Description of problem: In case if SHE-host has connection to iSCSI LUN, on which SHE-VM is deployed/being deployed, over multipath, allow reinitiating connectivity over another path, if initial path fails. Version-Release number of selected component (if applicable): ovirt-hosted-engine-setup-2.2.0-0.0.master.20171016160008.git55723e8.el7.centos.noarch How reproducible: 100% Steps to Reproduce: 1.Deploy SHE on pair of hosts, over iSCSI LUN, while iSCSI storage server exposing the LUN from multiple NICs. 2.From host with SHE-VM block one of the iSCSI paths with iptables, i.e. "iptables -A OUTPUT -p tcp --destination-port 3260 -d <IP address of one of the iSCSI NICs> -j DROP" 3.Check status of "hosted-engine --vm-status". 4.Check "iscsiadm -m session". Actual results: I'm getting SHE-VM stack in paused status if one of the paths being blocked. I'm getting all paths as active using "iscsiadm -m session", although at least one of the paths is blocked. Expected results: "iscsiadm -m session" should be refreshed and return up to date status. If one of the paths fails, then host should fail over to remaining best active path. Additional info: Sosreport host on which I've found the issue is attached.
Created attachment 1343361 [details] sosreport from alma04
This was the whole point of multiple targets
(In reply to Yaniv Kaul from comment #2) > This was the whole point of multiple targets I don't see this working. iSCSI initiator not shifting to other feasible targets on my pair of hosts.
After retesting the https://bugzilla.redhat.com/show_bug.cgi?id=1149579 it is working as expected and fully covers this very RFE. Moving back to previous state as was set in comment #2.