[ClusterLabs] Antw: After reboot each node thinks the other is offline.

[Greg Woods]

On Tue, Aug 1, 2017 at 2:05 AM, Stephen Carville (HA List) <
62d2a7ca at opayq.com> wrote:

> On 07/31/2017 11:13 PM, Ulrich Windl [Masked] wrote:
>
>  I guess you have no fencing configured, right?
>
> No. I didn't realize it was necessary unless there was shared storage
> involved.  I guess it is time to go back to the drawing board.  Can
> clustering even be done reliably on CentOS 6?


Yes, it can. I have a number of CentOS 6 clusters running with corosync and
pacemaker, and CentOS 6, while obviously not the latest version, is still
maintained and will be for at least a couple more years. But yes, you have
to have fencing to have a cluster. I believe there is a way to manually
tell one node of the cluster that the other node has been reset (using
stonith_admin I think), but without fencing you are likely to end up in the
state where you have to manually reset things to get the cluster going
again any time something goes wrong, which is not exactly the high
availability that  you build a cluster for in the first place.

--Greg
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.clusterlabs.org/pipermail/users/attachments/20170801/972b6dac/attachment-0003.html>