[ClusterLabs] "standby reboot" - Behavior

Jorge Fábregas jorge.fabregas at gmail.com
Sun Aug 30 13:15:41 EDT 2015


Hi,

This is on SLES 11 SP4 (Pacemaker 1.1.12).

I recently found out, within the crm shell, that there's the "reboot"
option that you can use with standby.  It will put the node in standby
node until it gets rebooted (I like it).  However when I use it, there's
no indication (anywhere) that the node is indeed in standby node.

All works as expected:  resources are moved away and you can't move
resources back into it until it gets rebooted.  It's just that I
expected it to appear as "standby" before the reboot occurs.

Here's an example:

crm(live)node# list
sles11a: normal
	standby=off
sles11b: normal
	standby=off


crm(live)node# standby sles11a
sles11a: normal
	standby=on
sles11b: normal
	standby=off

ok...


crm(live)node# online sles11a
sles11a: normal
	standby=off
sles11b: normal
	standby=off


crm(live)node# standby sles11a reboot
sles11a: normal
	standby=off
sles11b: normal
	standby=off


At this point all resources are moved away from sles11a and you can't
start any on it until it gets rebooted.  I even tried:  "crm configure
edit" but I don't see any indication of a reboot pending or a flag
somewhere indicating that it's indeed in standby node.  Maybe this is
handled outside the CIB I guess.

Thanks for any feedback.

Regards,
Jorge




More information about the Users mailing list