01-20-2014 07:24 AM
We have a qlogic HBA (qla2xxx) running on Red Hat that seems to be failing on OS startup. Exactly how its failing I am not sure. The card was connected to a Cisco UCS environment but was removed and is now connected to a cisco 9138 but fails at fabric login. The FCID showing up for this HBA is 0x652100 and seems to have been configured in with the card. I have set up a persistent fcid on the switch but the HBA doesn't seem to be getting this far. Can anyone tell me how I can troubleshoot this further? This is a production system so I am limited with my actions. We have an a and b fabric, the failed HBA is on a fabric. The other HBA is working fine.
01-20-2014 08:56 AM
Hello,
Is the device logging into fabric A ? You can use the following commands to verify -
show flogi database
show fcns database
show npv flogi (if this is an NPV enabled 9100)
Is the zoning set up correctly for this device (show zoneset active) ? Is it done on this 9100 or a different MDS (NPV Core) ?
Thanks,
Arun
01-20-2014 09:06 AM
No, the HBA is not logging into the fabric on A. I do not see an FCID for the port nor do I see it on the nameserver.
Here is a cat of all files in the /sys/class/fc_host/host0 directory which is the directory of the non-working adapter
cd /sys/class/fc_host/host0
[root@cporacle2 host0]# cat *
cat: device: Is a directory
0x1000080088020b32
cat: issue_lip: Permission denied
0x5001438000c55cbf
0x652100
0x5001438000c55cbe
Offline
NPort (fabric via point-to-point)
4 Gbit
cat: statistics: Is a directory
cat: subsystem: Is a directory
Class 3
HPAE311A FW:v4.04.05 DVR:v8.02.00.06.05.03-k
wwpn (World Wide Port Name)
cat: uevent: Permission denied
01-21-2014 12:31 AM
Hello
I think the hard coded FCID may be the issue. Can you turn it off and let it get one normally? Can you also provide an FCID of a working device in that switch?
Steven
01-21-2014 06:13 AM
Thanks Steven, that was my thinking, as well, but I don't know how I would "turn off" a hard coded FCID? Working FCID's on this MDS 9134 are domain 65 00 00 through 65 19 00. I've already assigned a persistent FCID to this card - the next one on the list, namely 65 1a 00. However, the card is still not logging into the fabric.
However, since my original post yesterday, I've noted that this card as been receiving "MAC bit errors" for quite some time so I am suspecting bad hardware somewhere along the storage path. I will try replacing the GBIC (SFP) and fixing other possible hardware issues.
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide