[Linux-HA] Recovery after a failure
achapela.rexistros at gmail.com
Tue Jul 3 11:02:00 MDT 2007
I'm testing Heartbeat 2 to make a MySQL master/slave system to takeover
a failure. Now I can test a service, test the network connectivity, etc
and all runs OK.
At this moment, I want the next thing:
Server A execute MySQL resource and IPADDR (the VIP shared address)
Server B execute MySQL in slave mode
When a failure happens in the server A (lost network connectivity,
resource MySQL is lost...) the VIP and the MySQL resource must be moved
to the server B, then if Server A goes up the resource can't go again to
the server A ( I don't want it goes). I want to supervise this action
and I will make it manually.
My problem is to write a rule and to execute the resource in the two
nodes. I execute resources in two nodes as a clone statement. This is my
config for the resource:
<primitive id="mysqld-child" provider="heartbeat" class="lsb"
<op id="mysqld-child-monitor" name="monitor" interval="20s"
<op id="mysqld-child-start" name="start" prereq="nothing"/>
(This is an example.. and could change in the future...)
And my second problem is How can I write a rule to don't run a resource
in a recovery node, even if it's "master" node ??
More information about the Linux-HA