0/1/2/0: Unable to access previously accessed device at nport ID 0x653d13.

[Stuart Abramson]

We have been "cleaning up" our EMC volume allocation (moving from old to new Symms and consolidating servers on single Symms). So we have been replacing "old" PVs on servers with "newer" PVs.

On the MC/ServiceGuard machines I have noticed the message above, and also this one:

Mar 2 15:51:05 tarr1 cmclconfd[3908]: Unable to open disk /dev/dsk/c25t0d5: No such device or address

THESE VOLUMES AREN'T USED IN ANY WAY! There isn't anybody accessing them.

Why am I getting these? Where are they coming from? How do I figure out what "0x653d13" is?

Here's an odd thing:
1. On some machines the disks don't show up at all in "ioscan", because we removed the volumes.
2. In some machines they show up in "ioscan" as CLAIMED.
3. In some machines they show up in "ioscan" as NO_HW.
4. In all cases if you "diskinfo", you get an error, because, of course, we removed them.

I found an old forum entry which referred to EMS - I can clean this message up with "EMS maintenance", whatever that is.

Does MC/ServiceGuard keep track of volumes on a cluster machine even if they aren't shared?

I suspect I can clean this up with a reboot - it isn't really a problem, it's just annoying.

[RAC_1]

MCSG does not keep track of all vgs on machine, but only under it's controll. It keeps track of vgs and not pvs. Does anyof the VGS (under MCSG) has that PV?? (Before EMC change or after EMC change)

[Stuart Abramson]

No. None of the volumes are involved in the cluster. Never were.

It also happens on non-MC/SG machines.

[RAC_1]

See the following thread and follow the procedure given.

[RAC_1]

Forgot the thread.

http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=127156

[joinsiva]

hi staurt,

i feel you are facing problem becoz of
fibre channel card connectivity at address 0/1/2/0 .

findout what card is ther in that address by,

#ioscan -funC fc

tract your hardware path and note down the device file for that card.

and do,

#fcmsutil .

analyse the output
check the nport id which u mentioned is same here...

check the link speed and driver state.

some time it happens like your fc card link works at 1Gb/s and your EMC with 2gb/s, even in this case you will face problem. that time you manually need to set the speed in the EMC to 1Gb/s.

even u can try, fcutil command for tesing your card communication over san. A short description for fcutil you can find here..

all the best.

sivakumar.p.

[joinsiva]

sorry forgot to attach you the link for fcutil command

http://docs.hp.com/en/B2355-90681/fcutil.1M.html

sivakumar.p.

[rose huang]

I am having the same events after I clean up my XP1024 devices. I clean up all my NO_HW by using rmsf -H. I am thinking about open a call with HP. Not sure if this is tied to the Tachyon card or the EMS. I recycled the diagnostic and the event re-appeared after 24 hours.