"Connection refused" trying to backup SQL Cluster 2

[adamcolliss]

Hi

I'm trying to back up a MS SQL Cluster (Sql 8.00.194/MDAC 2.71.9040.2) and I keep getting 'Connection refused' errors whether backing up the filesystem or SQL databases. The networker server account has access to both the SQL Cluster and filesystem. Version of networker server 7.2.2 Build 422, SQL Client 4.0.0.113.

I also have a another Cluster, which filesystem does backup correctly, but the SQL backups fail (they just hang, no error). Trying to launch the SQL Backup program on the SQL Server causes an error - 'nwmssql: sql logon failed'.

Any advice would be greatly appreciated.

 

[NWNinja]

Hi adam!
Regarding the "connection refused" errors this most likly has to do with the servers file on the client side. Make sure this file contains all known hostnames (short + FQDN) of the backup server. Using the ip chould also work fine. Make sure you end the file with a enter to avoid junk charachters at the end of the last line. Ofcourse you also need to verify the name resolution by using nslookup and so on. If not using DNS, makse sure the hosts files contain the correct entries.
On some occasions, you may also need to create the "pathownerignore" file on the client side. This is sometimes needed when backing up clusters.

Regarding the "logon failed" error, this simply means that the account you are starting the NetWorker services with, does not have enough privilages to access the SQL DB. Just start the services with an account that has access to SQL as well as the filesystem. Local System Account, wich is default, does usually not have enough privilages.

Regards
Maverick

 

[adamcolliss]

Hi
Thanks for the information, I have discovered the first Cluster (with the connection refused errors) has an additional network address for a heartbeat monitoring system. When I do a test savegrp for this Cluster I can see that Networker is resolving RDNS for the server to the address of the heartbeat monitoring network (which isn't linked to the same network as the Networker backup server). It also tries the NSLookup on the backup server resolves the Cluster to its correct DNS however.
Could this be causing the connection refused errors, and if so how can I force Networker to use the correct address? I've tried adding the correct address to the Hosts file and to the aliases of the client object in the server administrator but it still resolves the RDNS incorrectly.

 

[adamcolliss]

Managed to stop RDNS resolving to the incorrect address (someone had made changes to the DNS Server without making me aware). I've got the first cluster working correctly now by bypassing DNS and using hosts files and changing the servers file in RES.

 

[a060463xyz]

you should prevent the nics of the intercluster link from registering their ip-dresses with the dns-Server :
select the correct nics properties, select the tcpip-protocols properties, press the advanced button.
Now select dns and remove the mark in the checkbox "Register Adresses..."

After this only the Adresses of the public LAN of the Clusters should be visivle via dns and you can get rid of the local hosts files.

 

[adamcolliss]

Hi

Im still getting problems with the second SQL database (filesystem is fine). I can do a successful savegrp -p on the group the SQL object is a member of, but the scheduled backup just hangs and I have to manually stop the group which gives the error 'index was never started'. Trying to run the SQL client on the SQL Server still gives the error 'nwmssql: sql logon failed' as well. I am assured the user running the SQL Client has all of the required permissions to the database.