can not list file on some mountpoint

[yunardi]

Dear All,

I have a strange problem on my HP-UX 11.11. I can't do command ls for one mountpoint on my system. It only blinking and give no result. The ls process semms hung, and I can't killed the process, it's pid change to 1.

# ls -l /dmp >>> no result, blinking cursor only
# ps -ef|grep ls
root 25865 25876 0 15:11:42 pts/te 0:00 ls -l /dmp
# kill -9 25865
# ps -ef|grep ls
root 25865 1 0 15:11:42 pts/te 0:00 ls -l /dmp

When I do fuser onthis, there is some proces accessing this.

# bdf /dmp
Filesystem kbytes used avail %used Mounted on
/dev/vg11/lvol1 52428800 11604040 40506464 22% /dmp

# fuser /dmp
/dmp: 20571co 25916co 20274c 23564o 20575c 25865o 22331co 22335c 23602co 20530co 20539c

# fuser /dev/vg11/lvol1
/dev/vg11/lvol1: 20571co 25916co 20274c 23564o 20575c 25865o 22331co 22335c 23602co 20530co 20539c

Please help me, because oracle process is running, I cant reboot server/umount the FS.

Thanks,
EKO

[Pedro Cirne]

Hi,

If you just want to umount FS without a gracefully stop of the processes you can kill all processes acessing FS by:

fuser -k /dev/vg11/lvol1

Post the following:

fuser -u /dev/vg11/lvol1

Enjoy :-)

[Kevin O'Donovan]

Hi,

hanging and not responding is usually an I/O issue. is the filesystem on a local disk?

If you have glance installed, take a look at that, see is the disk being fully used.

Also check sar -d to see if the disk is busy most of the time.

I would suggest looking into how much use the filesystem is getting, and what disks its using. One way to improve performance would be to add another disk and stripe the logical volume across the 2. Do check out that disk though to see if it is just performing badly.

If you're using a SAN with multiple paths, try to load balance across the 2 paths better (assuming you're not using a program to do that already).

The pid of the ls command isn't 1 in your output, the parent process ID is 1. That can happen when you try to kill off a process, it'll be 'adopted' by the system process with pid 1.

hth,
Kevin.

[Eknath · Accepted solution]

Hi EKO,

It seems that the disk or the hardware is not responding which has /dev/vg11/lvol1 data on it. or there could be bad block on the disks whcih you are trying to access.
Run ioscan and see if all disks are present and claimed.

Cheers!!!
eknath

[V. Nyga]

Hi Eko,

try if 'ls' works.
Maybe you have a NFS problem (also if this directory is not NFS!)

Check this thread:
http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=501693

HTH
Volkmar

[Kevin O'Donovan]

You probably already know, but to see what disk(s) the logical volume is using, the command is:

lvdisplay -v /dev/vg11/lvol1 | more

vgdisplay -v vg11 will show you the disks in the volume group, and if there are one or more alternate links, then you have more than one path to the drive. If you're not using any software to load balance, then try to have all the paths used as primaries, e.g. if you have to paths try having path 1 primary for disk 1, path 2 primary for disk 2, path 1 for disk 3, etc.

Kev.

[yunardi]

Dear All,

Thanks very much for the rply.

I have checkd using vgdisplay -v and ioscan there is missing disk from external storage, status is NO_HW :D

# vgdisplay -v vg11
--- Physical volumes ---
PV Name /dev/dsk/c21t0d0
PV Status unavailable
Total PE 2157
Free PE 0
Autoswitch On

# ioscan -funCdisk|more
disk 312 1/0/14/0/0.0.0 sdisk NO_HW DEVICE EMC SYMMETRIX
/dev/dsk/c21t0d0 /dev/rdsk/c21t0d0
disk 313 1/0/14/0/0.0.1 sdisk NO_HW DEVICE EMC SYMMETRIX
/dev/dsk/c21t0d1 /dev/rdsk/c21t0d1
disk 314 1/0/14/0/0.0.2 sdisk NO_HW DEVICE EMC SYMMETRIX
/dev/dsk/c21t0d2 /dev/rdsk/c21t0d2
disk 315 1/0/14/0/0.0.3 sdisk NO_HW DEVICE EMC SYMMETRIX
/dev/dsk/c21t0d3 /dev/rdsk/c21t0d3
disk 316 1/0/14/0/0.0.4 sdisk NO_HW DEVICE EMC SYMMETRIX
/dev/dsk/c21t0d4 /dev/rdsk/c21t0d4
disk 317 1/0/14/0/0.0.5 sdisk NO_HW DEVICE EMC SYMMETRIX
/dev/dsk/c21t0d5 /dev/rdsk/c21t0d5

Regards,
EKO