RP4440 running HP-UX keeps hanging

[JM asks]

I have a rp4440 running hp-ux 11.23.

The server became unresponsive over the weekend and had to be rebooted.
The server stayed up for about 5 hours and became unresponsive again.
Engineering looked at it from a hardware perspective and could see no issues.
I did see the BOOT_SLAVE_CPU_FAIL BOOT_BOOT_FAILURE and MEM_SLAVES_FAILED
issued for the one on for the weekend, but they did not think nothing of it.
they looked at the tombstones and were pretty sure they were of no use.

I looked at the MP power settings and everything looked normal.
Nothing showed up during boot, or in syslogs or dmesg.
Any suggestions or comments would be appreciated.

thanks in advance.

[JM asks]

thats hpux 11.23

[Mel Burslan]

I don't quite understand PA-RISC hardware but your error message sounds like you had a hardware failure of some sort over the weekend. If you have a hardware support contract, it might be a good time to use it.

[Rgomes]

What does cstm or event.log say?

[vishnu.khandare]

Hi John,

Can u post lines from syslog.log or SL logs whats the exact error when the system goes down.
It might be an issues with RAM, we can say after checking logs only. Need to investigate to resolve.

Regards
Vishnu Khandare

[sarfaraj ahmad]

hi,

is there any event generated in the server?
please check,
/var/opt/resmon/log/event.log, syslog.log, dmesg and MP logs. r u getting any error in these files? is there any amber LED on the server physically. what is running on the server? when server becomes unresponsive then try to login into console through MP and check at which stage server staying?

[Manix]

Hello John,

If server was hung and you rebooted did
you got valid crash dump generated in
/var/adm/crash or ts99 file in /var/tombstomes
by initiating TOC "transfer of control" through MP.

IF we are rebooting the hung machine it`s a good & recommended idea by HP as well to
get some crash analyzed & narrow down the
SW /Hw issue.

I hope this may help.

Thanks

Manix

[JM asks]

FYI. I was able to swap the drives to another rp4440 and boot it up. The problem stayed with the original hardware, the server still becomes unresponsive when the swapped OS was brought up.

The engineering group is investigating. They are trying to get some good PIM data, but because it takes a TOC to get the system back it may be giving false info, so they say. Nothing shows in the MP logs (since it just happened again and nothing showed).

Also, when the server becomes unusable it becomes inaccessible from either network or MP/console . . . still.

[DeafFrog]

Hi John ,
In crontab :
00,5,10,15,20,25,30,35,40,45,50,55 * * * * /opt/perfscript
00,5,10,15,20,25,30,35,40,45,50,55 * * * * /opt/topscript

/ #more /opt/perfscript
date >> /tmp/perflogs
vmstat >> /tmp/perflogs
sar 1 5 >> /tmp/perflogs
echo ====================================== >>/tmp/perflogs
/ #more /opt/topscript
echo ====================================== >>/tmp/topout
top -f /tmp/topout

This will give a rough idea of whad happened back in time when server chokes.I suspect some malfunction application code.do u have glance/wma ?In that case can you post some glance extract for major metrics(google the forum , as to how)

Regards,