error 126 on HPUX

[Dang Huu Tung]

Hi,
Is the below error message from syslog? how can I fix it?

Thank you

SCSI: Async write error -- dev: b 31 0x050500, errno: 126, resid: 65536,
blkno: 8911488, sectno: 17822976, offset: 535429120, bcount: 65536.
blkno: 8911480, sectno: 17822960, offset: 535420928, bcount: 8192.
blkno: 8911472, sectno: 17822944, offset: 535412736, bcount: 8192.
blkno: 8911464, sectno: 17822928, offset: 535404544, bcount: 8192.
blkno: 8911456, sectno: 17822912, offset: 535396352, bcount: 8192.
blkno: 8911448, sectno: 17822896, offset: 535388160, bcount: 8192.
blkno: 8911440, sectno: 17822880, offset: 535379968, bcount: 8...
LVM: Recovered Path (device 0x1f030500) to PV 0 in VG 1.
SCSI: Read error -- dev: b 31 0x030700, errno: 126, resid: 8192,
SCSI: Async write error -- dev: b 31 0x050500, errno: 126, resid: 8192,
LVM: Performed a switch for Lun ID = 0 (pv = 0x000000004849e800), from raw device
0x1f030700 (with priority: 0, and current flags: 0x40) to raw device 0x1f050700 (with priority: 1, and current
flags: 0x0).
LVM: vg[3]: pvnum=0 (dev_t=0x1f050700) is POWERFAILED
LVM: Restored PV 0 to VG 1.
LVM: Restored PV 0 to VG 3.
SCSI: Read error -- dev: b 31 0x050500, errno: 126, resid: 2048,
blkno: 8, sectno: 16, offset: 8192, bcount: 2048.
LVM: Recovered Path (device 0x1f030700) to PV 0 in VG 3.
LVM: Performed a switch for Lun ID = 0 (pv = 0x000000004849e800), from raw device
0x1f050700 (with priority: 1, and current flags: 0x0) to raw device 0x1f030700 (with priority: 0, and current f
lags: 0x80).

[Mohanasundaram_1]

Hi,

Looks like this is a storage LUN which is causing this error. If there are many powerfail messages like this, consider changing the pvtimeout value of each LUN to 90 or 120. pvdisplay -v will show the pvtimeout value at present.

You can change this online and observe the problem.

I checked in /usr/include/sys/errno.h and did not find an err no.126. So I am not sure what that is.

But I guess, since different LUNs are giving this message, it is a good idea to check the pvtimeout.

Cheers,
Mohan.

[Bill Hassell]

I seem to remember that there are several 100-130 error codes that are defined by non-HP code, specifically volume managers and high end disk arrays. You'll need to use the disk array's diagnostics to find the problem.

[Naveej.K.A]

hi,

The error looks like there had been some problem with the controller on c5 and there has been a switch over to its alternate path. c5t0d5 and c5t0d7 has failed with scsi errors. and a switch over has been done to the scsi device c3t0d5 and c3t0d7.
It could be an issue with the cable connecting the scsi card and the storage device also

with best wishes
Naveej

[Dang Huu Tung]

Hi,
Thanks for your replies,
One more information, our storage device is VA 7100 with 7 external disks and we have two VA controllers, when I use ioscan these controllers status is CLAIMED.
Please give us your suggestion.

tks

Tungdh

[Naveej.K.A]

hi,

VA7100.....7 hard disks would form a single Redundancy group and hence a single Raid volume, You either have a RAid1+0 or an autoraid. You might be having several LUNs and each LUN will be assigned a device file like /dev/dsk/c5t0d0 and will have a backup link through the other VA controller.

when you use ioscan you can't see the controllers, but the LUNs...you can use the amrcfg and armdsp commands to the status of the controller, the battery, the fibre cable and so on.. get on to the manual pages of these commands and you will get a better idea.

with best wishes
Naveej

[Dang Huu Tung]

Hi,
Sorry for sending message late. We are trying to add one more 73GB Hard disk to VA, but that problem remains unchanged. The attached file is VA log that I collected from our system. One more thing, our system use autoraid mechanism with RAID 1+0 and RAID 5.
Could you please analyze this logs and give us your suggestions.

Thanks alot

Tung

[Dang Huu Tung]

Hi,
Here is another log files that I collect from my our system.
Take a look pls

Tung

[Dang Huu Tung]

Hi,
And here is Oldsyslog files that I collect from my our system.
Take a look pls

Tung

[Dietmar Konermann]

The error 126 is not really top secret...
#define EPOWERF 126

So essentially this error happens when an IO times out (for LVM: does not complete within the specified pvtimeout, 30 secs by default). Go ahead and log a call with your response center to get the hardware checked. Maybe a longer pvtimeout is able to cover the symptom.

Best regards...
Dietmar.

[Hai Nguyen_1]

Tung,

Use the GUI program "launcher" to diagnose your VA7100 (The command line program "armdisp" can display the status of the array as well.) Visually check the system status LED in front of the array. If it IS amber then the array is having some issue.

Hai

[Dang Huu Tung]

Hi All,

Thank you very much for your help and your support.
We tried to change PV time out and upgraded firmware for VA controller and Disks, now there is no error generated and the performance is acceptable.

Thanks again and regards

Dang Huu Tung

[Mohanasundaram_1]

HI ,


Good to know that your problem is resolved. Thanks for the update.

Would you also be kind enough to assign points to people who have helped you? This is as a courtesy and it will enable you to get responses in future for your questions.

Cheers,
Mohan.