problem mounting disk on cluster

[nipun_2]

Hi all,
This is the setup
Server - (contains disks used by all the nodes)-
Node1 - (contains disks used by all nodes)
Node2 - Satellite
Node3 - Satellite

Node1 had been turned on today only. All the other nodes were already on. Normally all the disks on Node1 are as a volume set with Disk2 logical. This is how everyone accesses data on the disks on Node1.

However, currently when I booted up Node1 it showed Disk2 on that node. But when I tried to access disk2 from other nodes -"there is no disk2"

I did sh dev and all the Nodes (except Node1 )showed

Node1$DKB0 MntVerifyTimeOut
Node1$DKB100 MntVerifyTimeOut
Node2$DKB200 MntVerifyTimeOut



Disk2 logical points to Node1$DKB100:

what should be my set of steps so that I can make it available as disk2 right now. I do not wish to make it a permanent change. So please let me know how I can undo the changes as well.

Note: Node1 containing disk2 is working fine and I can access all the data through there.
Thanks in advance
Nipun

[Uwe Zessin · Accepted solution]

So, NODE1 was shutdown, but the disks were not dismounted from all other nodes. The system was down for longer than system parameter MVTIMEOUT.


You can try to dismount the disks on the other nodes:
$ dismount /ABORT NODE1$DKB0:

and then manually remount, unless there are still some files open. Sometimes its possible to get rid of the mount with:
$ dismount /ABORT /OVERRIDE=CHECKS NODE1$DKB0:

but there are situations were your really have to reboot.

[Jan van den Ende]

Nipun,

set SYSGEN param MSCP_LOAD to 1 (one),
and SYSGEN param MSCP_SERVE_ALL to 5

.. you will need a reboot of node1

Alternatively, you can assign non-zero alloclass to each clusternode.
That will however require ALL nodes to reboot.

Proost.

Have one on me.

joe

[Jan van den Ende]

Oops -

should have read better before typing ahead.
Ignore my post - re-read Uwe..
Sorry.

Proost.

Have one on me.

jpe

[nipun_2]

As always...thanks a lot Uwe