I have an intermittent issue with one of my clients. Server and client are both 7.6.3.
The client is a DC and it has three IP addresses, all set to "preferred". This is due to historical reasons and can't be changed.
In the hosts file of the client, I have the following entries:
10.10.10.1 blah-dc1 blah-dc1.company.com
In the hosts file of the backup server, I have the same information. My client resource has got aliases of blah-dc1, blah-dc1.company.com, blah-dc1-2 and blah-dc1-5.
Normally, the backup works fine. However, every so often, the backup fails with:
* blah-dc1:DISASTER_RECOVERY:\ 39078:save: SYSTEM error: client `blah-dc1-5' is not properly configured on the NetWorker Server
It is always blah-dc1-5 it fails with.
The solution is to remove the blah-dc1-5 alias and restart the backup. It fails with the same error. Then, I add blah-dc1-5 back in as an alias (i.e. reversing what I've just done) and restart the backup. The backup is then successful.
No other changes have been made and this method has worked several times.
Does anyone have any ideas why this should be the case? I've never looked at the client resource within nsradmin when this happens, but the GUI clearly shows blah-dc1-5 as an alias, so I assume it is still a valid entry.
Could you please try using FQDN for blah-dc1-5 in the client's hosts file.
10.10.10.5 blah-dc1-5 blah-dc1-5.company.com.
remember to restart the NW service once done to take effect.
First some basic troubleshooting is required to get the clear picture on the root cause.
STEP 1: Check the forward and reverse nslookup of client for all the 3 IP and ensure both lookups are working fine.
STEP 2: Either give the FQDN for all the 3 IP or give short name for all the 3 IP (Giving FQDN for 1 IP is a wrong way).
STEP 3: Stop the Networker services of client and rename it's tmp and nsrladb directory. In the /nsr/res/servers file write the IP, Short name and FQDN of backup server. Start the nsrexec service back.
STEP 4: Run client side backup with -pv and -D9 option and give me the error message if error occurs.
Thank you all for your replies.
Both hosts files now have FQDNs for all addresses.
1 nslookup on backup server of blah-dc1 shows IP address of 10.10.10.5.
nslookup on backup server of blah-dc1.company.com shows IP addresses of 10.10.10.1, 10.10.10.2, 10.10.10.5
2 Both hosts files now have FQDNs (I don't know why I didn't have them in the first place)
3 /nsr/res/servers file already had shortname and FQDN of backup server. Services previously stopped and /nsr/tmp renamed (no difference), but /nsr/res/nsrladb NOT renamed.
4 Will run this command at the next failure (this morning's backup was successful)
Yes, the three IP addresses on blah-dc1 are on the same subnet, but the backup server is on a different subnet.
I will keep you posted.
Have you ever tried Rebooting the server.
I had the same issues sometime back, with a different error though, but at last reboot helped me to get rid of this.
Interesting thought - after all, a reboot of a Windows server does cure 99% of all faults
However, this is a DC that's otherwise working for a customer who strictly adheres to the RfC process whereas my fix of removing the alias, re-running the backup and then adding the alias back in has no impact to the customer.