|
Vijay K <vijay_k@xxxxxxxxxxxxx> wrote:
> have a load balancing setup working fine.
> Now when the swithover happens I need another resource to be taken over
> by
> the director that is going become master.
> Here I am facing a problem,I added the resource script in
> /etc/ha.d/resource.d/. and an entry in haresources.
> After this it so happens that when I want to stop the heartbeat in the
> active master machine so that the other machine takesover I am unable to
> stop it.If I remove this entry in haresources everything is working
> perfectly fine but when I add the I am facing this problem.
> Right now my haresources file has two virtual ips, cluster name and
> ldirectord and all these services are switching over perfectly.When I
> add
> this new service which was developed by me I am facing this problem.
> Is there any special ways to be adhered to when a new resource is to be
> added.
> Can somebody advice me on how to go about adding this new resource to
> haresources and make it work.
> Here is what my haresource file looks like :
> Clustername IPaddr::10.10.30.30 IPaddr::10.10.33.33 ldirectord::www
>
> - Ignored:
> monitorfes
>
> Monitorfes is the new resource I want to be taken over and the
> /etc/ha.d/resource.d/monitorfes file looks like
>
> case $1 in
> start) pkill lbmonitor >/dev/null 2>&1
> lbmonitor CLIENT 10.10.30.36
> exit 0;
> ;;
> stop) pkill lbmonitor >/dev/null 2>&1
> lbmonitor SERVER 10.10.30.36
> exit 0;
> ;;
> esac
> exit 1
>
>
> Is there something wrong in the script. Lbmonitor is the daemon process
> that
> will b called here
Look at IPaddr as an example, and make sure you implement
each action (start, stop, status...) in an idempotent way.
That is, running a comand many times should be the same as
running it once.
--
Horms
|
