[ClusterLabs Developers] How to prevent pacemaker from doing a system halt - 2 Node HA with customised fencing

[Vimal Menon]

Hello,

I have a 2 Node cluster which behaves active-passive. I have created a
custom capture the flag based fencing which stores a flag and action
messages on a 3rd machine (not part of the pacemaker cluster).

One observation is that when I ungracefully reboot one of the machines (say
Machine A) which owns the resources (a web server and multiple vips), and
after 'fencing action: off' in Machine B when 'fencing action: on' is in
progress - as soon as in the boot process of  machine A's pacemaker starts
up it gets halted (hard halt) by the pacemaker system, even before
pacemaker starts up completely.

How to prevent pacemaker from performing a system halt when using a
customised fencing script?

Thanking you in advance,

Regards,
Vimal
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.clusterlabs.org/pipermail/developers/attachments/20170123/09a70d02/attachment-0002.html>