Firewall?

There's a Cisco PIX between the client and the server. I have a timeout set on it of 90 minutes for connections with no traffic. The only one hour timeout it has set is for h225 connections, which I seriously doubt Networker would be using...

For anyone with PIX knowledge, here's the complete set of my timeout settings:

-----
timeout xlate 0:07:00
timeout conn 1:30:00 half-closed 0:05:00 udp 0:02:00 icmp 0:00:02
timeout sunrpc 0:10:00 h323 0:05:00 h225 1:00:00 mgcp 0:05:00
timeout mgcp-pat 0:05:00 sip 0:30:00 sip_media 0:02:00
timeout uauth 0:02:00 absolute
telnet timeout 60
ssh timeout 10
console timeout 0
-----

The sunrpc setting would seem the most relevant, but it's ten minutes...

Oh, as a further note, these systems were being backed up by an EDM backup system before, and didn't experience such a problem with it. Indeed, I've had full backups of ~150 GB drives run up to 21 hours on it without timing out (and the same client is now one of the ones experiencing this issue with the Networker server).

Message was edited by:
TSC

Is 'timeout:' defined in the clients .res files?

Vic

Which is your 7.3.2 version? Jumbo, Build11, Update1?

I believe you have done it, but I would also verify if really there's no packets being dropped by the firewall (happened to me a lot of times and Firewall admins took a long time to find where the problem was)

I would also suggest to check fw logs first for any drops as that would be primary suspect. What is strange here is that you get session which is valid and active killed.

Amaia, original problem is not related to inactive session but rather to active one.

You can put the infinity value "0" in the inactivity time out of the group to solve the inactivity time out issue.

Amaia.

I do not think that the problem is the save set´s lenght, it seems to be a conectivity problem , you can do FTP test from Networker Server to the clients and the other way around, you can check if your Network Cards in the machines are correctly forced 100 full duplex or auto-negotiated (depends if the Network is 100 or giga) and check if the ports on the switch are also correctly forced, you can check the events or messages if solaris for some conectivity errors,,, in the daemon does not apperar no error? either in the messages?

That's good test... another thing would be to go back to host behind fw and try to make a copy and see if it breaks after 1 hour. I have seen that before with fw1 and it was combination of tcp timeout on Solaris box and keep alive on fw1, but only inactive sessions were affected. What is possible here is that you have setting for open sessions longer than 3600 sec to be closed by force. I'm not sure what is the name of such setting, but I do know it exists as I remember once hearing discussion about that (but it was not related to backup rather pissed admin who was in the middle of something when his ssh went down).

There aren't any .res files under the /res folder. From reading in the manual, it looks like those only get created if you're using savepnpc? I'm just using the default setup, not using savepnpc.

Right now, I'm trying a backup of a similarly large filesystem that won't go across our firewall. For historical reasons, this server so far has only been backing up Windows systems outside our firewall. That should let me know if it's a firewall problem for sure, and if it is, then I'll see about getting logs.

Thanks again!

NetWorker 7.3.2.Build.364 Network Edition/50 on the server. I'm not sure where to check on the client, but I installed the client from the same CD as that server version.

I'm currently trying a backup of a machine that doesn't cross the firewall before involving the firewall admins... if that works okay, then it should be a lot easier to get them to believe the problem might be with them.

II have a similar problem and what is happing Networker on there networker server side is changing it TCPIP ports after 30+ gig and the firewall is saying this is not right. Why Networker is changing ports that are in session and are backing up data I have not found out as yet. If Networker would just use the ports it was wing for the first 30 gig I would not have any problem.
You may be running in to the same problem. Sorry I have not found a solution as yet.
My firewall guys are also slow and tell me what the fire wall is seeing. May be its the firewall admin¿s job description to be a bit slow and non-informative.

Are you sure that what fw team is seeing is not NW retry of saveset so it does use new port then? I really can't see why would NW change port at all (it sounds really silly to me) and I believe at that level NW simply uses OS calls and not something by NW specific. But if you put things on both side on debug mode and have those fw people work together with you - you will easily find out what is the truth.

We are seeing the exact same problem, did anyone find an answer to this problem