save: shutdown failed

[warlock0199]

Hello Everybody.

I have been struggling with this problem a while now. So I hope someone has the answer for me.

I have a win2003 bck server (7.4) and a Win2000 cluster Clients (7.4). They are on a different subnet and behind the firewall and in the DMZ.

I receive the following error if I want to run a backup of the cluster. All the backups start fine. But all backups that take longer than a hour to backup fail with this error:

save: RPC error: RPC send operation failed. A network connection could not be established with the host.
save: shutdown failed
save: save completion time: 7/23/2007 3:30:05 PM

It doesnt matter if I do a manual or scheduled backup. I get the same errors!

All the name resolution with ping and nslookup works fine both ways.

I can use telnet to connect with the reports. And the firewall personal says that the ports are open.

Does anyone have any idee what I can try to do?

 

[605]

It looks like a temporary network problem. From the Error Message Guide:

The Remote Procedure Call (RPC) services are unable to communicate with the server software, because the local or remote client network is down. Communication between the server and client was lost while the server was processing the request.

Unfortunately hard to investigate.

 

[Rif123]

Hi,

check with your fw personal again. It is common that firewalls disconnectes tcp-sessions after 1 hour of inactivity. I know PIX and FW1 have configuarable values for this. Fw personal should be able to tell if there are dropped packets in the firewall for illegal/unrecognized/closed tcp session IDs.

You can also try a NetWorker environmental variable. Can't remember exactly on top of my head, but it is called something like nsr_keepalive. Check the admin guide for this. I have not had much success with this though. What used to work at least up to 7.1.X was to set the group properties to verbose. That makes NetWorker more chatty and thereby holding the tcp-sessions alive. Not sure it will work in 7.3+ though since NetWorker was heavily redesigned.

 

[warlock0199]

Thanks for the reply guys.

Do you know the PIX firewall setting or command that change that 1 hour setting? I asked them again and they assure me that it is not the firewall.

I also used the NSR_KEEPALIVE_WAIT=300 setting. But it didn't help.

I will try to change the verbose setting and see what I can get from that.

 

[warlock0199]

Hello everybody.

I found the fix for the problem!

On the PIX firewall, there is a setting to change.

Change the Timeout Conn setting from the default 1:00 to 3:00 or 5:00 hours. Depends on how long your backups run.

Hope it helps somebody.