No reverse DNS lookup for server on client

A further addition - we use servers file with all aliases for the Networker server entered and the client has been restarted after editing this. All aliases of the client are listed in the client resource and we have specified both the production and backup network interface fields in the server network interface field without success.

Hello David,

The first error (from the server) says that you can't contact nsrexecd so let's start from that - if you do usual tests from backup server like:
- ping client
- traceroute client
- rpcinfo -p client
- rpcingo -t 390113 client
- echo print | nsradmin -p 390113 -i - -s client

does this work?

Second, how do you the backup - I assume backup network as you mentioned to have that one. Nevertheless, metadata will go over the production - is this open to server/storage node?

If your problem would be PTR record which does not exist for backup server on client you would see that in daemon.log on client side. The error you listed seems to suggest something else is the problem - it looks like backup server can't connect to client name at all. This brings me to the question if your client resource is created to point to the name which is assigned to IP on the backup VLAN/network or not? When you start the backup from server do you see connection being established on the clinet? (for simple analysis use netstat or for advanced one use sniffer to dump TCP handshake to make sure that works). Of course, the easiest way to address some of the concerns you listed would be to set temporary hosts table on client with entry for backup server, but as I said error seems to indicate something else is the issue. Do you have firewall on the client (Windows box) or between server and the client?

Heres what we see - ping and traceroute are fine; the other tests give the following results:

root@ssbak13# rpcinfo -t 390113 ukwap345
rpcinfo: ukwap345 is unknown service

root@ssbak13# rpcinfo -p ukwap345
program vers proto port service
100000 2 tcp 7938 rpcbind
100000 2 udp 7938 rpcbind
390113 1 tcp 7937 nsrexec

root@ssbak13# echo print|nsradmin -p 390113 -i - -s ukwap345
nsradmin: lost connection to server: Unable to send
nsradmin: Rebound to specific service
Program not registered
No resources found for query:
name: ;
type: ;

Any ideas?

My mistake. It should be
root@ssbak13# rpcinfo -t ukwap345 390113


Is this against management or backup network?


Looks like communication is not possible in both directions. Try to sniff where and how communication goes on network level and try to to the same then using some other tool (scp for example). Do you use server network interface for that client? If yes can you communicate with that address from client's IPs? Also, did you try the to verify if you see the same with hosts table as you said before you would try?

rpcinfo -t ukwap345 390113
program 390113 version 1 ready and waiting

root@ssbak13# rpcinfo -p ukwap345
program vers proto port service
100000 2 tcp 7938 rpcbind
100000 2 udp 7938 rpcbind
390113 1 tcp 7937 nsrexec

root@ssbak13# rpcinfo -p ukwap345-bak
program vers proto port service
100000 2 tcp 7938 rpcbind
100000 2 udp 7938 rpcbind
390113 1 tcp 7937 nsrexec

root@ssbak13# echo print|nsradmin -p 390113 -i - -s ukwap345-bak
nsradmin: lost connection to server: Unable to send
nsradmin: Rebound to specific service
Program not registered
No resources found for query:
name: ;
type: ;

I am hoping to get onto the box and try configuring hosts tables shortly.

Don't forget to run all these tests then again from that box against the backup server.

I configured the hosts file on the client and the backup worked fine; I also looked at the other server which supposedly was configured in DNS exactly the same and was working with other Windows clients and took the two differing tests back to the DNS Administrator. This time he checked a little more closely and found he had not put a reverse lookup entry in DNS (as I had originally told him!)