On Fri, Jan 23, 2009 at 11:45, Hell, Robert <[email protected]> wrote:
> Ok, thanks I now really believe it (sorry for that).
>
> Just to be sure that I got the third node solution right:
> This would be very similar to the quorum server - quorum is calculated by the 
> three cluster nodes.
> So if one site looses network connection, the remaining two cluster nodes 
> have quorum because they see each other. But what is happening with the third 
> cluster node? He stops all running resources and kills himself (fencing is 
> not available)?

It depends on your value of no-quorum-policy.
Newer versions of Pacemaker support no-quorum-policy=suicide (although
I just fixed a bug that prevented it from working correctly so you'd
want to get the latest source).

> Would it be an option to use Xen VMs as "third cluster nodes" for multiple 
> clusters on one server?

Silly me, I should have thought of that.
Sure.  We don't care what "hardware" your cluster runs on :-)

> -----Original Message-----
> From: [email protected] 
> [mailto:[email protected]] On Behalf Of Andrew Beekhof
> Sent: Freitag, 23. Jänner 2009 09:28
> To: General Linux-HA mailing list
> Subject: Re: [Linux-HA] Two-node clusters in split-sites
>
> On Fri, Jan 23, 2009 at 09:09, Hell, Robert <[email protected]> wrote:
>> Hi,
>>
>> I'm wondering that quorum server doesn't work, because Alan Robertson 
>> mentioned it in his LinuxWorld 08 talk. Can someone proof that it isn't 
>> working?
>
> I think Lars has done this in the past.
>
>> Third node solution: I wouldn't be able to use the third machine for more 
>> than one two-node cluster, right?
>
> correct
>
>> -----Original Message-----
>> From: [email protected] 
>> [mailto:[email protected]] On Behalf Of Michael 
>> Schwartzkopff
>> Sent: Donnerstag, 22. Jänner 2009 13:18
>> To: General Linux-HA mailing list
>> Subject: Re: [Linux-HA] Two-node clusters in split-sites
>>
>> Am Donnerstag, 22. Januar 2009 12:32:57 schrieb Hell, Robert:
>>> Hi,
>>>
>>>
>>>
>>> we are planning to run multiple two-node clusters (with different
>>> functions: database cluster, samba cluster, ...) in two different sites
>>> with a fast, but unreliable network connections. All cluster nodes
>>> replicate data with drbd.
>>>
>>> I'm specially afraid of split-brain in this scenario because reliable
>>> fencing is not possible in this split-site situation.
>>>
>>>
>>>
>>> After investigating multiple cluster solutions I think Linux-HA is the
>>> only project which provides a solution for our problem. We want to use
>>> Linux-HA with one cluster node in each site and an external quorum
>>> server (as described in the Heartbeat tutorial) without fencing.
>>>
>>> Has anyone experiences with a solution like this and is this a solutions
>>> which should work, is "supported" and is worth further investigations?
>>> And last but not least: is it possible to run one quorum server for
>>> multiple clusters (as mentioned before we have multiple two-node
>>> clusters, but I only want to run one quorum server).
>>>
>>>
>>>
>>> Many thanks and kind regards,
>>>
>>> Robert
>>
>> Hi,
>>
>> as far as I know the quorum server does not work. You would have to build a
>> third node and constraints to prevent resources running on that node.
>>
>> I would put the third node into a third location.
>>
>> Comments?
>>
>> Michael.
>>
>> --
>> 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: [email protected]
>> 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
>> [email protected]
>> http://lists.linux-ha.org/mailman/listinfo/linux-ha
>> See also: http://linux-ha.org/ReportingProblems
>> _______________________________________________
>> Linux-HA mailing list
>> [email protected]
>> http://lists.linux-ha.org/mailman/listinfo/linux-ha
>> See also: http://linux-ha.org/ReportingProblems
>>
> _______________________________________________
> Linux-HA mailing list
> [email protected]
> http://lists.linux-ha.org/mailman/listinfo/linux-ha
> See also: http://linux-ha.org/ReportingProblems
> _______________________________________________
> Linux-HA mailing list
> [email protected]
> http://lists.linux-ha.org/mailman/listinfo/linux-ha
> See also: http://linux-ha.org/ReportingProblems
>
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems

Reply via email to