backup Windows 2000 cluster

[rklerks]

Hello All,

I can't get our Windows2000 cluster backup on the right way.
Let me try to explain.
Networker server Sun Solaris 7 with version 6.1.3
Client version 6.1.3

node-A
node-B
virtual-server-C

If I try to backup the virtual-server-C it complains about the following:
savefs: path T: by default belongs to client node-B and NOT client virtual-server-C

If I "touch" the file within C:\Program\Files\nsr\bin\pathownerignore on node-B the backup will run but puts the client index on node-B
After this the backup allways goes full ???

Because the index is put on node-B and not on the virtual server.

I also put the Administrator user of both nodes in the Remote access field of the virtual server.

Our NT4 cluster is working like this and is running fine.

Any ideas ??

Thnx in advance...

Ron Klerks

 

[gupel]

Try to put save -c virtual-cluster-C in the "backup command" for the virtual node.

 

[wallace88]

You need to create the C:\Program\Files\nsr\bin\pathownerignore on all physical nodes in the cluster.

 

[rklerks]

Hi,

I tried the save -c virtual-cluster-C in the backup command... too bad, thanks anyway... it tells me the same about the pathownerignore file which I already tried in a earlier stage.
When I touch the pathownerignore file it will backup but puts the client index on the node-level instead of the virtual-node-level.
This not gives the opportunity to recover from the virtual-server.
Any other ideas... I hope so...

 

[cyberted]

Hi,

we have 3 Windows 2000 Cluster and no problems at all.

config:

Node A saveset all
Node B seveset all
Virtuel node C saveset "E:\"
Virtuel node D saveset "F:\"


If you have SP2 of windows 2000 you need to install SP3 or the getlocaldisk patch (or like this). it's a bug in SP2 of Windows.

If you don#t like to install a patch do the following:

Node A saveset SYSTEM STATE: SYSTEM FILES: SYSTEM DB: C:Node B saveset SYSTEM STATE: SYSTEM FILES: SYSTEM DB: C:Virtuel node C saveset "E:\"
Virtuel node D saveset "F:\"


Bye

cyberted

 

[rklerks]

Hello All,

We found the problem.... we didn't fix it yet but we are getting close.
We were able to build a w2000 test cluster and that works!!

On the production cluster they using Veritas Volume manager so they can use dynamic disks within the cluster.
We stripe on EMC level and give them 48 Gb striped disks.
With Veritas they can build and extend their diskgroups/volumes as big as they like.

So Veritas looks to be the problem...

We try to solve it soon.... hope anyone else got a new idea.

Regards,
Ron