topic Re: SCSI error in Operating System - HP-UX

SCSI error

Supporto Unix — Sun, 12 Jan 2003 12:23:32 GMT

vmunix: SCSI: Unhandled interrupt -- lbolt: 454689803, dev: cb013002

What is the meaning of that line?
I've no strange behaviour of the system.
What can i see to verify that all is ok?

tank you
Luca

Re: SCSI error

Eugeny Brychkov — Sun, 12 Jan 2003 17:18:10 GMT

Luca,
something wrong came on the bus confusing SCSI controller and driver from device c1t3d0. Please check if this deivce is supported and has supported firmware. Please post here: server's model, disk model and its firmware (diskinfo -v /dev/rdsk/c1t3d0)
Eugeny

Re: SCSI error

KCS_1 — Mon, 13 Jan 2003 00:58:20 GMT

Hi,

Look at under files:

system logs:
# vi /var/adm/syslog/syslog.log

# dmesg

# /var/opt/resmon/log/event.log

# ioscan -f

# ioscan -funCdisk

# diskinfo -v /dev/rdsk/c#t#d#

Re: SCSI error

RAC_1 — Mon, 13 Jan 2003 02:54:05 GMT

Re: SCSI error

Supporto Unix — Mon, 13 Jan 2003 09:22:13 GMT

Thanks for your answer.

I can give you these information:
1)
lsdev -c 203
Character Block Driver Class
203 -1 sctl ctl

2)
/dev/dsk/c1t3d0 does not exist,
neither
/dev/rscsi/c1t3d0 exists.

3)
The system is a SuperDome

4)
attach of the complete output of the error in the syslog (18 lines).

How it is possible having an error on a disk or a controller that doesn't exist?
I had think that hw was removed but the kernel maintain a info about it. So the kernel is unable to handle a such interrupt. It's an ipothesys.
What you suggest me?

Thank you a lot!!!!

bye
Luca

For Anil:
the error doesn't say me that there is a timeout error.
lbolt is a signed int that count the number of clock tick since the system was booted.
So error says me that an error occured at that time (lbolt).

Re: SCSI error

Eugeny Brychkov — Mon, 13 Jan 2003 10:25:37 GMT

Luca,
if there's no such device, I would recommend the following action plan: reboot server, and on bootup it will install device files (for new discovered devices) and remove all NO_HW entries (for disconencted hardware).
If it's not acceplable please attach the following info: 'ioscan -fn', 'll /dev/dsk', 'll /dev/rdsk'.
Please let us know on you decision/results
BTW, how often do you see this event in syslog?
Eugeny

Re: SCSI error

Supporto Unix — Mon, 13 Jan 2003 10:38:37 GMT

every time i execute an ioscan...
for example ioscan -fnClan... or ioscan -fnCclic...
or when i execute sam...

Re: SCSI error

Dietmar Konermann — Mon, 13 Jan 2003 10:48:16 GMT

Hi, Luca!

This really looks like something is talking with the initiator using an invalid target. The controller generates interrupts for a device where the driver does not expect to get interrupts from.

I assume this may be related to some hardware problem... may be even cabling/termination or similar. What is attached to this bus... do all those devices work as expected?

BTW, thanks for pointing out that there is no such thing like an "lbolt error". The lbolt kernel global simply indicates "ticks since boot".

So your error occured...

454689803/100/3600/24 ~ 52

... after ~52 days uptime.

Best regards...
Dietmar.

Re: SCSI error

Lalo_Weng — Tue, 14 Jan 2003 03:27:00 GMT

hello,

My suggestion is that if you can make sure that the device is not exist, you can remove the device file with 'rmsf'. Then you check the syslog.log following days to see if the error message is still there.

Re: SCSI error

Supporto Unix — Tue, 14 Jan 2003 13:56:13 GMT

Yes this is the thing I 've done. Infact there was an extern DAT drive that was removed. I removed the special files but error continue.
I thing the best thing is to reboot.

Thank you all

Ciao
Luca