Networker

Re: [Networker] How to tell how much physical tape is being used?

2007-04-24 16:55:47
Subject: Re: [Networker] How to tell how much physical tape is being used?
From: Darren Dunham <ddunham AT TAOS DOT COM>
To: NETWORKER AT LISTSERV.TEMPLE DOT EDU
Date: Tue, 24 Apr 2007 13:53:09 -0700
> Is there a way to determine how much physical tape space is being used 
> by a backup?

Not in a vendor-independent way.

> I'm probably not using 5 
> GB worth of tape space due to hardware compression, so it is possible 
> that both examples result in the same amount of physical tape space 
> being used? Maybe?

Probably similar.  Most modern tape drives seem to use something similar
to LZW that will give fair text compression and good speed.  Nowadays
they should also have filters so that if the compression is negative, it
writes out the uncompressed data.  I have no idea what RMAN does.  It is
possible that RMAN compression is able to optimize the DB format in ways
that the tape drive cannot.

> Is there a way to tell which method is actually using more tape or if 
> they're close?

You could RMAN backup to a file, (with with and without RMAN
compression) to estimate how it's doing.  Then you could run those
through simple LZW compressors to guess how the tape might behave.

You could also have a specific tape/software combo from a vendor that
can display the amount of tape used on a cartridge.

> I can query for the 'mediafile' numbers for the save sets, and I notice 
> that they always start 2 beyond the previous one, but a given saveset 
> could be of any size so this doesn't really tell me how much tape is 
> being used, right?

Right.  And they're "virtual" indexes anyway that don't see HW
compression except that that tape will hold more.

> Here's why I ask such a goofy (stupid?) question. OK, we ran this Oracle 
> backup (level full) with the 'compressed' option, and we noticed that 
> the backup size was very small (kinda nice!), but as a result, the drive 
> speeds were very slow. This is not surprising given that we probably 
> were underfeeding the drives since the client was sending less data due 
> to the Oracle compression. Next, we run the same thing but without the 
> 'compressed' option, and now the drives scream (again, not surprising 
> given that the client now has more data to send), but the backup sizes 
> were way larger. Overall, however, the backup completion times were 
> reasonably similar. I'm wondering if maybe in the second example that 
> because the data was not already compressed that the drives shrunk it 
> down and that maybe it's not really taking up any more space on the tape 
> than the same data in the first run. We recovered both sets, and the 
> total number of recovered database files and their sizes are the same in 
> both cases.  How can I determine, therefore, which method used less tape?

Without a specific tool from the tape drive vendor that will tell you,
you cannot be certain.  You could also just repeat the backup 'X' times
to see when the tape finally fills.  First one to fill the tape can be
assumed to be the bigger on-tape backup.

-- 
Darren Dunham                                           ddunham AT taos DOT com
Senior Technical Consultant         TAOS            http://www.taos.com/
Got some Dr Pepper?                           San Francisco, CA bay area
         < This line left intentionally blank to confuse you. >

To sign off this list, send email to listserv AT listserv.temple DOT edu and 
type "signoff networker" in the body of the email. Please write to 
networker-request AT listserv.temple DOT edu if you have any problems with this 
list. You can access the archives at 
http://listserv.temple.edu/archives/networker.html or
via RSS at http://listserv.temple.edu/cgi-bin/wa?RSS&L=NETWORKER