IO errors.

[Vijeesh CTK]

Hi gurus,

I am facing a strange problem in in HA environment( MCSG verion A.11.08.. I am having 2 server on a cluster. both same config..connected to a StorageTek FC disk array.
Both servers are with June 2000 HWCR patch. we installed HWE Sept 2002 and OnlineDiag Dec 2002 on one server (fail-over server )(active-passive mode ). after that we are getting a lot errors in that server for that cluster lock disk on the passive server in EMS and syslog.. in syslog it is getting
Feb 11 11:20:10 prod2 vmunix: SCSI: Read error -- dev: b 31 0x050700, errno: 126, resid: 1024,
Feb 11 11:20:10 prod2 vmunix: blkno: 8, sectno: 16, offset: 8192, bcount: 1024.

a lot errors like this.

and in iostat also.. other than 2 root disks it is showing the cluster lock disk

I am worried about this errors.

please help me

Cheers

Vijeesh CTK

[Steve Steel]

Hi


Probably a bad patch.

Which OS version is it.


Steve Steel

[Vijeesh CTK]

OS HP-UX 11.00

[Stefan Farrelly]

It sounds like your patch/diag upgrade has introduced new behaviour from the diags/drivers which is either showing errors which arent there or showing errors you simply couldnt see before.

What QPK bundle are you running ? it should match the diags - Dec2002.

You can try searching for any patches for the Dec2002 diags, or try a different version - say the Sep2002 diags to see if the errors keep happening.

[Armin Feller]

If you have a hw support contract with HP, please open a call at you local HP Responce Center. It seams that the disk is defect and should be replaced ;-(

But you should also check your patches (LVM, SCSI, XVFS).

[Steve Steel]

Hi

Look at patch 28131 . see list below

PHKL_28131
PHKL_18543
PHCO_23651
PHCO_21187
PHKL_17038
PHKL_21392
PHKL_20016
PHKL_20674
PHKL_25475
PHKL_24187

This is the latest scsi + dependencies. Install these and the message should go away. If it does not log a HW call.


Steve Steel

[Vijeesh CTK]

Hi Stefan

No QPK installed.. after installing HWE it started giving errors.. Now little bit worried on installing QPK.
Becoz this is a production server. and i wont get any downtime.. for even an hr also.. to look after reboot after QPK installation and problem solving..

CTK

[T G Manikandan]

0x050700 is c5t7d0.

check the disk whether it has gone bad.

Also check the latest lvm patches


Thanks

[Vijeesh CTK]

HI TGM

I am not getting any errors on the same disk on the other server..( active node in cluster ) but only on passive node ( this is cluster lock PV ).

before installing the STM Dec 2002 . there was no error reflecting. so I think uninstalling STM Dec2002 will solve this issue..

please help me guys n gals

Cheers

CTK