Veritas-bu

Re: [Veritas-bu] Backing up multiple Oracle db's from a single host

2009-05-01 13:33:56
Subject: Re: [Veritas-bu] Backing up multiple Oracle db's from a single host
From: Wayne T Smith <wts AT maine DOT edu>
To: VERITAS-BU AT MAILMAN.ENG.AUBURN DOT EDU
Date: Fri, 01 May 2009 13:30:07 -0400
One NBU policy per DB works best for us.  You don't necessarily have to 
have multiple scripts, as long as your script "knows" what DB to backup 
and what policy/schedule to use.

Having separate policies makes scheduling changes and reruns easier.  It 
allows us to see exactly which DB is being backed up in the Admin 
Console.  It also allows us to limit and adjust schedules to minimally 
affect performance (the "nnn" DBs on a client don't all start their 
backups at once).

Cheers, Wayne

nellis wrote, in part,  on 2009-04-30 5:23 PM:
> Hello everyone, first I'll go ahead and apologize for such a long post.
>
> I'd like to ask everyone; what works best for those of you backing up 
> multiple Oracle db's from a single host. I have several Oracle 10g servers I 
> backup. Each has anywhere from 1 to 20 db instances. I use a single policy 
> and a common path & script name to backup the Oracle servers. This for me 
> keeps the number of policies I have to manage down to a minimum. 
>
> Our dba's have asked for one policy to be setup per DB. The reasoning behind 
> this is that they want the flexibility to restart a failed backup without the 
> restart initiating a backup of all DB's on servers. I told them that what 
> they should be doing on failed DB backups is from the "activity monitor", 
> right click on the failed Job ID and select restart. 
>
> Here is where my lack of "in depth knowledge" of how NBU works with the 
> Oracle RMAN API when a single stream fails. 
>
> They asked, how does NBU know to just restart the backup for the DB that 
> failed? The best "high level" reply I could come up with was that NBU tracks 
> which DB piece it's backing up in a log and when a restart is initiated, it's 
> able to request a backup of just the DB which that piece was associated with 
> via calls to the Oracle RMAN API. 
>
> In other words, restarting a single failed Jod ID from "activity monitor" for 
> a backup stream associated with an Oracle backup does not mean that NBU looks 
> at the policy and executes the backup script specified in the backup 
> selection for all servers or just one server which is a member of that policy.
>
> Am I wrong? 
> How does a restart work with Oracle? 
> How do you backup Oracle servers which host multiple instances?
> What are the advantages and disadvantages if any with one policy per instance?
>
> Having a policy for each to me sounds like a nightmare in more than one way 
> but I'd REALLY appreciate anyone's input here.
>
> Thanks in advance to everyone that replies!
>
> +----------------------------------------------------------------------
> |This was sent by Norman_Ellis AT Discovery DOT com via Backup Central.
> |Forward SPAM to abuse AT backupcentral DOT com.
> +----------------------------------------------------------------------
>
>
> _______________________________________________
> Veritas-bu maillist  -  Veritas-bu AT mailman.eng.auburn DOT edu
> http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu
>   
_______________________________________________
Veritas-bu maillist  -  Veritas-bu AT mailman.eng.auburn DOT edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu