Disk problem

[Podgornik Igor]

Hi everyone

I have problem with my disk. We have EVA5000 and superdome with hp-ux B.11.23.My College who works with eva present me a disk with 200 GB of capacity. The task was to extend a vgoradata with additional 200 GB of new disk space.

Command which I use was:

insf -e (to install and assign device files)

pvcreate /dev/dsk/c20t0d7

After that i want to extend a vgoradata and i use:

vgextend /dev/vgoradata /dev/dsk/c20t0d7

And there is start to be a problem. When I press enter to send a command immediate after that a got a message:

Volume group "/dev/vgoradata" has been successfully extended.

That is ok , but I cant get back a prompt? Ok I think maybe is working something in background so I leave it alone and start another terminal session with server. He prompt me with user and password and let me in, but when I type "df-k" a cant get output of command and it look like he's freeze .No other terminal session work after that. Ping command was successfullys and server reply on the ping. So we try to connect on to server over console but he also refuse us to connect . After that we restart it.

In attach. you can find a cfg2html report of the server and you can see that there is problem with vgoradata and c20t0d7 disk .

Any suggestion!?

[IT_2007]

use "-f" option when you do pvcreate.

That disk might not initialized properly.

[Steven E. Protter]

Shalom,

Best way to recognize a disk is to reboot the system.

This runs the command insf -C disk

Which can also beused to properly recognize the disk.

Its also possible that you have a problem with the volume group structure which will force you to rebuild the volume group with the following general process.

vgexport
vgcreate with the -p option for physical volumes.

The problem is lvm assumes you will be spreading your available pe's over 255 disks. Rarely is this the case, so to get a more realistic spread you set the physical limit on the vg lower providing more flexibility. Note the only way to change this is to vgcreate again so set it carefully if you need to.

SEP

[Calandrello]

this environment this in cluster?

[Podgornik Igor]

no it's not in cluster mode

[Calandrello]

Igor this schemes works with NFS client or server? this normal behavior and when NFS is had customer and it lose the communication with the server

[Stephen Doud]

First - if the system hangs, don't just reboot it, perform a TOC (Transfer of Control) to capture a memory dump and have HP analyze it. You don't want a hang to re-occur do you?

2nd - c20t0d7 doesn't show up as a member of vgoradata, so, like a previous writer who noted that the '-f' option was not used with pvcreate, try again with the '-f' option. Then vgextend it into vgoradata and use vgdisplay -v vgoradata to confirm it's finally in there before newfs or extendfs etc.

Also, if this is not a Serviceguard situation, this thread should not be in this category. HP-UX System Administration would be a better category to get more suggestions.