Errors from LVM

[Ninad_1]

Hi,

I got some errors from LVM some time back.I noticed the errors when I saw the syslog.log which I am attaching. Please can anyone tell me what these errors messages mean and what do I need to do ?

The sample errors are as below. The syslog.log file is also attached.

Apr 5 17:12:37 server1 LVM[23611]: Volume Group configuration for /dev/vg_migr_04 has been saved in /etc/lvmconf/vg_migr_04.conf
Apr 5 17:18:26 server1 vmunix:
Apr 5 17:18:26 server1 vmunix: SCSI: Read error -- dev: b 31 0x1c1400, errno: 126, resid: 16384,
Apr 5 17:18:26 server1 vmunix: blkno: 3916224, sectno: 7832448, offset: -284753920, bcount: 16384.
Apr 5 17:18:31 server1 vmunix: SCSI: Read error -- dev: b 31 0x1c1000, errno: 126, resid: 1024,
Apr 5 17:18:31 server1 vmunix: blkno: 8, sectno: 16, offset: 8192, bcount: 1024.
Apr 5 17:18:34 server1 vmunix: ct_query failed. hw_path = 0/2/0/0
Apr 5 17:18:31 server1 vmunix:
Apr 5 17:18:39 server1 vmunix: ct_query failed. hw_path = 0/10/0/0
.
.
Apr 5 17:18:54 server1 vmunix: LVM: Performed a switch for Lun ID = 0 (pv = 0x0000000042f02000), from raw device 0x1f1ce000 (with priority: 0, and current flags: 0x40) to raw device 0x1f2ee000 (with priority: 1, and current flags: 0x0).
.
.
Apr 5 17:18:54 server1 vmunix: LVM: VG 64 0x070000: PVLink 31 0x1c0500 Failed! The PV is still accessible.


Regs,
Nad

[William Wan_1]

Hi Ninad,

It seems that one or more of the disks in your VG vg_migr_04 has became unavailable.
This can be caused by many different reason.
It seems that the path to the storage box has connectivity problem. you can see this because more than 1 path is not availale.

ct_query failed. hw_path = 0/2/0/0
ct_query failed. hw_path = 0/10/0/0


Run a new ioscan and search for any NO_HW.
Check the storage box if all disks / controllers are available.
Also check the SAN switch connectivities if this is a SAN box.

If you want us to check the error, please provide us the following information.

vgdisplay -v vg_migr_04
ioscan -fn
If this is a SAN storage, which one is it?
Show us the configuration information of the SAN box. For example if you have a Virtual Array, use

armdsp -a

Which hp-ux is this? 11.0 ? 11i
Give us the patch information
swlist -l bundle

A well patched server for SCSI / FC is very important.

Hope this can help you a bit

Regards,

William

[Alex Lavrov.]

Hey,
to confirm that it's really a disk failure do:

dd if=/dev/rdsk/cXtXdX of=/dev/null bs=8192K

It will read the data from the disk and put it in /dev/null, this way you check that the disk is OK or not. If dd gets stuck for a long time or giving errors, you can be sure it's a disk failure and the only thing left to do is to replace it.


Good luck,
Alex.

[Florian Heigl (new acc)]

Apr 5 17:18:34 server1 vmunix: ct_query failed. hw_path = 0/2/0/0
Apr 5 17:18:31 server1 vmunix:
Apr 5 17:18:39 server1 vmunix: ct_query failed. hw_path = 0/10/0/0

I think Your zoning was faulty at that moment, it appears that 0/2/0/0 and 0/10/0/0 couldn't reach their other end - at least if this is a FC Fabric configuration.

You will much more detailed info in stm's logtool!