On 1/12/2010 at 08:25 PM, Frank Lazzarini <[email protected]> wrote: > Hi all, and a happy new year to all of you! > > I've got the following setup, Ubuntu 9.10 karmic with the following > packages. Drbd v8.3.3, Pacemake v1.0.5 (i know it's not the latest) and > samba. The setup is a two node failover system which serves as > fileserver mainly. Everything works fine as it is, if I switch off the > first node every resources jumps to node2 and backwards. My nodes are > called ewadata1 and ewadata2. Here is my configuration. > > node $id="76799c18-3c56-48fd-9c76-770c35fc9222" ewadata2 \ > attributes standby="off" > node $id="b8e36dc5-9757-49a4-9120-874db26a29ad" ewadata1 \ > attributes standby="off" > primitive IP ocf:heartbeat:IPaddr \ > params ip="10.65.68.254" \ > meta is-managed="true" target-role="Started" > primitive drbd_r0 ocf:heartbeat:drbd \ > params drbd_resource="r0" \ > op monitor interval="9s" role="Master" timeout="20s" \ > op monitor interval="10s" role="Slave" timeout="20s" > primitive fs-drbd ocf:heartbeat:Filesystem \ > params fstype="ext4" directory="/data" device="/dev/drbd0" \ > meta target-role="Started" > primitive samba lsb:samba \ > op monitor interval="20s" timeout="10s" \ > meta target-role="Started" > ms ms-drbd_r0 drbd_r0 \ > meta clone-max="2" notify="true" globally-unique="false" > target-role="Started" > location l_prefer_IP IP \ > rule $id="l_prefer_rule_IP" inf: #uname eq ewadata1 > location l_prefer_fs-drbd fs-drbd \ > rule $id="l_prefer_rule_fs-drbd" inf: #uname eq ewadata1 > location l_prefer_samba samba \ > rule $id="l_prefer_samba_rule" inf: #uname eq ewadata1 > location ms-drbd_r0_master ms-drbd_r0 \ > rule $id="l_prefer_rule_drbd_r0" $role="master" 100: #uname eq ewadata1 > order fs-drbd_before_samba inf: fs-drbd:start samba:start > order ms-drbd_r0_before_fs-drbd inf: ms-drbd_r0:promote fs-drbd:start
Strictly speaking you probably want some more constraints here. There doesn't seem to be an ordering constraint to ensure the IP comes up before Samba. Also, you should colocate the IP, Samba and Filesystem, otherwise it's theoretically possible Samba could fail on ewadata1, then try to migrate to ewadata2, but leaving the Filesystem running on ewdata1 (try "crm resource migrate samba ewadata2" if you want to see this in action). > property $id="cib-bootstrap-options" \ > stonith-enabled="false" \ > no-quorum-policy="ignore" \ > dc-version="1.0.5-3840e6b5a305ccb803d29b468556739e75532d56" \ > cluster-infrastructure="Heartbeat" \ > last-lrm-refresh="1263209944" > > > When I open up crm_mon I get an error ... samba_monitor_0 > (node=ewadata2, call=118, rc=1, status=complete): unknown error ... I've > tried cleaning up the resource multiple times but the error reappears > over and over again. Somehow it tries to monitor a resource that isn't > running yet. My question do I have to setup some rule for master/slave > although samba isn't a master/slave resource, or how can I get rid of > this error? Pacemaker will run a monitor op once prior to starting a resource, to make sure it's not already running. Possibly you're seeing this monitor op, but the init script is returning failure instead of "not running" for some reason. You'll need to check the logs (grep for "call=118, rc=1" will probably get you close) to see if there's any indication of what's failing. > P.s. I've tried changing from the lsb script to the heartbeat ocf script > of samba but when I use the ocf script of samba with this configuration > samba doesn't start on node2 when node1 is offline. There's an OCF RA for Samba? I know there's one for CTDB, but didn't think anyone had written a standalone Samba RA. Regards, Tim -- Tim Serong <[email protected]> Senior Clustering Engineer, Novell Inc. _______________________________________________ Linux-HA mailing list [email protected] http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
