03-22-2017 06:23 AM
Hi. I am currently runing OAT tests for an Exchange 2013 delivery. We have had lots of issues with GRT recovery, mainly that some mailboxes do not appear in the recovery console (but are definitely on that database) We get No Entity Found issues as well. We also have an issue with displayed database content. For example, this morning I moved a user off database B and put them on database A, then moved my own mailbox from DB A to DB B (DAG). After an hour, I ran a DAG back up (currently about to run a DB recovery from corrupted database) then when I interrogated the back up in Backup Archive and Restore, the databases were showing the mailboxes as being in their pre migration databases!! Why is NBU displaying this when the mailboxes are clearly in different databases? Is NBU querying a cached catalog? What is going on? GRT is a disaster for us.
03-22-2017 07:12 AM - edited 03-22-2017 07:21 AM
Two questions, two different answers.
Mailbox not showing up in the BAR GUI. There was a regression in 7.7.3, fixed in 8.0. The mailbox you don't see is the first one that's found during enumeration. There is a support case on this, with an EEB. Contact your support representative. Mention EEB 3911765.
Your second question is the result of how NetBackup had to change for Exchange 2013 and up. Much of the critical information needed for GRT is no longer in the Exchange database. It's only in AD. Therefore, NetBackup can't glean it from a backup image of the database. It needs to be gathered live. Unfortunately, the PowerShell cmdlets needed for this are slow and CPU intensive. It's as though Microsoft does a really bad outer join on our query. To overcome this, NetBackup does a periodic discovery of the mailboxes on each database. The NetBackup Discovery Framework service (nbdisco) discovers other data, too. For Exchange mailboxes, nbdisco sets a timer and gathers the mailbox information every four hours.
The simplest way to instantly refresh the discovered mailbox data on your master server is to restart the NetBackup Discovery Framework service on each client. At startup, nbdisco resets the Exchange timer and gathers the mailbox information.
Because the mailbox gathering is expensive on system resources, (it's better in 7.7.3 compared with previous releases by about half), nbdisco only gathers the information for the databases that were last backed up on the given client. If a database is backed up on a client for which the master server doesn't have the discovered mailbox list, then NetBackup gathers the mailbox list live during snapshot processing, for that database only.
03-22-2017 07:50 AM
Hi, thanks so much for the quick response.
So if I restart the disco service on all the servers, that will resolve the issue? we have 12 maibox servers in the DAG, so I think I'll script a batch job to do this in future.
03-22-2017 07:56 AM
Yes, or just wait a multiple of 4 hours from the last time the NetBackup Discovery Framework service started between creating, deleting, or moving mailboxes, and running a GRT backup.
03-22-2017 07:58 AM
just restarted the discovery service on all exchange servers and when launching a GRT restore, still get the no entity found response.
databases are showing correct information now though.
03-22-2017 08:06 AM
Do you mean you took a new GRT backup after restarting the discovery service? Did you give the discovery a few minutes to perform its task?
I'm not at all clear what you mean about getting "no entity found." I think you need to work with Veritas support to walk through this third issue.
03-22-2017 08:16 AM
ah, i didnt run a fresh backup, will do that now.