I'm still having no luck with this cluster.
In hope that I just did something inconsisent between the nodes, I removed the
cluster (scinstall -r) from all nodes, along with all iSCSI configuration, and
have started over again. In short, I am back to the same problem (a node sits
indefinitely waiting for quorum, and I have no idea how to debug this).
My process is:
pkg install ha-cluster-full (on all 4 nodes)
on one node (mltproc0): /usr/cluster/bin/scinstall
Selecting install for just this node, typical mode, input all 4 node names. The
only place where I'm not 100% certain on what I'm doing is when entering the
private interconnects. I am entering the name of the vnic (vmltx1 in my
original post), telling it that it IS a dedicated interface (the default), and
then overriding the default to tell it that it IS Ethernet.
After I enter all data, allow script to reboot the first node.
Once the first node comes all the way back up, then go to the next node and do
scinstall, tell it to join an existing cluster, similar answers all the way
through, allow it to reboot, and the go to the next.
At this point, after all 4 nodes are in the cluster, I ran
/usr/cluster/bin/clsetup, which did not ask any questions before presenting the
main menu.
I rebooted mltproc1, w/o any problems (it rejoined the cluster on startup).
I then rebooted mltproc0 - and it just sits at
NOTICE: CMM: Cluster doesn't have operational quorum yes; waiting for quorum.
>From another node:
root at mltproc1:~# /usr/cluster/bin/clquorum status
=== Cluster Quorum ===
--- Quorum Votes Summary from (latest node reconfiguration) ---
Needed Present Possible
------ ------- --------
3 3 4
--- Quorum Votes by Node (current status) ---
Node Name Present Possible Status
--------- ------- -------- ------
mltproc0 0 1 Offline
mltproc1 1 1 Online
mltstore1 1 1 Online
mltstore0 1 1 Online
--- Global Quorum Health Check (current status) ---
Node Name Health Check Type Entities Status
--------- ----------------- -------- ------
--
This message posted from opensolaris.org
