NFS or swap problems

[savus]

Hello,

I have a problem with a HP-UX B.11.11.
Sometimes I have a error message:

"Pid XZY killed due to text modification or page I/O error". This happed once when I tried to initialize a SCSI disk and once with a program that was started from a NFS share.

Do I have problems with the swap space or is this a configuration error in the NFS client?
I runed xstm on the disks to check them for harware failures but it seems that they are OK.

Can somebody please help me ?
Regards,
Stefan

[avizen9]

hello stefan
it doesnt look swap issue, it should be your nfs share network issue.
but also need to cehck your nfs configuration.

what is swapinfo output, thanks,

[Michal Kapalka (mikap)]

Hi,

check this link :

http://forums13.itrc.hp.com/service/forums/questionanswer.do?admit=109447627+1243413425323+28353475&threadId=701754

maybe it helps you.

mikap

[savus]

Hello, here is the reqested output of swapinfo command:

Kb Kb Kb PCT START/ Kb
TYPE AVAIL USED FREE USED LIMIT RESERVE PRI NAME
dev 1048576 0 1048576 0% 0 - 1 /dev/vg00/lvol2
reserve - 266208 -266208
memory 356424 115944 240480 33%

[Michal Kapalka (mikap)]

hi,

check dmesg + /var/adm/syslog/syslog.log if you have some errors regarding swap or NFS not responding.

mikap

[savus]

Hi Mikap,

thanks for the link, I already read the tread - did not helped me. I checked also dmesg - seems fine. A few days ago I had big problems with nfs logged in /var/adm/syslog/syslog.log but now it seems there are ok (I changed in the NSSwitch the first source to local files and not NIS).
Now very rare I see mesages like:

May 25 16:58:53 HOSTxxx vmunix: NFS server (pid1109@/home) not responding still t
rying
May 25 16:59:07 HOSTxxx vmunix: NFS server (pid1109@/home) ok

Regards,
Stefan

[Michal Kapalka (mikap)]

Ok,

this messages are normal, if the NFS server is not available, so NFS should be ok, did you check a performance on the system with glance ?? or some your company monitoring, maybe the system was owerloaded, if our SAP team, dont care about performance we recieve lot of messages related to the SWAP ....

Deferred swap reservation failure pid: 15435

Its really very dificult to say, where s the problem, but the I/O error are mostly related to the HW problems.

mikap

[Hakki Aydin Ucar]

DÄ°d you check them :
/usr/bin/rpcinfo -s
/usr/bin/nfsstat -m

eventually you found nothing makes sense from logs, you can try reboot nfs server.

[savus]

Hello, the outputs are attached in the file.
The nfsstat comand does not return anything.

[Michal Kapalka (mikap)]

hi,

did you check the NFS server which are exporting the filesystems for your server ???

showmount -e hostname

if you have a network team ask them if some network error was happen during the time.

check if all filesystem are mounted.

mikap

[Hakki Aydin Ucar]

yes nfsstat gives not much clues.

try this command if you can see your nfs server(s) mounted :

# bdf | grep
:/nfs/ifep01 104574976 51433936 52373272 50% /nfs/

if not mounted try:

mountall OR
mount -a

if not works consider a reboot your nfs server.

[Dave Olker]

> May 25 16:58:53 HOSTxxx vmunix: NFS server (pid1109@/home) not responding still trying
> May 25 16:59:07 HOSTxxx vmunix: NFS server (pid1109@/home) ok


These are not messages about a remote NFS server not responding. These are messages that AutoFS is having problems. Are you using AutoFS? Do you need to be using AutoFS?

Dave

[savus]

Hello Dave,

Yes I use and really need the auto FS.

I disabled the NFS client (then reboot) for a test, and tryed to initiliaze the SCSI disk from SAM =>
Same error "Pid XZY killed due to text modification or page I/O error" - so I guess that NFS is ok...
Then I think there is something wrong with the swap space or a hardware issue.

Regards,
Stefan

[avizen9]

HI,
what is server model?
if you feel there some hardware issue, take downtime and run full hardware Diagnose test on it, if you are not sure how, then can take help from hp, thanks,

[Dennis Handly]

>once with a program that was started from a NFS share.

If the program is on NFS, then you can get this error if someone fiddles with the program on the server.

Other times I've gotten this error if I just control-C to abort the program on NFS.