I'm having some troubles getting Sun Cluster 3.2u1 or u2 working properly with our quorum devices. I have a 2 node cluster, interconnected with crossovers to nge2 and nge3 on each machine. They're both connected to two J4400s (each node has a SAS connection to each J4400), and can see and use all disks properly. When I install/setup the cluster, things go fine until I reboot one of the nodes. Instead of the other node attempting to failover, it panics, as it can't obtain quorum. clquorum shows i have 2 nodes present out of 3, even when both servers and the quorum device are all online and contributing a vote each. It seems that the quorum device on our J4400 isn't working properly, even though it thinks it is.
clquorum status shows:
Cluster Quorum ===
--- Quorum Votes Summary ---
Needed Present Possible
------ ------- --------
2 2 3
--- Quorum Votes by Node ---
Node Name Present Possible Status
--------- ------- -------- ------
cletus 1 1 Online
brandine 1 1 Online
--- Quorum Votes by Device ---
Device Name Present Possible Status
----------- ------- -------- ------
d1 1 1 Online
cluster show displays
Quorum Devices ===
Quorum Device Name: d1
Enabled: yes
Votes: 1
Global Name: /dev/did/rdsk/d1s2
Type: shared_disk
Access Mode: scsi2
Hosts (enabled): cletus, brandine
When I check /var/adm/messages on the cluster nodes, i see:
Feb 3 15:51:00 cletus genunix: [ID 310555 kern.notice] NOTICE: CMM: Quorum device 1 (/dev/did/rdsk/d1s2) added; votecount = 1, bitmask of nodes with configured paths = 0x3.
Feb 3 15:51:00 cletus genunix: [ID 515411 kern.notice] NOTICE: CMM: Registered key on and acquired quorum device 1 (gdevname /dev/did/rdsk/d1s2).
Feb 3 15:51:00 cletus genunix: [ID 250885 kern.notice] NOTICE: CMM: Quorum device /dev/did/rdsk/d1s2: owner set to node 1.Feb 3 15:51:25 cletus genunix: [ID 795311 kern.warning] WARNING: CMM: Issuing a NULL Preempt failed on quorum device /dev/did/rdsk/d1s2 with error 2.
All of my searching has come up with not much. I'm new to Sun Cluster and Solaris, so I know I'm doing something wrong. Most likely, it's painfully obvious. Any help would be much appreciated.