Updating my ISCSI SAN

[msmart]

Strange issue. I updated my SAN the other day which brought it offline for about 30 seconds during a power cycle of the controllers. 30 seconds is well within the configured timeouts in iscsi.conf. After the controllers came back online all of my VMs were fully functional but proxmox itself indicated that the LVM/ISCSI storage was offline and would not let me perform any operations on it. This included live migration. The only solution I found was to shut down all vms on each PVE node in my cluster and reboot the node. It seemed like easily avoidable downtime.

My PVE 1.x cluster that connects to the same SAN did not have this problem. It recovered just fine. 1.x also has the pvesm enable command that I suspect would have fixed the issue without a reboot. Is there anything I can do in 2.X to recover from the offline state next time I have to update my SAN firmware? Is there anything analogous to pvesm enable in 2.x?

Thanks,
Matt

 

[mir]

If the session still exists the following reinitiate your connections: iscsiadm -m node -R

If the session is not active. Eg. the above failed create a new session:
iscsiadm -m node
iscsiadm -m node -T <target> -p <portal> -l

Example
# iscsiadm -m node
172.16.2.2:3260,2 iqn.2010-09.org.openindiana:vmotion1:vdata1
192.168.2.10:3260,1 iqn.2004-04.com.qnap:ts-212:iscsi.qdisk.c8c1d1
192.168.2.10:3260,1 iqn.2004-04.com.qnap:ts-212:iscsi.qnap.c8c1d1

Output is to be read: <portal>,<session number> <target>
The first listed node is my SAN which I want to make a connection to
iscsiadm -m node -T iqn.2010-09.org.openindiana:vshare -p 172.16.2.2:3260 -l

If the above command returns with success then your pve-node will see your LUN again.

The above must be done on every node in the cluster.