Unsolved
This post is more than 5 years old
1 Rookie
•
3 Posts
0
6958
Cannot open a save session with NetWorker server
I am having issues with several servers backed up by EMC networker. The server is version 8.2.1.0 build 681 power edition/137 and I chose 1 client to T-shoot and it is 7.4 SP5 build 726. Running on 2008 R2 x64. I admit I am a total newbie so I am pretty lost here. The previous engineer left and noone was taught anything but how to change tapes. The info from the logs are:
cr-***-a01.tlr.thomson.com: save failed to authenticate with nsrmmd on host cr-***-a01.tlr.thomson.com: Timed out.
107204:save: 'Backup renamed directories' disabled, renamed directories will not be backed up
98519:save: Unable to setup direct save with server cr-***-a01.tlr.thomson.com: Failed to connect to storage node on cr-***-a01.tlr.thomson.com: save failed to authenticate with nsrmmd on host cr-***-a01.tlr.thomson.com: Timed out.
107204:save: 'Backup renamed directories' disabled, renamed directories will not be backed up
98519:save: Unable to setup direct save with server cr-***-a01.tlr.thomson.com: Automatically cancelling the save request. 'save' was unable to start the active session in 86564 seconds. The elapsed time is over a day. This may be due to the date/time being set incorrectly on the client..
90018:save: Cannot open a save session with NetWorker server 'cr-***-a01.tlr.thomson.com': Automatically cancelling the save request. 'save' was unable to start the active session in 86564 seconds. The elapsed time is over a day. This may be due to the date/time being set incorrectly on the client.
cr-**sql-a03r:index: retried 1 times
I have been checking times on the clients are they within 2 minutes and I can ping and NSlookup the client names. The backups reach 100% and sit there for 2 days and then fail. also the save set that is failing is the :index, everything else says it is successful.
Kleinenbroich
86 Posts
0
May 16th, 2017 06:00
Hi ratnfested,
according to my understanding, there is an incompatibility in the authentication between NetWorker client and NetWorker server.
This from the EMC NetWorker 8.1 rlease notes:
"By default, NetWorker hosts and daemons use the nsrauth mechanism to authenticate components and users, and to prevent user and host impersonation. You can configure compenent authentication to use oldauth, but this authentication method is not secure.".
I expect, that NetWorker server is using the "nsrauth" mechanism and the client "oldauth".
According to my understanding, you've two options:
To change the auth setup check the NetWorker "Security Configuration Guide" (search for oldauth).
Regards
Michael
bingo.1
2.4K Posts
0
May 16th, 2017 07:00
The pain is if you you must change tapes. I hope you have a library to do that automatically. As you run Power Edition my guess is that such has been installed. Please check the 'alerts' area for potential tape handling requests.
You said: ".. also the save set that is failing is the :index, everything else says it is successful."
Index backups are running locally on the backup server - consequently a data transfer should be no issue.
However, if setup best practices, their data is usually sent to a separate pool. So if you only have one tape drive you or the library should swap the tapes. So ance again - look at the alerts box, it should tell you more.
ratnfested
1 Rookie
1 Rookie
•
3 Posts
0
May 16th, 2017 07:00
Hi. Thanks for the response. I did check the documentation nd the local agent settings are all set to 0.0.0.0/0,nsrauth/oldauth. I assume that means they can use either. And the jobs to mostly complete it is just the .Index that is failing.
ratnfested
1 Rookie
1 Rookie
•
3 Posts
0
May 16th, 2017 08:00
I believe Bingo has nailed this one. I was starting to suspect maybe the indexes were written to another media pool. You said the right things to get me to check the logs under the devices and sure enough 19 jobs were waiting to write to a media pool that was full up. So I labeled a couple more Virtual tapes for that media pool. I think things are coming around now. Thank you all very much for the quick responses.
wlee
263 Posts
0
May 17th, 2017 09:00
Suggest focusing on the index backup failure. The issue of nsrmmd timing out may be a secondary effect of the index backup issue.
Does all index backups fail?
For testing, you should test with a NetWorker client that is installed with a supported version. The minimum supported version currently is NetWorker 8.2.
While the index backups appear hung, in NMC are there any messages under the Alerts section? This may help you to understand why it appears hung.
You need better information on what is causing the index backups to fail. Suggest testing using the following steps:
Hopefully this will help you to determine what the index backups are waiting for.