Media Server Failover

[H_Sharma]

Hello Experts, I have a windows client and need to restore some SQL database. It gets failed with Network issue (41) Error code every time and taking only 1 media server i.e for sv5.net . 19/03/2016 09:17:32 - Info bptm(pid=15924) INF - Waiting for positioning of media id 0T67L4 on server sv5.net for reading. May I change Media server sv5.net to another for this client restore only? Thanks & Regards,

[revarooo]

If another media server has access to the back end storage (tape or disk) then you can add to FORCE_RESTORE_MEDIA_SERVER = original_media_server new_media_server

Have a look at this Technote:http://www.veritas.com/docs/000044025

[Marianne · Accepted solution]

SQL restore needs a LONG LONG Client Read Timeout on the media server so that SQL can initialize the database. Something like 3 hours (7200) is normally sufficient. You can read up about this in NBU for SQL Guide. Media server failover on Windows master is in Host Properties - > Master - > Universal settings - > Media Host override

[H_Sharma]

Thanks Marianne. Do we need to restart any Master/Media Netbackup services after making the below changes? Media server failover on Windows master is in Host Properties - > Master - > Universal settings - > Media Host override

[Marianne]

I cannot remember if a restart is needed. Try it without restart and see what happens. If original media server is again chosen, use cluster commands/utilities to restart NBU. Before you do that - the media server is hardly ever the reason for Client Read Timeout during SQL restore. The reason is the time needed for SQL to initialize the database. Please search for 'timeout' in NBU for SQL Guide. Rather increase the timeout on the media server or read up in above manual about initializing of SQL db.