12-30-2014 06:25 AM
We currently have:
Netbackup 7.1.0.2 Master Server
Windows 2003
Netbackup 7.1.0.2 Media Server
Red Hat Linux 2.6.9
2 Unit tape Robot
Both the master and media server have access to the robot and the 2 tape units. Also the Red Hat media server has 16TB of local disk storage for backup. When the DSSU policy takes the images from the local disk storage on the media server to tape the jobs stream across the network to the master server instead of using the fiber from the media server. Some of the jobs were using the media server and some were using the master. I came across the article https://www-secure.symantec.com/connect/forums/duplicate-job-running-wrong-media-server so I ran the nbemmcmd with the common_server_for_dup required command. Now all the jobs use the master instead of using both servers.
How can I force the media server to send the DSSU policy images directly to the tape drives instead of streaming across the network to the master server?
Solved! Go to Solution.
12-31-2014 07:03 PM
When you said "Both the master and media server have access to the robot and the 2 tape units.", are they sharing the same tape library? If so, which one is the robot control host?
In your DSSU relocation destination storage unit, it should use the tape storage unit configured for the media server instead of the one configured for master server.
12-30-2014 03:01 PM
Most efficient way to duplicate is creating "storage lifecycle policy" there you can select alternate read server
http://www.symantec.com/business/support/index?page=content&id=HOWTO49103&profileURL=https%3A%2F%2Fsymaccount-profile.symantec.com%2FSSO%2Findex.jsp%3FssoID%3D1419980439856tmuPictjwWJBaOz0V08SUnJ9Lwk6eWW4BAdC4
12-30-2014 04:14 PM
As a matter of interest, are you using storage unit groups ?
What do you have set for the final destination storage unit ?
12-31-2014 05:48 AM
I have been looking into storage lifecycle policies and vaulting but haven't implemented anything yet.
Currently we have 2 storage units. 1 is the tape drives and the other is the 2 storage units that we have. The storage units are setup as a temporary staging before sending the images to tape. The media server is set as the media server and not the master but when the high water mark is reached the duplication job sends the images to the master over the network and then writes them to the tape drive.
12-31-2014 07:03 PM
When you said "Both the master and media server have access to the robot and the 2 tape units.", are they sharing the same tape library? If so, which one is the robot control host?
In your DSSU relocation destination storage unit, it should use the tape storage unit configured for the media server instead of the one configured for master server.
01-02-2015 01:08 AM
Try setting this EMM:
nbemmcmd -changesetting -MUST_USE_LOCAL_DRIVE 1 -machinename {server}
01-06-2015 05:36 AM
Thank you watsons.
I checked the final destination storage unit and it had the tape storage unit for both the master and the media server. Also the storage unit was set to fail over so that is why it was always picking the master server. I created a new tape storage unit and put the media server tape drive first then the master server tape drive and set to fail over. It looks like once the current duplication jobs completed the new jobs started going to the tape unit from the media as I would like.
01-12-2015 10:49 AM
I have been testing storage life cycle policies and have a question.
How can I see that the jobs are going to tape and secondly can I set the duplication to happen after 1-2 days instead of a week? The lowest time frame is 1 week.
01-13-2015 07:45 AM
Please open a new thread for this qustion.