends on your setup. heartbeat doesn't care about quorum if there are
> only two nodes in the cluster. If you have more nodes you can disable this
> feature and and/or use stonith.

Is disabling the quorum a binary decision, or can I choose the percentage?

> Anyway: I would advise you not to use heartbeat any more, but to upgrade to
> pacemaker.

We want to use pacemaker, however pacemaker in turn uses heartbeat (?) ...

... except that pacemaker can also use openais instead of heartbeat,
is that what you mean?

Regards, Colin
_______________________________________________
Linux-HA mailing list
Linux-HA [at] lists
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems


misch at multinet

Oct 2, 2009, 5:08 AM

Post #4 of 7 (936 views)
Permalink
Re: N+N Cluster [In reply to]

Am Freitag, 2. Oktober 2009 14:00:20 schrieb Colin:
> On Fri, Oct 2, 2009 at 1:14 PM, Michael Schwartzkopff <misch [at] multinet>
wrote:
> > Am Freitag, 2. Oktober 2009 13:05:42 schrieb Colin:
> >> is there any experience on using Heartbeat with a N+N cluster, i.e. a
> >> cluster with more than 2 nodes split over two fire-areas?
> >>
> >> As far as I can gather from the documentation, this setup has the
> >> problem that if one fire-area is offline, then the other fire-area
> >> will either stop working immediately, or as soon as one of its
> >> computers has a problem (I read somewhere in the documentation that >
> >> 50% are needed in a group of nodes to carry on, however in a two-node
> >> setup that would not work...?) which somehow defeats the intention of
> >> a fire-area to improve availability...
> >
> > It depends on your setup. heartbeat doesn't care about quorum if there
> > are only two nodes in the cluster. If you have more nodes you can disable
> > this feature and and/or use stonith.
>
> Is disabling the quorum a binary decision, or can I choose the percentage?

no-quorum-policy=
no-quorum-policy (enum, [stop]): What to do when the cluster does not have
quorum. Allowed values: stop, freeze, ignore, suicide

> > Anyway: I would advise you not to use heartbeat any more, but to upgrade
> > to pacemaker.
>
> We want to use pacemaker, however pacemaker in turn uses heartbeat (?) ...
>
> ... except that pacemaker can also use openais instead of heartbeat,
> is that what you mean?
>
> Regards, Colin

pacemaker can use both. Teh difference is that openais is maintained.

--
Dr. Michael Schwartzkopff
MultiNET Services GmbH
Addresse: Bretonischer Ring 7; 85630 Grasbrunn; Germany
Tel: +49 - 89 - 45 69 11 0
Fax: +49 - 89 - 45 69 11 21
mob: +49 - 174 - 343 28 75

mail: misch [at] multinet
web: www.multinet.de

Sitz der Gesellschaft: 85630 Grasbrunn
Registergericht: Amtsgericht München HRB 114375
Geschäftsführer: Günter Jurgeneit, Hubert Martens

---

PGP Fingerprint: F919 3919 FF12 ED5A 2801 DEA6 AA77 57A4 EDD8 979B
Skype: misch42
_______________________________________________
Linux-HA mailing list
Linux-HA [at] lists
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems


colin.hch at gmail

Oct 2, 2009, 6:00 AM

Post #5 of 7 (937 views)
Permalink
Re: N+N Cluster [In reply to]

On Fri, Oct 2, 2009 at 2:08 PM, Michael Schwartzkopff <misch [at] multinet> wrote:
> Am Freitag, 2. Oktober 2009 14:00:20 schrieb Colin:
>> On Fri, Oct 2, 2009 at 1:14 PM, Michael Schwartzkopff <misch [at] multinet> wrote:
>
>> > Anyway: I would advise you not to use heartbeat any more, but to upgrade
>> > to pacemaker.
>>
>> We want to use pacemaker, however pacemaker in turn uses heartbeat (?) ...
>>
>> ... except that pacemaker can also use openais instead of heartbeat,
>> is that what you mean?
>
> pacemaker can use both. Teh difference is that openais is maintained.

Ok, thanks for the notice -- the heartbeat page did seem a little
outdated, however we had planned on using it due to some prior
experience (several years back).

(Now I only have to find out what OpenAIS -- and corosync? -- exactly
are and do, and what their relationship is, the documentation section
on their homepage is limited to a quickstart guide...)

Thanks, Colin
_______________________________________________
Linux-HA mailing list
Linux-HA [at] lists
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems


misch at multinet

Oct 2, 2009, 6:03 AM

Post #6 of 7 (937 views)
Permalink
Re: N+N Cluster [In reply to]

Am Freitag, 2. Oktober 2009 15:00:09 schrieb Colin:
> On Fri, Oct 2, 2009 at 2:08 PM, Michael Schwartzkopff <misch [at] multinet>
wrote:
> > Am Freitag, 2. Oktober 2009 14:00:20 schrieb Colin:
> >> On Fri, Oct 2, 2009 at 1:14 PM, Michael Schwartzkopff <misch [at] multinet>
wrote:
> >> > Anyway: I would advise you not to use heartbeat any more, but to
> >> > upgrade to pacemaker.
> >>
> >> We want to use pacemaker, however pacemaker in turn uses heartbeat (?)
> >> ...
> >>
> >> ... except that pacemaker can also use openais instead of heartbeat,
> >> is that what you mean?
> >
> > pacemaker can use both. Teh difference is that openais is maintained.
>
> Ok, thanks for the notice -- the heartbeat page did seem a little
> outdated, however we had planned on using it due to some prior
> experience (several years back).
>
> (Now I only have to find out what OpenAIS -- and corosync? -- exactly
> are and do, and what their relationship is, the documentation section
> on their homepage is limited to a quickstart guide...)

Basically they do the same as heartbeat: Provide cluster communication and
cluster membership information.

When you install pacemaker, just install openais (or corosync for debian
lenny) instead of heartbeat. Configuration is simple. Just adjust the network
configuration to your needs. And then start openais (or corosync).


--
Dr. Michael Schwartzkopff
MultiNET Services GmbH
Addresse: Bretonischer Ring 7; 85630 Grasbrunn; Germany
Tel: +49 - 89 - 45 69 11 0
Fax: +49 - 89 - 45 69 11 21
mob: +49 - 174 - 343 28 75

mail: misch [at] multinet
web: www.multinet.de

Sitz der Gesellschaft: 85630 Grasbrunn
Registergericht: Amtsgericht München HRB 114375
Geschäftsführer: Günter Jurgeneit, Hubert Martens

---

PGP Fingerprint: F919 3919 FF12 ED5A 2801 DEA6 AA77 57A4 EDD8 979B
Skype: misch42
_______________________________________________
Linux-HA mailing list
Linux-HA [at] lists
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems


andrew at beekhof

Oct 6, 2009, 7:41 AM

Post #7 of 7 (874 views)
Permalink
Re: N+N Cluster [In reply to]

On Fri, Oct 2, 2009 at 3:00 PM, Colin <colin.hch [at] gmail> wrote:
> On Fri, Oct 2, 2009 at 2:08 PM, Michael Schwartzkopff <misch [at] multinet> wrote:
>> Am Freitag, 2. Oktober 2009 14:00:20 schrieb Colin:
>>> On Fri, Oct 2, 2009 at 1:14 PM, Michael Schwartzkopff <misch [at] multinet> wrote:
>>
>>> > Anyway: I would advise you not to use heartbeat any more, but to upgrade
>>> > to pacemaker.
>>>
>>> We want to use pacemaker, however pacemaker in turn uses heartbeat (?) ...
>>>
>>> ... except that pacemaker can also use openais instead of heartbeat,
>>> is that what you mean?
>>
>> pacemaker can use both. Teh difference is that openais is maintained.
>
> Ok, thanks for the notice -- the heartbeat page did seem a little
> outdated, however we had planned on using it due to some prior
> experience (several years back).
>
> (Now I only have to find out what OpenAIS -- and corosync? -- exactly
> are and do, and what their relationship is, the documentation section
> on their homepage is limited to a quickstart guide...)

I'd hold off on corosync for a bit, seems there were some regressions
from the previous stable series that prevent it from playing nicely
with Pacemaker.

You might find the following useful though, its for Fedora 11 but the
openais parts are pretty generic.
http://www.clusterlabs.org/mediawiki/images/9/9d/Clusters_from_Scratch_-_Apache_on_Fedora11.pdf
_______________________________________________
Linux-HA mailing list
Linux-HA [at] lists
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems

Linux-HA users RSS feed   Index | Next | Previous | View Threaded
 
 


Interested in having your list archived? Contact Gossamer Threads
 
  Web Applications & Managed Hosting Powered by Gossamer Threads Inc.