workstation keeps crashing

[Paul Sperry]

The server is a Visualize C3000
running hp-ux 11.11 with up to date patches.
I am thinking that this is a HardWare problem.
I can exersize the hard drive (which is brand new) for 12 hours straight using xstm with no problem. Some times the server stays up for days sometimes only hours, Then it "crashes".
The only error I can get is on the console listed below


netfmt Error reading from input file. (205)
read() returned error code : 5 - I/O error.
Check your input parameters and try again.
Current input file position (in bytes): 128


Nothing in the syslog no crash dump.
When I try to issue a command at the console
nothing happens. All I can do is power cycle the system to get it back up. I have a duplicate server with matching kernel params patches etc that is fine. Any Ideas?
Thanks In Advance.

[Sanjay_6]

Hi Paul,

The netfmt command is used to format the log files generated by the nettl command. So this should not be the culprit. Are you getting the crash dump which you can analyse.

you can alsoe try this tool from software.hp.com to analyze the pim info (stuff in your tombstone/ts99).

http://software.hp.com/portal/swdepot/displayProductInfo.do?productNumber=pimdecode

Hope this helps.

Regds

[Steven E. Protter]

Seems like a bad disk.

dmesg

check /var/adm/syslog.log for lbolt or disk timeouts.

You may need to replace a disk.

If there is no error, run mstm,cstm, or xstm and run the excersize command on all disks. You should find a bad one.

diskinfo might provide some good cross reference information.

SEP

[Sundar_7]

Paul,

Sounds like a bad block for me in the root disk.

use dd to figure out

# dd if=/dev/rdsk/cXtYdZ of=/dev/null bs=1024k

If you see any IO related error messages from dd, then you have a bad hard drive

- Sundar.

[Steven E. Protter]

Duh.

I forgot.

Did you do q4 analysis on the dump? You can send it to HP and they'll let you know what patch to install.

Instructions, hopefully my modifyied version attached.

SEP

[Andrew Rutter]

sounds like it could be a hardware fault, cpu/vrm or memory. try swapping the discs round and see if it stays with the os to verify it.
does it pass the memory tests in stm aswell?
it could also be the firmware, are they different on both boxes? the latest is 5.0 and fixes various issues.


Andy

[MarkSyder]

Hi Paul.

Have you run elm to see if it is posting any reports when it crashes?

Mark Syder (like the drink but spelt different)

[Sยภเl Kย๓คг]

Hi Paul
This seems to be a Hardware problem,mainly to suspect the disk only.However check for the proper SCSI termination and so on
regards
SK