I am inconsistently getting the following error in /var/adm/messages when I try to switch a resource group which contains a zone to another node:
Dec 2 11:43:22 mltproc1 SC[,SUNW.HAStoragePlus:8,common_shares,common_zpool,hastorageplus_postnet_stop]: [ID 471757 daemo n.error] cannot unmount '/common_pool0/common_zone/root' : Device busy Dec 2 11:43:22 mltproc1 SC[,SUNW.HAStoragePlus:8,common_shares,common_zpool,hastorageplus_postnet_stop]: [ID 952001 daemo n.error] Failed to export :common_pool0 I've done something to cause this, because the resource group was able to be migrated freely earlier. Now it works sometimes, but not other times and I haven't been able to correlate when it will switch correctly and when it won't. The only way to clear this is to issue '/usr/cluster/bin/clrs clear common_zpool' and then must be careful to do a 'clrg switch' to the node where the zpool is still lingering (i.e. switch it to the same node where it was last running), or else I end up with corruption because two nodes are accessing the zpool via iSCSI at the same time (found that out the hard way!) What is strange is that sometimes immediately after the hassle of a failed stop and clearing and restarting the resource group on the same node, I can do a switch to a new node without any problem. -- This message posted from opensolaris.org
