How do I know which disk is faulty?

[Henry Chua]

disk 10 8/0/3/0.2.0 sdisk CLAIMED DEVICE SEAGATE ST39204LW
/dev/dsk/c3t2d0 /dev/rdsk/c3t2d0
disk 11 8/0/3/0.3.0 sdisk CLAIMED DEVICE SEAGATE ST39204LW
/dev/dsk/c3t3d0 /dev/rdsk/c3t3d0
disk 4 8/0/3/0.4.0 sdisk CLAIMED DEVICE SEAGATE ST39173W
/dev/dsk/c3t4d0 /dev/rdsk/c3t4d0
disk 0 8/4/19/0.5.0 sdisk CLAIMED DEVICE IBM DDRS-39130WS
/dev/dsk/c0t5d0 /dev/rdsk/c0t5d0
disk 1 8/4/19/0.6.0 sdisk CLAIMED DEVICE IBM DDRS-39130WS
/dev/dsk/c0t6d0 /dev/rdsk/c0t6d0
disk 2 8/12/5.0.0 sflop CLAIMED DEVICE TEAC FC-1 HF 07
/dev/floppy/c1t0d0 /dev/rfloppy/c1t0d0
/dev/rfloppy/c0t1d0
disk 3 8/12/5.2.0 sdisk CLAIMED DEVICE TOSHIBA CD-ROM XM-6201TA
/dev/dsk/c1t2d0 /dev/rdsk/c1t2d0


SOrry Guys, I posted this last time, but just can't figure it out.. one of these disk is faulty, I figure it is c0t5d0 but how can I be sure?.. and with this info how can I identify the physical location of this disk.

[Sridhar Bhaskarla]

Hi Henry,

If you are talking about the following thread

http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=722740

Your c0t5d0 is showing a size of 0 bytes in your diskinfo output. So, this disk seems to be bad. Is this part of any VG?.

vgdisplay -v |grep c0t5d0

If so, try 'pvdisplay -v /dev/dsk/c0t5d0' and see if there are any LVs associated with it. If there are any LVs, then check if the data is intact. I had seen disks with bad firmware showing bad 'diskinfo' information. So, post your 'pvdisplay,vgdisplay and bdf' outputs before we confirm it as a bad one.

-Sri

[Sanjay_6]

Hi,

Do a diskinfo on each disk and the faulty one may give you an error,

diskinfo -v /dev/rdsk/cxtydz

You can also try to do a dd for each of the disk,

dd if=/dev/dsk/cxtydz of=/dev/null bs=1024k

The faulty one should give you error.

What about your syslog, Is anything getting logged into syslog.

what about EMS / onlineDiag. Don;t you have onlineDiag installed on the system. If you do that should say something in syslog / root email.

Looks like the disks at c0t5d0 and c0t6d0 are internal and the other disks at c3... are external. What is your setup / what type of storage ?.

Hope this helps.

Regds

[Romen Q. Cordial_1]

Hi Henry,

First of all, how do you know that you have disk faulty, does somebody alarm you or application people or user complain that they can't access data. Check the ff:
1. /var/adm/syslog/syslog.log & OLDsyslog.log,
2. type dmesg - this will tell you if you have any problem.
3. check /var/opt/resmon/log/event.log
4. bdf - to check if the Filesystem are all mounted.
5. strings /etc/lvmtab - to see what LVM has been assign to the disk you think is faulty.
6. ioscan -fnC disk check again if still claimed.
7. to identtify the physical location you can check this by understanding the path of this disk 0 8/4/19/0.5.0 from back panel of your computer you can trace it, or if it is a storage disk, normaly the light is on steady yellow light not blinking.

[T G Manikandan]

Just do a

vgdisplay -v

check the the Actual PV and Claimed PV.


Also,check your /var/adm/syslog/syslog.log

The log will give you the messages as which disk has the problem.

[Henry Chua]

Hi Guys,

Sanjay is right in saying that it is indeed an internal harddisk. But Since I have 2 internal drive, how can I make sure which one it is without goin thru a game of mastermind.. Romen, u were saying that I can identify the system drive by looking at the hardware path.. is it possible to elaborate on this?

thank you all for your support..

[Sridhar Bhaskarla]

Hi,

What is the hardware model?. One of the tricks I use is to dd the disk and watch the LED on the disk.

Since c0t6d0 is working fine, do a

dd if=/dev/dsk/c0t6d0 of=/dev/null bs=1024k

Then watch both the drives. The one that is continuously "on" is c0t6d0 and the one that is off is c0t5d0. Stop 'dd' and make sure the drive goes off. Repeat the above again couple of times to confirm.

-Sri

[Sยภเl Kย๓คг]

1.Try diskinfo
2.What is the LED status on the disk,is it green or orange color.
3.Physical location can be identified by the Hardware Path given.

regards
SK
suniilhcl@rediffmail.com

[Isralyn Manalac_1]

Hi Henry,

The diskinfo -v /dev/dsk/cXtXdX will provide you the information you need. If the output comes up with 0 bytes, then you have a bad disk. After that, do an ioscan and jot down the hardware address of the faulty disk.

Regards,

Ira

[Henry Chua]

Thank guys for the info.. I have identified the faulty harddisk.

As I have 2 similar system, thus I attempted to used the good system drive to replace the fualty one. However I got this message on the boot-up screen.

vgchange: warning: couldn't query physical volume "dev/dsk/c0t5d0". The specified path does not correspond to physical volume attached to this volume group.

what does this means?

[Sridhar Bhaskarla]

Henry,

This was already part of a VG. Since you replaced the disk, VG activation will complain as it doesn't have corresponding VG headers. What you have to do is to restore them once the system is completely up.

The restoral process is dependent on how the disk was being used. There are three scenarios.

1. This disk is under vg00 and is a mirror boot disk.
2. This disk is not in vg00 and mirrored.
3. This disk is unmirrored in it's VG.

Case 1: Search forums for 'boot mirror' and you will find many threads indicating solutions.

Case 2: Since the disk is mirrored, your data is protected, all you have to do is to restore the mirrors. Use the following process. Replace vgxx with the correct VG (you can get it from 'strings /etc/lvmtab')

vgcfgrestore -n vgxx /dev/rdsk/c0t5d0
vgchange -a y vgxx
vgsync vgxx

Case 3: The data is lost on all the LVs that used this disk. Restore the VG configuration and restore the data from the backups.

vgcfgrestore -n vgxx /dev/rdsk/c0t5d0
vgchange -a y vgxx
mount -a

Look at the filesystems that do not have data and restore them from backups.

-Sri