RMAN configuration suggestions to help save our storage space

backupdisk-spaceoraclermanstorage

To preface…
Our Senior Oracle DBA has just recently passed away, and I've been taking over. My level of experience is just a junior level. So there is a lot I still don't understand. So I appreciate all the tips I can get, as this is a very frightening position for me.

Our technical issue…
My company has just run out of storage space available in the mounted filesystem, and it is preventing us from recording backups any longer.

I have done my due diligence and cleaned up any obsolete data, but it did not help. The sys admin team is suggesting that we reduce the number of days that RMAN retains our backups. Retention Policy is currently set on default 7 days, I have considered changing this to 5, however 7 days just seems simpler as we hardly have any activity running over the weekend.

My question…
What are your suggestions on things I can do to help clear up more space? What RMAN configurations should I consider looking through? Please advise with any thoughts…

We are running in archivelog mode, we are also doing incremental backups, (and full DB backups), all of which are failing and not recording backups due to the space issue. So there is currently no way to restore anything. Personally, I believe 7 days is ideal and simpler. Much easier to monitor when backups are recorded every weekend.

Best Answer

I have considered changing this to 5, however 7 days just seems simpler as we hardly have any activity running over the weekend.

I think you may be misunderstanding what that means.
Oracle will keep every file it needs to be able to Recover your database back to a point in time five days ago. You can take backups as often as you like during those five days and any that are no longer required will be automatically "dropped off".

So the question is not how long are you keeping backups for, but what kind of backups are you taking and how oftenare you taking them?

Also, what does your Recovery Strategy say? This document, agreed with Management, must include things like:

  • How long you can take to get a broken database up and running again (formally, the Recovery Time Objective, RTO), and
  • How much Data you're allowed to lose in doing so (the Recovery Point Objective, RPO).

Find these metrics and ask yourself - can your current setup support them?
If the answer in "No" then you have a very convincing argument to get more disk space, rather than trying to wriggle around in what you've got.

Related Question