Warning LVM

[Dimitri di Paolo]

hi,
the first message I recived (from diagnostic command dmesg) in dicember was the follow :
***************
SCSI: Abort abandoned -- lbolt: 21911, dev: 1f012000, io_id: 10026db, status: 200
SCSI: Abort abandoned -- lbolt: 21941, dev: 1f012000, io_id: 10026df, status: 200
SCSI: Abort abandoned -- lbolt: 21970, dev: 1f012000, io_id: 10026de, status: 200
SCSI: Resetting SCSI -- lbolt: 22072, bus: 1
SCSI: Reset detected -- lbolt: 22072, bus: 1
DIAGNOSTIC SYSTEM WARNING:
The diagnostic logging facility has started receiving excessive
errors from the I/O subsystem. I/O error entries will be lost
until the cause of the excessive I/O logging is corrected.
If the diaglogd daemon is not active, use the Daemon Startup command
in stm to start it.
If the diaglogd daemon is active, use the logtool utility in stm
to determine which I/O subsystem is logging excessive errors.
LVM: Recovered Path (device 0x1f012000) to PV 0 in VG 0.
LVM: Restored PV 0 to VG 0
***************
but the server seem working well until now.
Now the server works only in single user mode, mounting the file system manually (only after the command vgchange with -a y option) but sometime I am reciving the follow Warning and then the server blocks (no prompt):
****************************************
DIAGNOSTIC SYSTEM WARNING:
The diagnostic logging facility has started reciving excessive errors from the I/O subsystem. I/O error entries will be lost until the cause of the excessive I/O logging is corrected.
If the diaglogd daemonis not active, use the Daemon Startup command in stm to start it.
If the diaglogd daemon is active, use the logtool utility in stm to determine which I/O subsystem is logging excessive errors.
LVM: vg[0]: pvnum=0 (dev_t=0x1f012000) is POWERFAILED.
*****************************************

May I know if it is a hardware or software problem?
Please, help me.

[Mark Grant]

It looks like you have a disk in vg00 that is failing. It is definately a hardware issue and you need to get it sorted by your friendly HP engineer.

[Fabio Ettore]

Ciao Dimitri,

problem is about device DISK (1f) and device file /dev/dsk/c1t2d0 (012000).

As Mark said it is an hw problem.

Best regards,
Ettore

[T G Manikandan · Accepted solution]

WHat OS version you are running

The problem could be

1.Bad hard disk c1t2d0
just do a diskinfo on that disk
2.check your scsi cabling and scsi termination
3.Load all the required scsi patches

4.Try increasing the timeout of the disk using

pvchange

[Elmar P. Kolkman]

The reason the system kept running is probably you have mirroring on your lvols or the disk is not used yet.
You could try to do a 'vgchange -a y -q n '...

But like the other said, to solve this you need to check and possibly replace the failing disk.

[Kent Ostby]

The "DIAGNOSTIC SYSTEM WARNING" message here is the key. It pretty much tells you that you have a HW problem of some sort.

As to which device, you can use the last six digits of the "dev_t" line above (i.e. (dev_t=0x1f012000) as follows:

ll /dev/dsk/* | grep 012000

In most cases, this will fall out as /dev/dsk/c1t2d0 or something like that.

You can then see which volume group this belongs to by doing:

pvdisplay /dev/dsk/c1t2d0

Given that the system refers to it as PV0 of VG0, it could be your root disk, but go through the pvdisplay step above to make sure.

Best regards,

Kent M. Ostby

[Dimitri di Paolo]

Thanks very much everybody.
Sorry for incomplete informations.
The server is a HP 9000/800 L1000, OS HP-UX 11.00.
I have increased the timeout up to 90 sec (pvchange command) on c1t2d0 (boot disk, 9Gb, no mirror) and now it seems that the warning message "LVM ... powerfailed" are terminated.
if I mount all file system and then I go to init 3, the server up regolarly but after login it blocks (no prompt for password).
At this point, I ask :"maybe the problem is a software problem"?

[Fabio Ettore]

Instead I press on an hw problem. Even if you don't obtain LVM powerfailed messages anymore it doens't mean that problem on the bad disk is solved.
What about "DIAGNOSTIC SYSTEM WARNING:" messages? Are they terminated too?
I suggest you to open an HP hardware call in order to check the disk /dev/dsk/c1t2d0.

Best regards,
Ettore

[Sridhar Bhaskarla]

I agree with Ettore. Software problems do not cause POWERFAILED messages. It's sure to do with your disk c1t2d0. You may need to get it replaced. A mirror would have saved you. Try remove and put it back. Not that it will fix the problem but you may be lucky enough to get it up until you can take a make_tape_recovery tape.

-Sri

[Anoop P_2]

Remember, while u could be lucky to get ur disk probs solved by pulling it and putting it back, you could also be equally unlucky by doing so, left with a non-bootable machine due to a complete disk failure!

[Mic V.]

I agree with the two previous posters. Yes, I've done the pull/replace trick to get a disk running again, sometimes several times. But the disk does end up dying.

I once got a bad batch of hot-pluggable drives of questionable origin that slowly went bad in this way over time.

Mic

[donne007]

I Agree with the poster!
But if you suspect the HDD to be faulty Try this out

1. # diskinfo /dev/rdsk/cxtxdx
Which should give you some info abt the disk i it's good!

2. # dd if=/dev/dsk/cxtxdx of=/dev/null
you should get an o/p like n file in & n file out.. BUT BE CAREFULL WHILE RUNNING THIS AS THIS MAY TAKE YOUR DISK FOR A TOSS !!!


Good luck
Asif

[Dimitri di Paolo]

Hello everybody.
Thank you for your informations. I changed disk, backplane SCSI adapter and backplane disk but after I running "sea" command, I have always ALERT LEVEL:7=RESERVED and 0x5800387000FF6292 ....... -type11 error.

Help me please! I don't know what's the problem.

Tank's

[Dimitri di Paolo]

I have changed disk vg00 and Core I/O card and now the server is OK !!!
Sorry for my questions, thanks very much everybody.