Problem in Changing retention policy

[pd77]

Hi All,

I was tring to change the retention policy of the HACMP cluster Servers and it throwed the following errors:

"An Alias of this client is already an alias for client 'hostname'.This is not recommended . Reapply only if client is in a cluster configuration" .

My networker Server Version is 7.1.3 and it is running in AIX 5.3 OS .
HACMP client Networker version is 7.1.3 and it's OS is AIX 5.3 .

Can any one please help me how to get rid off this error.

Thanks in Advance,
Pd77.

 

[605]

This refers to your special configuration. Have a look at the client's 'Alias' list to verify it. I suppose this is o.k. because you are running a cluster scenario. So just reapplying should make the message disappear.

 

[pd77]

Thanks for the reply 605, My alais lists are
1. Titian
2. Fully qualified name of Titian
3.newport
4.Fully qualified name of newport

Actual host name is "titian" and this server has two ethernet card. In First card, two IP address has been configured . One IP is pointing to this newport and other one is pointing to the titian. Cluster name of this server is "wmm". I am new to this organisation and i really don't know how they configured this server in the past.

I was doing Client config-> Selected client--> change browse and retention policy.. It gave the error message (specified in my earlier post) and i pressed OK and press Apply botton and it asked me to save or cancel the changes, i saved the changes.But nothing happening.

 

[wamason]

this should have worked...if you you hit ok/save immeditaely after the message, the change should have taken effect. Try once more.

 

[pd77]

Hello wamason...

I tried several times.. But still it is not working...

 

[a060463xyz]

To my knowledge the cluster behaviour changed from NW 7.0 to 7.1.x.
With 7.0 it was possible to give two physical clients the same alias to backup a cluster. This works no longer.

Now you have to create a new clientobject with the name of the cluster-ressource. Add root@node1 and root@node2 to the remote accesslist of the new client.
Dont use saveset All but specify the cluster filesystems e.g. /u01.
This will consume a client connection license.

You can avoid this by customizing the save command :
save -c <node>
<node> can be titian or the name of any other cluster node.
The you will add the savesets of the cluster to the cfi of a physical machine.But this can hamper the ability to perform incrementell or differential backups.
So if you have enogh connection licenses dont customize the save command.