ZCS in RHCS

[IPMHanwei]

Hi All,

I have installed ZCS cluster in xen environment.
Have installed it in both mail1 and mail2 server to run on RHCS.

We are not able to start the cluster service and the owner should be mail1.
But it is left blank.

Any idea how should we start to troubleshoot?

We are totally new to clustering.
================================================
[root@mail1 ~]# clustat
Member Status: Quorate

Member Name ID Status
------ ---- ---- ------
mail1.ipmirror.com 1 Online, Local, rgmanager
mail2.ipmirror.com 2 Online, rgmanager

Service Name Owner (Last) State
------- ---- ----- ------ -----
service:cluster.ipmi (none) stopped
================================================
[root@mail2 ~]# clustat
Member Status: Quorate

Member Name ID Status
------ ---- ---- ------
mail1.ipmirror.com 1 Online, rgmanager
mail2.ipmirror.com 2 Online, Local, rgmanager

Service Name Owner (Last) State
------- ---- ----- ------ -----
service:cluster.ipmi (none) stopped

[uxbod]

Welcome to the forums

What version of ZCS and Redhat ?

[IPMHanwei]

Hi,

We are using ZCS 5.0.4 in RHEL 5(64bits).

Kindly assist.
Thanks

[phoenix]

Originally Posted by IPMHanwei

We are using ZCS 5.0.4 in RHEL 5(64bits).

RHCS5 isn't currently supported, only RHCS4 is at the moment. Post the output of the following command:

zmcontrol -v

also update your profile with that information.

[IPMHanwei]

HI,

[zimbra@mail1 root]$ zmcontrol -v


Release 5.0.4_GA_2101.RHEL5_64_20080321141727 RHEL5_64 NETWORK edition

We are setting up the clustering use a GUI name LUCI.

Thanks

[phoenix]

This is not yet a supported platform for Network Edition. You must install it on a RHCS4 or wait for the resolution to this bug: Bug 18197 - RHEL5 support for cluster/RHCS otherwise you will not receive support for this configuration.

[Rich Graves]

If you are determined to use Xen rather than bare hardware or VMWare, then you might be better off with the unsupported workarounds than RHEL4.

Also consider whether you really need clustering. After some messing around with RHEL4 clustering, I decided that pager and manual failover works for me. All of our (few) outages have been due to user error or software bugs, mostly due to attachment indexing. In every instance that comes to mind, had clustering software determined that the service was unresponsive and failed over to the other node, it likely would have made the problem worse.

[IPMHanwei]

Hi,

For the time being, i would prefer to go for the manual failover.
We need to store our data in a SAN storage.

Which are the path that we need to mount to the SAN storage?
/opt/zimbra
/opt/zimbra-cluster/mountpoints/cluster.ipmirror.com

How do we softlink a existing directory to the san storage?

Thanks