[Debian-ha-maintainers] PCS problem

Richard B Winters rik at mmogp.com
Thu Jan 28 17:26:06 UTC 2016


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

Hello,

January 28 2016 6:17 AM, "Adam Kuśmirek" <amkusmirek at gmail.com> wrote:
> Hi,
>
> I used to work with pcs tool when managing Corosync/Pacemaker clusters.
>
> I installed new Jessie/HA system and created cluster using crm tool as described in Clusters from
> Scratch; Debian style document.

So everything you installed was from ppa.mmogp.com ?


> Then I realized that pcs is available in ppa.mmogp.com repository.

And soon in unstable!


> I installed it but when trying to run pcs status i get the error with traceback:
>
> root at ubi2-manager-a:~# pcs status
>
> Cluster name:
>
> Traceback (most recent call last):
>
> File "/usr/sbin/pcs", line 155, in <module>
>
> main(sys.argv[1:])
>
> File "/usr/sbin/pcs", line 147, in main
>
> status.status_cmd(argv)
>
> File "/usr/lib/python2.7/dist-packages/pcs/status.py", line 13, in status_cmd
>
> full_status()
>
> File "/usr/lib/python2.7/dist-packages/pcs/status.py", line 53, in full_status
>
> if utils.corosyncPacemakerNodeCheck():
>
> File "/usr/lib/python2.7/dist-packages/pcs/utils.py", line 1976, in corosyncPacemakerNodeCheck
> cs_nodes = getCorosyncNodesID()
> File "/usr/lib/python2.7/dist-packages/pcs/utils.py", line 1955, in getCorosyncNodesID
> cs_nodes[node_list_node_mapping[m.group(1)]] = m.group(2)
> KeyError: '0'
>
> When I run pcs stonith show everything looks ok:
>
> root at ubi2-manager-a:~# pcs stonith show
> fence_ubi2-manager-a   (stonith:fence_virsh):  Started ubi2-manager-b
> fence_ubi2-manager-b   (stonith:fence_virsh):  Started ubi2-manager-a
>
> Can you help ?
>
> Regards
> Adam Kusmirek

If you did install everything from ppa.mmogp.com, then first and foremost you would want to kill pacemaker_remoted, I tend to do a 'ps aux | grep pacemaker', and kill the process listed.

Next I tend to have to start the cluster via:  'pcs cluster start' (on each node)

After a few moments 'pcs status' works good:

- -
# pcs status
Cluster name: mmasxha
Last updated: Thu Jan 28 12:13:09 2016
Last change: Wed Nov 25 07:03:18 2015
Stack: corosync
Current DC: mmasx.mmogp.com (1) - partition with quorum
Version: 1.1.12-561c4cf
2 Nodes configured
1 Resources configured


Online: [ mmasx.mmogp.com mmasx2.mmogp.com ]

Full list of resources:

 ClusterIP	(ocf::heartbeat:IPaddr2):	Started mmasx.mmogp.com

PCSD Status:
  mmasx.mmogp.com: Offline
  mmasx2.mmogp.com: Offline

- -


Please note that the PCSD status is offline in my paste above, though it works for me as well.



If you installed some packages not from ppa.mmogp.com then it is possible that this is your problem, as all the versions of packages in ppa.mmogp.com were created some time ago, but for each other - since then the pacemaker package has been changed, as well as corosync.  I have not tested crmsh nor pcs against the latest pacemaker/corosync packages.

It is possible that the changes in the pacemaker/corosync packages now have things in different places - and crmsh/pcs depend on certain file locations (determined at build time) in order to do some of their tasks properly.

I can check into this, but there very well might not be any issue with regard to that - in which case we can do our best to determine and help to alleviate the issue you are facing.

Please advise - and best regards,


Rik
-----BEGIN PGP SIGNATURE-----
Version: OpenPGP.js v1.2.0
Comment: http://openpgpjs.org
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=whSN
-----END PGP SIGNATURE-----



More information about the Debian-ha-maintainers mailing list