Questions on adding sub-repository to a ovm 2.2 pool.

Question

I am adding a sub-repository to a single server pool. The root repository is on an iscsi lun that is multipathed.
I logged in to the new target using iscsiadm and was given new devices sdd and sde. I obtained the scsi_id of the new lun and updated my multipath.conf file as follows. The friendly name ovmdevrepos3 is the new iscsi target. Ovmdevrepos2 is the original repository.
multipaths {
multipath {
wwid 36090a03810777ced37f214010000b080
alias ovmdevrepos2
multipaths {
multipath {
wwid 36090a0381077fca09efe94010000309b
alias ovmdevrepos3
I did a service multipathd reload and now running a multipath -ll shows the following.
mpath1 (36090a03810777ced37f214010000b080) dm-1 EQLOGIC,100E-00
[size=600G][features=0][hwhandler=0][rw]
\_ round-robin 0 [prio=0][active]
\_ 3:0:0:0 sdc 8:32 [active][ready]
\_ 4:0:0:0 sdb 8:16 [active][ready]
ovmdevrepos3 (36090a0381077fca09efe94010000309b) dm-2 EQLOGIC,100E-00
[size=800G][features=0][hwhandler=0][rw]
\_ round-robin 0 [prio=0][active]
\_ 6:0:0:0 sdd 8:48 [active][ready]
\_ 5:0:0:0 sde 8:64 [active][ready]
I don't know why it is now referring to the first device as mpath1 and not ovmdevrepos2. Prior to running the multipath reload, I had:
/dev/mpath0
/dev/ovmdevrepos2
/dev/ovmdevrepos2p1
Now I have:
/dev/mpath1
/dev/mpath1p1
/dev/ovmdevrepos3
I don't understand why the devices are getting renamed and if this will impact my environment negatively. I have not rebooted or restarted the multipathd service yet.
I have not tried to create an ocfs2 filesystem on the new device yet. I'm not even sure which device to reference when doing so.
Any help in understanding what's going on here would be greatly appreciated.

Answer 1

Hi,
You need to follow the steps mentioned in the link:
http://middlewaremagic.com/weblogic/?p=971
<b><font color=maroon>1. Looking at server 02, I do not see the domain installed. The domain exists only on 01. So when I do the install of Weblogic 10 on server 03and 04, do I bypass the creating of the domain? </font></b>
It is always a best practice to have the same domain directory absolute path on every remote Managed Server Box. BEA has provided a utility to make sure this happens. The utility is called as "Pack/Unpack utility". Which is described in the Step7). of above Post.
<b><font color=maroon>2. Can I copy the license from 02 to 03 and 04?</font></b>
Yes, If u have the "licence.bea" file then u can use it in all Managed Server Box in the same location where it was present in the original location.
<b><font color=maroon>3. So what steps do I need to do? Install weblogic and JDK on 03 & 04, go into admin server on 01 and add new managed servers unix machines, then add them to the existing cluster. Then start nodemanagers on the new servers and then start the managed servers in the admin console. Sounds easy, but im sure its not. What am I missing?</font></b>
Yes U need to Install JDK and WebLogic on all the ManagedServer remote Boxes. Then u can create ManagedServers with the IP Address mentioned of the remote Boxes where u want to run them. U need to Create a Machine sothat u can assign NodeManagers to your Managed Servers for remote Start/Stop purpose. But u need to make sure that u follow the Step8) Mentioned in the above link to Enroll your NodeManager from each remote ManagedServer Box.
NOTE: to run nmEnroll() WLST function u must have the same Domain Directory structure in all the ManagedServer Physical Boxes. U might have already done pack/unpack of your Domain...while doing this step.
Then anytime u can add your ManagedServers to the Clusters.
Thanks
Jay SenSharma