andrew at beekhof
May 6, 2012, 10:04 PM
Post #6 of 6
(530 views)
Permalink
|
On Tue, May 1, 2012 at 8:38 PM, Mark Busuttil <markbusu [at] gmail> wrote:
> Thank you all for your help, I have isolated the issue and it seems that its
> working correctly, it was miss configuration from my end, with the help of
> "crm configure property no-quorum-policy="ignore"
>
> We are nearly working as expected, however could someone confirm if the
> following assumption is correct:
>
> Node1 is running a resource and fails
> Node 2 starts running the resource
> Node 2 fails, the resource would not attempt to move onto Node 1, unless i
> executed:
>
> crm resource cleanup haproxy
Depends which action fails, how it fails (return code) and what the
operation's on-fail attribute is set to.
http://clusterlabs.org/doc/en-US/Pacemaker/1.1/html/Pacemaker_Explained/s-ocf-return-codes.html
has some details on the return code aspects.
>
> Thanks in advance!
>
> On Mon, Apr 30, 2012 at 9:51 PM, Jake Smith <jsmith [at] argotec> wrote:
>>
>> What do you mean the cluster IP remains?
>>
>> No errors in the logs?
>>
>> Is the haproxy LSB compliant? see
>>
>> http://www.clusterlabs.org/doc/en-US/Pacemaker/1.1/html/Pacemaker_Explained/ap-lsb.html
>>
>> Output of crm_mon -fr1?
>>
>> Jake
>>
>>
>>
>>
>> ________________________________
>>
>> From: "Mark Busuttil" <markbusu [at] gmail>
>> To: "Jake Smith" <jsmith [at] argotec>, pacemaker [at] oss
>> Sent: Monday, April 30, 2012 3:44:55 PM
>> Subject: Re: [Pacemaker] HAproxy resource not moving
>>
>>
>> Hmm unfortunately seems to be the same issue... the cluster IP seems to
>> remain... however the ha proxy resource just goes offline and does not
>> move.. thanks
>>
>> On Apr 30, 2012 8:16 PM, "Jake Smith" <jsmith [at] argotec> wrote:
>>>
>>>
>>> ----- Original Message -----
>>>
>>> > From: "Mark Busuttil" <markbusu [at] gmail>
>>> > To: pacemaker [at] oss
>>> > Sent: Monday, April 30, 2012 2:06:42 PM
>>> > Subject: [Pacemaker] HAproxy resource not moving
>>>
>>> > Hey guys,
>>>
>>> > I recently started testing out Pacemaker + Corosync for a clustering
>>> > solution for HA proxy. My configuration is as follows
>>>
>>> > crm(live)configure# show
>>> > node node1.local
>>> > node node2.local
>>> > primitive ClusterIP ocf:heartbeat:IPaddr2 \
>>> > params ip="********" cidr_netmask="32" \
>>> > op monitor interval="30s"
>>> > primitive test lsb:haproxy \
>>> > op monitor interval="10s"
>>> > colocation haproxy-with-public-IPs inf: test ClusterIP
>>> > order haproxy-after-IP inf: ClusterIP test
>>> > property $id="cib-bootstrap-options" \
>>> > dc-version="1.1.6-3.el6-a02c0f19a00c1eb2527ad38f146ebc0834814558" \
>>> > cluster-infrastructure="openais" \
>>> > expected-quorum-votes="2" \
>>> > stonith-enabled="false"
>>>
>>> Might be because you've lost quorum and the cluster has no directions
>>> without quorum... this might fix:
>>>
>>> crm configure property no-quorum-policy="ignore"
>>>
>>>
>>> HTH
>>>
>>> Jake
>>>
>>> > If both my nodes are up, the VIP responds as requested to the HA
>>> > proxy requests.. and seems to always stick to node2.
>>>
>>> > Online: [node2.local node1.local ]
>>>
>>> > ClusterIP (ocf::heartbeat:IPaddr2): Started node2.local
>>> > test (lsb:haproxy): Started node2.local
>>>
>>> > However, if i break HA proxy on node 2 or reboot that machine, the
>>> > cluster dies and doest move over to node 1 and the crm_mon displays:
>>>
>>> > Online: [node1.local ]
>>> > OFFLINE: [ node2.local ]
>>>
>>> > Not sure what im missing here, however i would assume that both
>>> > resources should move over to Node1..
>>>
>>> > Thanks in advance for any help you can provide. --
>>> > Regards,
>>> > Mark Busuttil
>>>
>>> > _______________________________________________
>>> > Pacemaker mailing list: Pacemaker [at] oss
>>> > 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://bugs.clusterlabs.org
>>>
>>> _______________________________________________
>>> Pacemaker mailing list: Pacemaker [at] oss
>>> 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://bugs.clusterlabs.org
>>
>>
>>
>> _______________________________________________
>> Pacemaker mailing list: Pacemaker [at] oss
>> 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://bugs.clusterlabs.org
>>
>
>
>
> --
> Regards,
> Mark Busuttil
>
> _______________________________________________
> Pacemaker mailing list: Pacemaker [at] oss
> 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://bugs.clusterlabs.org
>
_______________________________________________
Pacemaker mailing list: Pacemaker [at] oss
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://bugs.clusterlabs.org
|
