Strange.. if everything is configured correctly this should not happens (unless some kind of bug is involved). Can you post your configuration as nsradmin output?

True especially when the first group runs on storage nodes and the successive ones with same config & for the same clients do not.

Do you want me to capture anything specific on nsradmin?

Client, pool and group configuration and util file used by backint for both (correct and incorrect) backups. I assume this is all scheduled backup?

The outputs captured by nsradmin could not be complete, the screen scrolls through and does not capture all details.

Observation - There are three SN for DDS. One of them i.e. the development server is a standalone and it easily backs up on SAN using rd=. The file servers are on cluster. What we are doing is that we have configured SN(client resource) for the virtual name and the devices under the jukebox are with the physical names.

This has been working fine for the file servers for the last one month. Earlier we re using the same for NW 6.x and it has been working for years.

Is it possible that it is not able to resolve the physical host name under the storage node field. Infact, I tried changing the SN field under the client(VN) to the corresponding HN, FQDN & IP of the physical node but nothing helped.

Do nsrsup and copy relevant data. I'm not sure if I got 100% your config, but did you try to use curphyhost for clustered server which is cluster between SN if I understand correctly?

No I have not tried the curphyhost option. nsrsup creates a 90-95 MB file which I have never been able to open.

In this cluster we have two physical and two virtual nodes. E: of phyA is accessed as E: of virtualA. F: of phyB is accessed as F: of virtualB. Actually E & F aremounted from a HP MSA Storage. So if PhyA fails, users will aceess both E & F from the VirtualB.

Let me try and capture informaiton tomorrow.

Well, without getting into NW configuration info it's going to be hard to check and suggest anything.

It means that cluster is not being recognized correctly or not configured correctly within NetWorker configuration. What NW version do you use?

A new observation - It is directing the backup of the virtual client to the Storage node mentioned in the Storage Node resource of the respective physical client. The problem is that it then stores the Save Sets in the name of the physical client and not the virtual client.

I am not sure if deleting the client resource for the physical client would help. Would it still use the device configured for the physical nodes?

Just to add, using curphyhost did not help at all - it went to LAN.

The conclusion is simple - it is directing backup to the device being used as Storage Node for the Client Resource of the Physical client - Any help would be great.

7.3.2 Jumbo 11.

But for a long time we have not done any changes to the cluster config. of NETworker. All we had done was that for some other maintenance jobs, we had reboot our servers which includes these cluster servers as well.

Even after the reboot the first backup went off well on the Fibre properly but all backups beyond that are creating problems.

Just to make you aware there is now jumbo of the jumbo (called jumbo update 1)

One thing that I have see in the past is that when you reboot clusters NIC order changes within MS and that causes hell for some reason (eg. ASR restore is touchy to what is first and what is second network connection). I'm not saying that is the issue, but that was first thing I thought when you mentioned reboots. I would double check hosts and DNS just to be sure anyway. You can also run save from them in debug mode to see what is going on really.

SOme one from Support told me yesterday but I believe we still do not have it on the Powerlink to download.

What do you want me to check in DNS? Backups were on Fibre so if DNS things have changed, backups in any case were Fibre based. Does this effect?

You have it at default update location:
ftp://ftp.legato.com/pub/NetWorker/Updates/732JumboUpdate1/

Name convention make affect alias host list. If something is mixed or NW thinks virtual is the same as phy you may get issues like that.