Re: [Veritas-bu] backing up by application

If you use an external scheduler such as $Universe or Tivoli Workload Scheduler (a/k/a Maestro) then you can make all the backups dependent upon one another although I wouldn’t. There is no need to rerun the middle tier backup just because the DB backup fails or more importantly no reason to rerun the DB backup because the middle tier fails. In the past I’ve worked at locations where we used these schedulers mainly to deal with shutting down middle tier before stopping DB then backing up both. Also often enough you’re actually doing a BCV copy of the DB from one host, importing and mounting that copy on another and then backing up the BCV mounted. These scheduling tools are great for insuring all the different pieces have been done before the DB backup is run on the BCV copy.

At my current job we don’t have such a tool so we generally script things like that. For most of our systems we have at least 2 policies: 1 for the OS which excludes all application and database directories and 1 or more for each of the distinct “environments” (e.g. PROD, DEV, STAGING) that exist on the systems. Since Oracle is typically broken up into at least one middle tier (application) server and one database server we would usually have 1 policy for backing up its middle tier and one 1 for backing up its database.

We also name our policies so that they are somewhat descriptive:
hostname-OS

hostname-APP-ENV

hostname-DB-ENV

Where hostname would be the name of the server and ENV would be the name of the specific environment.

From: veritas-bu-bounces AT mailman.eng.auburn DOT edu [mailto:veritas-bu-bounces AT mailman.eng.auburn DOT edu] On Behalf Of Ed Wilts
Sent: Tuesday, January 13, 2009 7:37 AM
To: VERITAS-BU AT mailman.eng.auburn DOT edu
Subject: Re: [Veritas-bu] backing up by application

On Mon, Jan 12, 2009 at 8:16 AM, kharris <netbackup-forum AT backupcentral DOT com> wrote:

We are redesigning our backup infrastructure to backup files based on applictaion as opposed to server.One of the challenges is identifying wether netbackup can go accross the network and pull files from multiple systems provided an identifier is supplied and bucket them.We are well aware of the policy,tape pool,as well as multiplexing aspects of this.We are trying to see if this is possible.

You can certainly "bucket them" with a keyword but that's only a small part of the solution. Scheduling your backups will be the real problem - an individual backup policy may only contain a file list that's common to all of the servers in the policy and you can have only one policy type (Standard, Windows, etc.) per policy. In other words, an application that spans an Oracle database on a Unix cluster with a Windows file cluster is going to be really difficult to back up on an application-consistent basis since you'll need a minimum of 4 NetBackup policies just to do those backups.

You also have to decide what to do if one of the policies doesn't run or fails - does the entire application need to be rerun, or just the failed pieces? That's something we can't answer for you.

It's all doable although you may need an external scheduler to help with this because of the limitations of what can go into a single NetBackup policy and the reruns that need to be done if part of the backup set fails.

.../Ed

----------------------------------
CONFIDENTIALITY NOTICE: This e-mail may contain privileged or confidential information and is for the sole use of the intended recipient(s). If you are not the intended recipient, any disclosure, copying, distribution, or use of the contents of this information is prohibited and may be unlawful. If you have received this electronic transmission in error, please reply immediately to the sender that you have received the message in error, and delete it. Thank you.
----------------------------------

_______________________________________________
Veritas-bu maillist  -  Veritas-bu AT mailman.eng.auburn DOT edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu