topic Re: Error message in syslog in Operating System - HP-UX

Error message in syslog

Gordon_3 — Wed, 23 Jun 2004 06:02:13 GMT

Hi all,

Our machine just undergo PV link failover which syslog has log the error, actually I want to know how to interpret below err message, from first 2 line, Path failure indicate the c21t0d0 & c21t0d2 has path problem ( 0x1f150000 & 0x1f150200 ), but why the next 4 line indicate another path c20t0d0 - c20t0d4 has problem?? From the result, it seems the c20 is the primary path and c21 is fail path, can anybody teach me how to interpret the result?? thx.

Gordon

Jun 23 17:41:21 hkgdevs vmunix: LVM: Path (device 0x1f150000) to PV 0 in VG 4 Failed!
Jun 23 17:41:21 hkgdevs vmunix: LVM: Path (device 0x1f150200) to PV 1 in VG 4 Failed!
Jun 23 17:41:21 hkgdevs vmunix: LVM: vg[4]: pvnum=0 (dev_t=0x1f140000) is POWERFAILED
Jun 23 17:41:21 hkgdevs vmunix: LVM: vg[4]: pvnum=1 (dev_t=0x1f140200) is POWERFAILED
Jun 23 17:41:21 hkgdevs vmunix: LVM: vg[4]: pvnum=2 (dev_t=0x1f140100) is POWERFAILED
Jun 23 17:41:21 hkgdevs vmunix: LVM: vg[4]: pvnum=3 (dev_t=0x1f140300) is POWERFAILED
Jun 23 17:41:28 hkgdevs vmunix: LVM: Restored PV 2 to VG 4.

Re: Error message in syslog

Muthukumar_5 — Wed, 23 Jun 2004 06:55:26 GMT

hai,

Because of linkfailover , Physical Volume 0 and 1's link from Volume group 4 failed. It is indicated that in the first 2 linee.

On Volume group 4, four physical volumes are indicated as POWERFAILED because of linkfail.

Check the pvdisplay,vgdisplay to know about the current physical volumes and volume managers.

Use pv* man pages and vg* man pages.

Regards,
Muthukumar

Re: Error message in syslog

Bruno Ganino — Wed, 23 Jun 2004 07:04:50 GMT

If this is useful

http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=615883

http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=240774

Bruno

Re: Error message in syslog

Bruno Ganino — Wed, 23 Jun 2004 09:09:15 GMT

Hi,
There are PV parameters that can be changed.
A frequently used parameter is the IO timeout parameter.
This parameter tells LVM how long to wait for disk transactions to complete before taking the device offline. This is accompanied by POWERFAILED messages on the console.
Certain disk need a higher timeout value.

# pvchange -t value device

For check if a LV has extents on that PV the key can be obtained using the -v -k option of lvdisplay command.

HTH
Bruno

Re: Error message in syslog

Ashwani Kashyap — Wed, 23 Jun 2004 09:31:11 GMT

from above the problematic disks indeed are c20t0d0,c20t0d1,c20t0d2 and c20t0d3 .

0x1f140200
0x = Just Hex
1f = dec =31 = Major number for scsi disk
14= dec 20 = the C number for the disk
0 = Target number
0 = Lun number

Next two zeros mean nothing .

How ever your problem might kust be of timeouts .

Change the timeouts of the problematic disks online by
#pvchange -t 180 /dev/dsk/c#t#d#

Re: Error message in syslog

Steven E. Protter — Wed, 23 Jun 2004 09:45:07 GMT

Looks to me like a good failover from one path to another.

The system didn't go down did it? No loss of service?

If so it seems to validate that you got PVLINKS setup correct.

Pat on the back, nice job.

Looks like one of your fiber cards may be a little flakey or scsi if its not a san.

SEP

Re: Error message in syslog

Gordon_3 — Thu, 24 Jun 2004 01:21:16 GMT

Hi all expert,

Yes u right, the PV link failover is success, current PV timeout for disk setting is 90. Also thx for pointing out how to interpret the hex code above, but I still don't understand why first 2 line complain 0x1f15, but next 4 line complain about 0x1f14 path? To me, it seems both first 2 line and the next four line are complaining to a failure, but on different path, no matter it say "Path" or "Disk Powerfailure", this cause I am so confusing, thx.

Gordon