[Openais] First node reboots when 2nd node joins cluster

[Andrew Beekhof]

On Wed, Sep 30, 2009 at 2:36 AM, Steven Dake <sdake at redhat.com> wrote:
> On Tue, 2009-09-29 at 15:37 -0400, Errol Neal wrote:
>> Hi. I'm a new openais user. I followed:
>>
>> http://www.clusterlabs.org/mediawiki/images/9/9d/Clusters_from_Scratch_-_Apache_on_Fedora11.pdf
>>
>> And I was very successful in getting a cluster setup. I'm having a few issues that I'm trying to sort out and I was wondering if I could get some help.
>>
>> Here is my configuration:
>>
>> node axigen1
>> node axigen2
>> primitive axigenFS ocf:heartbeat:Filesystem \
>>         params device="/dev/vgAxigenMailStore01/lvAxigenMailStore01" directory="/var/opt/axigen" fstype="ocfs2" \
>>         meta target-role="Started"
>> primitive axigenFilters lsb:axigenfilters \
>>         op monitor interval="120s" timeout="30s" \
>>         meta migration-threshold="2"
>> primitive axigenIP ocf:heartbeat:IPaddr2 \
>>         params ip="*.*.*.*" cidr_netmask="23" nic="eth0" \
>>         op monitor interval="21s" timeout="5s" \
>>         meta migration-threshold="2"
>> primitive axigenServer lsb:axigen \
>>         op monitor interval="120s" timeout="30s" \
>>         meta migration-threshold="2"
>> primitive dlm ocf:pacemaker:controld \
>>         op monitor interval="120s"
>> primitive o2cb ocf:ocfs2:o2cb \
>>         op monitor interval="120s" group axigenGroup axigenServer axigenFilters clone axigenFS-clone axigenFS \
>>         meta interleave="true"
>> group axigenGroup axigenFilters axigenServer
>> clone axigenFS-clone axigenFS
>> clone dlm-clone dlm \
>>         meta interleave="true"
>> clone o2cb-clone o2cb \
>>         meta interleave="true"
>> colocation axigenFS-with-o2cb inf: axigenFS-clone o2cb-clone
>> colocation axigenGroup-with-axigenFS inf: axigenGroup axigenFS-clone
>> colocation axigenGroup-with-axigenIP inf: axigenGroup axigenIP
>> colocation o2cb-with-dlm inf: o2cb-clone dlm-clone
>> order start-axigenGroup-after-axigenFS inf: axigenFS-clone axigenGroup
>> order start-axigenGroup-after-axigenIP inf: axigenIP axigenGroup
>> order start-o2cb-after-dlm inf: dlm-clone o2cb-clone
>> property $id="cib-bootstrap-options" \
>>         dc-version="1.0.5-462f1569a43740667daf7b0f6b521742e9eb8fa7" \
>>         cluster-infrastructure="openais" \
>>         expected-quorum-votes="2" \
>>         stonith-enabled="false" \
>>         cluster-delay="60s" \
>>         last-lrm-refresh="1254246905" \
>>         no-quorum-policy="ignore"
>> rsc_defaults $id="rsc-options" \
>>         resource-stickiness="100"
>>
>>
>> I'm running a mailsystem on a two node cluster backed by an lvm'ed ocfs2 filesystem. I recompiled lvm2 v 2.0.49 to support openais as the clvmd type.
>>
>> When I restart the second node in this cluster, the first one reboots. Is there something in my configuration that says for this to happen?
>>
>> Thanks,
>>
>
> Errol,
>
> I don't know enough about pacemaker's configuration to tell you how to
> proceed.  openais itself doesn't reboot nodes.  That is a function of
> "stonith" which is a component of pacemaker.

No stonith agents are defined either, so its not pacemaker that's
causing the resets.
Possibly OCFS2?  Perhaps its using its internal stack.