[Linux-HA] Master-slave, stopping a slave.
dk at in-telegence.net
Wed Jul 8 01:07:41 MDT 2009
c smith wrote:
> I currently implement DRBD with Pacemaker. The DRBD resource is configured
> as a multi-state Master-slave resource in which node1 is the default master
> and node2 is the default slave. I am putting together a backup system that
> will run some automated scheduled tasks on node2 (assuming both nodes are
> online). Such tasks include disconnecting the DRBD resource temporarily and
> promoting it to primary/standalone. I'm wondering, is there a way to
> manually stop the child clone on that node before starting the backup, then
> restarting it after? Something along the lines of `crm resource ms-drbd:1
> I have yet to test the system as it is but, from what I understand, if I
> manually disconnect the resource, Pacemaker's monitor functions will not be
> happy and likely throw monitor errors and/or try to restart and reconnect
> the resource. The Pacemaker's user guide does not go into detail regarding
> managing multi-state child resources. Is it possible to do without stopping
> the entire m-s resource?
It has been stated that is it not intended to mess with the internals of
child-management. Never to use child-instance-ids in constraints and so
on. They will just cause you unwanted trouble.
I'd probably just unmanage the drbd resource, then the cluster will not
monitor it. Then do whatever you need to do and afterwards, let the
cluster manage drbd again.
Look into the "is-managed" meta attribute.
More information about the Linux-HA