Discussion:
[Veritas-bu] NetBackup Oracle Intelegent Agent
gregsih
2015-01-29 08:28:28 UTC
Permalink
We have NetBackup 7.6.0.4 running on Red Hat Linux 6.6 as our master server with a requirement to backup a number of Oracle 12c Databases (also running on RHEL 6.6) using the new Symantec OIA product. The databases are all RAC clusters, which the agent cannot cope with, so we are backing each database up as a single node.

Does anybody have any experience of this and if so were any issues encountered.

+----------------------------------------------------------------------
|This was sent by ***@ghcomputing.co.uk via Backup Central.
|Forward SPAM to ***@backupcentral.com.
+----------------------------------------------------------------------


_______________________________________________
Veritas-bu maillist - Veritas-***@mailman.eng.auburn.edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu
Lightner, Jeff
2015-01-29 13:49:07 UTC
Permalink
I'm not sure what OIA is but you can backup Oracle RAC using Oracle RMAN in tandem with the NetBackup Oracle agent. With RMAN you don't have to take the database online. Our DBAs here set their backup to the node but I've suggested more than once that they use the VIP instead because that will float to another node if the one they're backing up would go down.

I don't think you'd be able to ever backup from multiple nodes at the same time for the same RAC cluster since each node is mounting the SAME disks as the other nodes in that cluster.

For larger databases we do disk array copying (e.g. EMC BCV or SRDF or Hitachi Shadow Images) to mirror each disk (or LUN) from the Production server to other disks (or LUNs) within the array. When that is fully synchronized we put Oracle into hot backup (or shut it down completely once a week for "cold" backup), split the mirror off, take Oracle out of hot backup (or restart it) on Production. We then mount the mirror copy on a separate media server and do standard (i.e. flat file, not Oracle/RMAN) backup. Most of the shops I've worked in did this.

One place I worked instead setup separate servers that had a cloned database mounted in standby mode. We had full redo logging turned on and copied Production logs from there to the standby node then applied the logs to that standby database. That meant you could get Production back online by taking the standby out of standby and making it the Production node. (We did however, put in a 6 hour delay between shipping the logs to the standby server and actually applying them to help prevent any faux pas that took down Production from getting immediately applied to standby.)

-----Original Message-----
From: veritas-bu-***@mailman.eng.auburn.edu [mailto:veritas-bu-***@mailman.eng.auburn.edu] On Behalf Of gregsih
Sent: Thursday, January 29, 2015 3:28 AM
To: VERITAS-***@MAILMAN.ENG.AUBURN.EDU
Subject: [Veritas-bu] NetBackup Oracle Intelegent Agent

We have NetBackup 7.6.0.4 running on Red Hat Linux 6.6 as our master server with a requirement to backup a number of Oracle 12c Databases (also running on RHEL 6.6) using the new Symantec OIA product. The databases are all RAC clusters, which the agent cannot cope with, so we are backing each database up as a single node.

Does anybody have any experience of this and if so were any issues encountered.

+----------------------------------------------------------------------
|This was sent by ***@ghcomputing.co.uk via Backup Central.
|Forward SPAM to ***@backupcentral.com.
+----------------------------------------------------------------------
Loading...