Hi,

On Mon, Feb 21, 2011 at 05:44:34PM +0000, Brett Delle Grazie wrote:
> Hi,
> 
> If I have two DRBD resources:
> resourceA
> resourceB
> 
> What's the best way to configure the constraints in pacemaker so that
> a tomcat resource, which serves both of them when mounted,
> will not be started unless they are both present? My current
> configuration (below) fails to migrate correctly when drbd_B fails due
> to network glitch
> and tomcat fails to stop.  In this case, the demote on drbd_B fails
> because the directory is still mounted and I end up with a split brain
> scenario.  Basically I want to force drbd_A and drbd_B to be master on
> the same system.
> 
> Currently I have:
> node nodeA
> node nodeB
> 
> primitive drbd_A ocf:linbit:drbd \
>         params drbd_resource="A" \
>         op start interval="0" timeout="240" \
>         op stop interval="0" timeout="100" \
>         op monitor interval="30"
> primitive drbd_B ocf:linbit:drbd \
>         params drbd_resource="B" \
>         op start interval="0" timeout="240" \
>         op stop interval="0" timeout="100" \
>         op monitor interval="30"
> 
> primitive fs_A ocf:heartbeat:Filesystem \
>         params device="/dev/drbd/by-res/A" directory="/mnt/A"
> fstype="ext3" options="defaults,noatime" \
>         op start interval="0" timeout="60" \
>         op stop interval="0" timeout="60" \
>         op monitor interval="60" timeout="40" depth="0"
> primitive fs_B ocf:heartbeat:Filesystem \
>         params device="/dev/drbd/by-res/B" directory="/mnt/B"
> fstype="ext3" options="defaults,noatime" \
>         op start interval="0" timeout="60" \
>         op stop interval="0" timeout="60" \
>         op monitor interval="60" timeout="40" depth="0"
> 
> primitive tomcat_tc1 ocf:heartbeat:tomcat \
>         params tomcat_user="tomcat" catalina_home="/opt/tomcat6"
> catalina_base="/home/tomcat/tc1"
> catalina_pid="/home/tomcat/tc1/temp/tomcat.pid"
> catalina_rotate_log="NO" script_log="/home/tomcat/tc1/logs/tc1.log"
> statusurl="http://localhost/manager/serverinfo";
> java_home="/usr/lib/jvm/java" \
>         op start interval="0" timeout="60" \
>         op stop interval="0" timeout="20" \
>         op monitor interval="60" timeout="30" start-delay="60"
> primitive vip_1 ocf:heartbeat:IPaddr2 \
>         params ip="10.xx,xx,xx" nic="bond0" iflabel="3" \
>         op monitor interval="10" timeout="20"
> group grp_1 fs_A fs_B vip_1 tomcat_tc1
> 
> ms ms_drbd_A drbd_A \
>         meta master-max="1" master-node-max="1" clone-max="2"
> clone-node-max="1" notify="true"
> ms ms_drbd_B drbd_B \
>         meta master-max="1" master-node-max="1" clone-max="2"
> clone-node-max="1" notify="true"
> 
> location loc_ms_drbd_A ms_drbd_A 100: nodeA
> location loc_ms_drbd_B ms_drbd_B 100: nodeA

You'd want to add $role=Master in these two. You could also
collocate them:

colocation two_drbd inf: ms_drbd_A:Master ms_drbd_B:Master

> colocation A_on_drbdA inf: grp_1 ms_drbd_A:Master
> colocation B_on_drbdB inf: grp_1 ms_drbd_B:Master
> 
> order A_after_drbdA inf: ms_drbd_A:promote grp_1:start

The order for _B missing?

> Is there a better way to configure this?

Otherwise, it seems fine to me.

One alternative is to put the two drbd resources in a group then
make a m/s resource of that group:

group drbd_AB drbd_A drbd_B
ms ms_drbd drbd_AB
colocation tomcat_on_drbd inf: grp_1 ms_drbd:Master
order tomcat_after_drbd inf: ms_drbd:promote grp_1

But in that case the two drbd will start sequentially.

Thanks,

Dejan

> Thanks for any help  /pointers.
> 
> -- 
> Best Regards,
> 
> Brett Delle Grazie
> 
> _______________________________________________
> Pacemaker mailing list: Pacemaker@oss.clusterlabs.org
> http://oss.clusterlabs.org/mailman/listinfo/pacemaker
> 
> Project Home: http://www.clusterlabs.org
> Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
> Bugs: 
> http://developerbugs.linux-foundation.org/enter_bug.cgi?product=Pacemaker

_______________________________________________
Pacemaker mailing list: Pacemaker@oss.clusterlabs.org
http://oss.clusterlabs.org/mailman/listinfo/pacemaker

Project Home: http://www.clusterlabs.org
Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
Bugs: http://developerbugs.linux-foundation.org/enter_bug.cgi?product=Pacemaker

Reply via email to