process running and getting killed

[Murali_16]

Dear support,

I have a problem.
I have HP unix 11i (B.11.23 U)
I have tried to extend the vol group by adding another harddisk. later I started the process to create a logical volume. however, the process seems to be struck as it is running for more thatn 10 hrs.
Please advise as how to end this process namely
root 29917 1 0 17:41:15 ? 0:00 mkfs /dev/vglog/rlvoraorigA 2621
I have tried to kill using the kill -9
command , but it did not work.

Please advise as I am struck up in the middle of my work.
Thank you in advance
Murali

[Steve Steel · Accepted solution]

Hi


looks as though it did nothing and is waiting for IO

Then it will not detect the signal and thus only a run level change or reboot will kill

Does it show in top as active


Steve Steel

[harry d brown jr]

the disks are what?

Striping, mirroring, raid what?

volume group looks like what?

The commands you used to add disk to VG and to extend FS ?

live free or die
harry d brown jr

[Kent Ostby]

When a command is doing intensive I/O, it will often mask interrupts so that it doesn't leave data in a corrupted state.

That is probably what is happening here.

Unfortunately, if you really are having a problem with it, it will simply hang unless you reboot the box.

[Bill Hassell]

lvcreate is a very short process, perhaps 1-2 seconds even on a busy machine. Since kill -9 fails, this is definitely an I/O problem. kill is a signaling system and doesn't really remove the process. Instead, the process will see the signal and respond to the termination request. But a process that is waiting on I/O is not running. Since disks respond quickly, this problem is due to a stucj I/O which is a failure in one of the disks, probably the one you added. You can check syslog.log to see if there are any errors (just to confirm the problem) but there are certain handshake failure that can't be logged and just hang.

There's nothing you can do at this point except to reboot. I'd recommend going into single user mode and diagnosing the new disk. Start with diskinfo. If that fails, the disk is useless, get a new one. If that's OK, run dd across the disk surface as in:

dd in=/dev/rdsk/c1t2d0 of=/dev/null bs=128k

(substitute your disk's device name for c1t2d0) This should complete without errors. Ideally, you can watch the disk activity light and if it gets stuck, you'll know the disk is bad.

[Murali_16]

Thank you Steve.

I rebooted the system and the problem is solved.

Thanks again,
regards,
Murali

[Murali_16]

The problem is solved and I am thankful for the prompt and useful replies I got.