Domain name change

Have you done checks with nslookup too? Have you cleared your DNS cache which may still be referencing the old name? Also use route command from the client to see whether the client is trying to communicate over the correct network. Also check that the aliases field for both the client and the client of the Networker server are correctly configured.

Nslookup and route, and also tracert are all fine. I flushed the DNS on both sides, but still get that error.
The aliases are configured fine as far as I could verify. Every communication between the hosts work fine except by Networker backups (and maybe recoveries).

What is stranger is that manual backups save the data but give that nsrexecd error.

If we are seeing this situation then we know it is a communication issue; the nsrexecd process is only used when the Networker server wishes to initiate a backup on the client, so if it has a problem contacting this then it has a problem starting the backup!

Try running a probe:

savegrp -vvvp -c client group

and see what sort of information we get back. If that does not give enough information then replace -vvv with -D9 for debug mode...

It may also be useful to know what kind of relationship exists between the two domains...

Is there anything in daemon.log on server and client at the time when this happens which is related to described situation? Also, did you try to restart NW services on the both end after the change?

He stated this:

So I added the winserver01.test2.com in the hosts files and in the aliases in client configuration.

But I was thinking, did you remove the old alias from the client configuration (I think it would not be happy if you left it there but not 100% on it)

Could we possibly see a breakdown of the configuration:

name of client
name of server
aliases of client
aliases of server
hosts file entries on client
hosts file entries on server

Output for savegrp -vvv -p would also be useful here as we would be able to see where Networker decides it has a problem!

Did you fill in alias field on client with short and long name of a client?

OK, after some time I got the output of savegrp:

savegrp -D9 -G DiarioFS
04/06/06 08:21:07 savegrp:
Calling clnttcp_create function
04/06/06 08:21:07 savegrp:
Calling clnttcp_create function
04/06/06 08:21:07 savegrp: socket bound to port 18394
04/06/06 08:21:07 savegrp: Adding port range (reserved): 600-1023
04/06/06 08:21:07 savegrp: socket bound to port 978
04/06/06 08:21:09 savegrp: can't get my primary domain controller info, The netw
ork path was not found.
04/06/06 08:21:12 savegrp: RPC Authentication: error in LookupAccountSid: Invali
d argument
04/06/06 08:21:12 savegrp: RPC Authentication: an old auth server issued the cre
d, it is not checksummed
04/06/06 08:21:12 savegrp: lgto_auth for `nsrd' succeeded
04/06/06 08:21:12 savegrp: RPC Authentication: Client successfully authenticated
using AUTH_LGTO
04/06/06 08:21:12 savegrp: There is no group named 'DiarioFS'
04/06/06 08:21:12 savegrp: pathname: 'C:\Program Files\Legato\nsr\tmp'

Of course there is a DiaroFS group... That LookupAccountSid error is in KB regarding DNS, but as I stated before, all communication using short or long names are fine...

OK, let's focus on one client only - resolving one client we will resolve others too.

From server do following:
echo print | nsradmin -p 390113 -i - -s client_resource_name and give us output here.

Then go to client and do ipconfig /all and give that output.

Then again from client give us hosts table and DNS output for back and forward resolution for each IP on the client and server.

At the end, give us nsradmin output for client name and aliases associated.

Oh yes, nslookup output (and hosts) from server for all those IPs is also good to have so... if you give us that too.

I don't remember seeing NW using by default such low port range before - what ports do you have configured?

Let's focus on the client - did you check the hosts table? Did you enter the new hostname.domain in alias of short_name client resource? What do you see in daemon.log on the client? You said that nsrexecd error is given when doing client initiated backup - what kind of error? Was this domain name changed only for clients or for the server too?

Do you think that low port range could give us some problem during the backup? Such slow performance or something like that...

Can you give us the same for client? And explain a bit what is production and what is backup LAN (IP and hostname)?

Dedicated as backup network?

OK, so server is ping brqoc-s-0000017 (10.100.5.3) and client is abbsv006 (10.100.5.4). I guess abbsv006.americas.abb.com used to be abbsv006.br.abb.com before, right? Given that you use dedicated network for backup there is remaining question of production network (used for metadata). So you will need to involve that in tests as well. Do you use backup server interface for the client resource and if yes to what does it point (along with that check the route set for the client).


So, clients is not able to contact itself via nsrexecd. I wonder if you see the two nsrexecd processes in task manager. Can you check that?

Also, try from client to test rpc communication with nsrexecd (rpcinfo -t abbsv006 390113 1) and see if you get response. You can do the same then for portmapper service.

Actually, route is for the client right? Can you give us rpcinfo for it?

I left the default port ranges, didn't changed it. Only one client was moved to another domain. This is what I have, brqoc-s-0000017 is the server, and abbsv006 is the client with the problem, and remember that backup is done throug a dedicated Network.

Hosts file from the NW Sever:
10.100.5.5 abbsv001 abbsv001.br.abb.com
10.100.5.4 abbsv006 abbsv006.americas.abb.com
10.100.5.7 peoplep peoplep.br.abb.com

Hosts file from the client:
10.100.5.3 BRQOC-S-0000017 BRQOC-S-0000017.br.abb.com
10.100.5.4 abbsv006 abbsv006.americas.abb.com
138.226.1.10 americas.abb.com

And some test commands from the client:

(1) ping brqoc-s-0000017

Pinging BRQOC-S-0000017 [10.100.5.3] with 32 bytes of data:
Reply from 10.100.5.3: bytes=32 time<10ms TTL=128
Reply from 10.100.5.3: bytes=32 time<10ms TTL=128

(2)ping brqoc-s-0000017.br.abb.com

Pinging BRQOC-S-0000017.br.abb.com [10.100.5.3] with 32 bytes of data:
Reply from 10.100.5.3: bytes=32 time<10ms TTL=128
Reply from 10.100.5.3: bytes=32 time<10ms TTL=128


And the commands from the server side:

(1) ping abbsv006

Pinging abbsv006 [10.100.5.4] with 32 bytes of data:
Reply from 10.100.5.4: bytes=32 time<10ms TTL=128
Reply from 10.100.5.4: bytes=32 time<10ms TTL=128

(2)ping abbsv006.americas.abb.com

Pinging abbsv006.americas.abb.com [10.100.5.4] with 32 bytes of data:
Reply from 10.100.5.4: bytes=32 time<10ms TTL=128
Reply from 10.100.5.4: bytes=32 time<10ms TTL=128


And I'll run the commands suggested as soon as I can. The error I receive was during a manual backup, as follow:

c:\>save -vvv -s brqoc-s-0000017 -b teste c:\temp
NetWorker: Cannot contact nsrexecd service on abbsv006.americas.abb.com, Service not available.

But backup finished fine in this test, and that's one thing I only saw today....

Well, I wrote too much...

It depends really... if you have firewall on the box it might. NW by default should not use such low port range. I know earlier version did have problems or they would use very low ports if unable to bind to default one (7937 for nsrexecd and 7938 for portmapper; this last one could be even your issue in case that something else blocked that port and thus the error message you see - use tcpview for sysinternals to see what process is using those ports).