So several days ago I had a major San failure and lost half the volumes access wise. So my Esx hosts lost about 40 of their 80 volumes for several hours. These hosts also are connected to another San which was functioning properly.
The vms on the good San were working fine and the vms on the bad array but on the volumes that were still visible were working good too.
However all my hosts at this point were disconnectEd in vcenter. I could not connect directly to them with a thick client. Nor login directly from the console with f2 or altf1.
I am certain I was in an APD situation. However I was under the impression that after 140 seconds the hosts would flush the io and declsre th paths dead and move on. that appears not to happen So i must be misunderstood.
So now i question having multiple arrays attached because of you havr major issues with one array it appears you are pretty much dead until it's fixed, so you can't manage any vms or anything on the good array.
Am I misunderstanding thr APD behavior and the 140s timeout?
we are running v5.5 update2 and some other patches.