Question : About the use of the STONITH module of glue in pcs.

[HideoYamauchi]

Hi All,

We installed pcsd0.9.148 on RHEL7.2.
In addition, we install glue component.

• http://hg.linux-ha.org/glue/

Is the external module included in glue available as STONITH module in pcs?
In addition, how becomes the command-line of create when we use glue?

Best Regards,
Hideo Yamauchi.

[tomjelinek]

Hi Hideo,

If the external module implements the same interface as fence agents from https://github.com/ClusterLabs/fence-agents project then it should work.

Hope that helps.

[HideoYamauchi]

Hi Tomas,

Furthermore, let me confirm it.

glue is installed in the next directory.
[root@rh72-01 ~]# ls /usr/lib64/stonith/plugins/external/
drac5 dracmc-telnet hetzner hmchttp ibmrsa ibmrsa-telnet ipmi ippower9258 kdumpcheck libvirt nut rackpdu riloe ssh vcenter vmware xen0 xen0-ha

The list command does not display a module of glue, is this a specification?

[root@rh72-01 ~]# pcs stonith list
fence_alom - Fence agent for Sun ALOM
fence_amt - Fence agent for AMT
(snip)
fence_zvmip - Fence agent for use with z/VM Virtual Machines

The create command does not make STONITH with a module of glue when I do not appoint force option.
Is this a specification, too?

[root@rh72-01 ~]# pcs stonith create test external/ssh params pcmk_reboot_timeout=300 pcmk_reboot_retries=1 hostlist="rh72-02"
Error: Unable to create resource 'stonith:external/ssh', it is not installed on this system (use --force to override)

[root@rh72-01 ~]# pcs stonith create test external/ssh params pcmk_reboot_timeout=300 pcmk_reboot_retries=1 hostlist="rh72-02" --force

[root@rh72-01 ~]# pcs status
Cluster name: cluster1
Last updated: Thu Feb 25 08:08:29 2016 Last change: Thu Feb 25 08:07:50 2016 by root via cibadmin on rh72-01
Stack: corosync
Current DC: rh72-02 (version 1.1.14-70404b0) - partition with quorum
2 nodes and 1 resource configured

Online: [ rh72-01 rh72-02 ]

Full list of resources:

test (stonith:external/ssh): Started rh72-01

PCSD Status:
rh72-01: Online
rh72-02: Online

Daemon Status:
corosync: active/disabled
pacemaker: active/disabled
pcsd: active/disabled

Best Regards,
Hideo Yamauchi.

[tomjelinek]

Hi Hideo,

I am sorry, pcs does not support external STONITH agents currently. We may take a look at this in future.

Best Regards,
Tomas

[HideoYamauchi]

Hi Tomas,

Okay!

We do not use the STONITH module of glue in pcs.
We wait for future support of pcs.

Many thanks,
Hideo Yamauchi.

[HideoYamauchi]

Because a discussion was over, this is closed.

[tomjelinek]

Reopening the issue so we can track the request for adding external stonith agents support to pcs.

[HideoYamauchi]

Hi Tomas,

Sorry...
Okay!

Best Regards,
Hideo Yamauchi.

[moonracer]

Hi,
I am glad I found this issue, although not glad that the issue itself exists ;)
I am currently building a Dual Primary setup on CentOS 7.2 with pcs where I need to use external agent for STONITH (external/hetzner).
Since pcs doesn't support external STONITH agents a Dual Primary Setup is not an option anymore.
It would be really great to have a possibility of using external STONITH agents in pcs.
Or can you see other way to realize STONITH and not using external/hetzner?
Best Regards,
Marek

[tomjelinek]

Hi @moonracer,

You can try creating your STONITH resources in pcs with --force flag. Since pcs is merely a configuration tool and pacemaker is the one doing the heavy lifting here, it might work for you.

Best Regards,
Tomas

[tomjelinek]

Hi @HideoYamauchi and @moonracer,

I just pushed a commit 491ac08 which overhauls pretty much all code dealing with resource and stonith agents. With this patch pcs gets all information about the agents from pacemaker. This means pcs now supports the same set of agents as pacemaker does, including external and cluster-glue agents.

Best Regards,
Tomas

[HideoYamauchi]

Hi Tomas,

Great!!
I confirm your correction.

Many thanks!
Hideo Yamauchi.

[HideoYamauchi]

Hi Tomas,

Even if stonith module of glue did not appoint "force" option, I confirmed that I could set pcs command.

[root@rh72-01-pcsd ~]# pcs stonith create stonith1  external/ssh params pcmk_reboot_timeout=300 pcmk_reboot_retries=1 hostlist="rh72-01"

[root@rh72-01-pcsd ~]# pcs status
Cluster name: test
WARNING: corosync and pacemaker node names do not match (IPs used in setup?)
Stack: corosync
Current DC: rh72-01-pcsd (version 1.1.15-e174ec8) - partition with quorum
Last updated: Wed Oct 12 11:19:57 2016          
Last change: Wed Oct 12 11:19:46 2016 by root via cibadmin on rh72-01-pcsd

1 node and 2 resources configured

Online: [ rh72-01-pcsd ]

Full list of resources:

 test   (ocf::heartbeat:Dummy): Started rh72-01-pcsd
 stonith1       (stonith:external/ssh): Started rh72-01-pcsd

(snip)

Cheers,
Hideo Yaamuchi.

[tomjelinek]

Hi Hideo,

Thank you for testing this in such a short time!

I'm closing this issue because it has been successfully resolved.

[HideoYamauchi]

thanks!