Operating System - HP-UX
1833887 Members
2142 Online
110063 Solutions
New Discussion

Error from the I/O subsystems in syslog

 
Robin_45
Occasional Advisor

Error from the I/O subsystems in syslog

Hi All

I am robin, I have one rp 3440 server, os is version 11i v1, I ran nickel script but it did not complete it hang, finally I killed the process manually. After that when i check syslog I have found a error ( Mar 9 11:22:52 bldhkbk1 vmunix: errors from the I/O subsystem. 21 I/O error entries were lost.). now although server is performing normal but I am not sure about this error, is there any relation with nickel or this is a just coincident with nickel script ?

However, i like to take nickel out and also want to resolve this error

Can anyone help me


Thanks and Regards
Robin
2 REPLIES 2
Steven E. Protter
Exalted Contributor

Re: Error from the I/O subsystems in syslog

Shalom Robert,

dmesg

or....

vi /var/adm/syslog/syslog.log

look for lbolt type errors.

lbolt errors indicate serious problems with i/o, perhaps a bad disk or cable or termination problem.

If you have recently swapped a hot swap disk, then lbolt can be ignored.

SEP
Steven E Protter
Owner of ISN Corporation
http://isnamerica.com
http://hpuxconsulting.com
Sponsor: http://hpux.ws
Twitter: http://twitter.com/hpuxlinux
Founder http://newdatacloud.com
Andrew Merritt_2
Honored Contributor

Re: Error from the I/O subsystems in syslog

The message you refer to is the second of a pair, when the problem starts you should see something like:

Aug 18 03:13:40 hsc100m3 vmunix: DIAGNOSTIC SYSTEM WARNING:
Aug 18 03:13:40 hsc100m3 vmunix: The diagnostic logging facility has started receiving excessive
Aug 18 03:13:40 hsc100m3 vmunix: errors from the I/O subsystem. I/O error entries will be lost
...

That usually means a problem with hardware. To see more about what caused it, run 'logtool' in STM, there should be errors logged from the device in question, since some should have got logged before the buffer diaglogd reads from filled up. Was the first message seen before or after the nickel script started?

You may also get EMS events logged in /var/opt/resmon/log/event.log from this device, but not always (as the message in syslog says, diaglogd is not able to process the large number of I/O events, so events will be lost, and if the events lost are those which would have triggered an EMS event, then of course you won't get those either).

What is the 'nickel' script? If it's exercising hardware in some way, and that hardware is faulty, then that could well trigger the errors.

Andrew