[Linux-HA] OCF drbd, reboot and migration

Philip Gwyn liste at artware.qc.ca
Wed Sep 19 20:51:09 MDT 2007


I've solved my previous problem with drbd : use the group's ID, not a
primitive's ID.

cib.xml : http://www.awale.qc.ca/ha-linux/collect-2/cibadmin-Q.xml
ha.cf   : http://www.awale.qc.ca/ha-linux/collect-2/ha.cf
drbd.conf http://www.awale.qc.ca/ha-linux/collect-2/share.conf

However, I now have 2 problems :

1- If I reboot a node, drbd becomes wedges in a FAILED state.  I have to do
   crm_resource -C -r share-drbd_0:0 -H host-1.awale.qc.ca
  to unwedge it.  This is worse if the node boots up alone; because drbd is
  FAILED, no other resource can be run.

Booting only one node gives me: 
        http://www.awale.qc.ca/ha-linux/collect-2/crm_mon-1r
        http://www.awale.qc.ca/ha-linux/collect-2/ha-debug


2- How do I manually migrate resources from one node to the other?
    The following don't cause a migration (assume everything is running on
     host-2)
    crm_resource -M -r share-FS -H host-1.awale.qc.ca  
    crm_resource -M -r DW -H host-1.awale.qc.ca

  However, I've just noticed that if share-drbd_0:0 is the Master and on host-2
  doing
    crm_resource -M -r share-drbd_0:0 -H host-1.awale.qc.ca

Is this the best/only way?

-Philip



More information about the Linux-HA mailing list