08-28-2010 09:25 AM
From my practical experiences I have upgraded 62 Dirs and SWs from 6.1.x>6.2.x
I have done nothing like disabling QOS and then enabling..
BTW you did not mention these HBAs belong to which OS. I had written in case of AIX, do an rmdev and cfgmgr..
Even if these r G-ports, do u have any issue like LUNs are not visible on that path.did u do nodefind and nsshow.
Thanks for the output of the cmd. Those show nothing.
Could you pls see , if it shows something in errdump.
08-29-2010 01:31 AM
1. Ask GLUX to reset adaptor using diag on an RL server, if there is one on any of these 3 ports. If this works ask them to do this for others.
2. Login as root and check CPU utilization of the switch. 6.2.2 has many bugs related to high CPU utilization, we had the similar problems when I upgraded to 6.1.0h - it took me one month to fix all the problems. Most of them were related to high CPU utilization on the switches and director.
to check CPU utilization
run the command
$vmstat 5 5
08-30-2010 06:28 AM
So the best way is remove the fibre cables of the servers from the SW ports and check again the warnings..But howcome in the other fabric there is no warning. Are these servers on single HBA/path.
BTW u have closed the thread.
Thanks a lot for the information