Server - Cluster - VMware Fence Device

[MarcusCaepio]

Hi all,

I'm new to this forum and need urgently help with a problem:
So, I'm trying to build an active/active cluster with CentOS 7 (VMware), which uses a shared storage for websites.
Currently I got 2 nodes. I followed the tutorial on http://www.thefallenphoenix.net/2015/02 ... -centos-7/ to build the pacemaker cluster, which works so far:

Code: Select all

[root@gfs01 ~]# pcs status
Cluster name: gfs_cluster
WARNING: no stonith devices and stonith-enabled is not false
Last updated: Tue May  5 13:55:22 2015
Last change: Tue May  5 13:55:21 2015
Stack: corosync
Current DC: gfs02 (2) - partition with quorum
Version: 1.1.12-a14efad
2 Nodes configured
0 Resources configured


Online: [ gfs01 gfs02 ]

Full list of resources:


PCSD Status:
  gfs01: Online
  gfs02: Online

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

Now, I want to install a GFS2 filesystem on a LUN, which is mounted via VMware to both nodes as /dev/sdb
I followed another tutorial, seen on http://www.thefallenphoenix.net/2015/02 ... stered-fs/ and here begin the problems:

When I want to create the resources for DLM and CLVMD, after creating, the resources are stopped and I don't know how to get them running.

Code: Select all

[root@gfs2-node1 ~]# pcs resource create res_dlm ocf:pacemaker:controld op monitor interval=3s on-fail=fence clone interleave=true ordered=true
[root@gfs2-node1 ~]# pcs resource create res_clvmd ocf:heartbeat:clvm op monitor interval=3s on-fail=fence clone interleave=true ordered=true

[root@gfs01 ~]# pcs status resources
 Clone Set: res_dlm-clone [res_dlm]
     Stopped: [ gfs01 gfs02 ]
 Clone Set: res_clvmd-clone [res_clvmd]
     Stopped: [ gfs01 gfs02 ]

I think, the error is, that I don't have any fencing devices configured yet. But because I am in VMware, I have to use fence-vmware-soap and I don't get it configured.

Can you guys tell, how to configure the fence device configured and get the resources running?

Code: Select all

[root@gfs01 ~]# pcs stonith create stonith01 fence_vmware_soap
Error: missing required option(s): 'port, ipaddr, login' for resource type: stonith:fence_vmware_soap (use --force to override)

Which port, IP and login parameters are needed???


Kind regards,

Marcus

[aks]

I think that talks to the ESX servers.
Anyway I couldn't find the docs for RHEL 7, but here's the docs for 6: https://access.redhat.com/documentation ... ap-CA.html at least it might give you a clue...

[TrevorH]

The port, IP and login parameters are the ones that give you access to the SOAP interface of your VMWare host machine.

[MarcusCaepio]

I think that talks to the ESX servers.
Anyway I couldn't find the docs for RHEL 7, but here's the docs for 6: https://access.redhat.com/documentation ... ap-CA.html at least it might give you a clue...

Thanks for the link. I guess, between RH6 and 7 are a lots of differences in the cluster technology.

The port, IP and login parameters are the ones that give you access to the SOAP interface of your VMWare host machine.

Thanks for the tipp, I'll check this.
Is it necessary to have fencing devices configured to get the resources DLM and CLVM started?
And if so, are there any other types of fencing I can use for the VMs to get the resources started?

[aks]

Yes fencing is required for STONITH (shoot the other node in the head) - a fundamental part of this type of clustering. I suppose you could do it over the network, but you would really need both network and disk heartbeats (just in case something goes wrong with the network or the disk).