Tape Usage

[ojmagg]

We upgraded from 5.1 mp 6 to 6.5 a few months back.  The process it's self went smooth, however, we are still having some issues...

We seem to be using a few extra tapes per day...one extra (for a total of 2) for our catalog.  When we were on 5.1, we only sent 1 off-site.  Then, after weekend, we are sending off 15 tapes.  Again, on 5.1 we used to only send off 10 after a weekend...

Anyone else notice a change in tape usage after moving to 6.5?

[scored_well2]

Have you checked your backup policies for any changes? For example, jobs running more than once during a window, software compression turned on? I have seen upgrades happen where the 5.1 policy came out the other side of the upgrade to 6.0 with different settings.  I have not tried an upgrade to 6.5 though.

Message Edited by Simon Caudwell on 01-04-2008 07:01 AM

[Stumpr2]

---

ojmagg wrote:

We upgraded from 5.1 mp 6 to 6.5 a few months back. 

---

a few months back? That doesn't give much to compare. If the images from a few months back still exist then you can compare size using

bpimagelist -U -client <clientname> -d MM/DD/YY -e MM/DD/YY

compare from a month before then run it again for the month of December

bpimagelist -U -client <clientname> -d 12/01/07 -e 12/31/07

[ojmagg]

Thanks for the replies!

I've been over the setting...they appear to be same as when we were on 5.1...

Here's the output from bpimagelist -U -client <client> -d MM/DD/YY -e MM/DD/YY:

Backed Up         Expires       Files       KB  C  Sched Type   Policy
----------------  ---------- -------- --------  -  ------------ ------------

01/02/2008 03:00  INFINITY     261043 42429911  N  Full Backup  Linux 

10/02/2007 03:01  INFINITY     240671 40518658  N  Full Backup  Linux

09/02/2007 03:02  INFINITY     240811 71663008  N  Full Backup  Linux

08/02/2007 03:20  INFINITY     223070 93752005  N  Full Backup  Linux

07/02/2007 03:22  INFINITY     221678 97044131  N  Full Backup  Linux

06/04/2007 08:56  INFINITY     223228 94675249  N  Full Backup  Linux

05/02/2007 03:00  INFINITY     222830 98771890  N  Full Backup  Linux

04/02/2007 03:10  INFINITY     221117 95262636  N  Full Backup  Linux

03/02/2007 03:03  INFINITY     218702 90555715  N  Full Backup  Linux

02/02/2007 03:17  INFINITY     216826 87465102  N  Full Backup  Linux

01/02/2007 03:00  INFINITY     210022 37068399  N  Full Backup  Linux

Looks like a 2 gig swing...but, that appears normal, given past performance...any other thoughts?

[Stumpr2]

Any new retention levels or media servers?

[ojmagg]

Nope...fyi...just 1 media/master...

[Rakesh_Khandelw]

With 5.1 catalog backup usese/overwrites/rotates the same set of tapes, where as with 6.0/6.x you may be writing to different tapes for catalog backups (with 6.0/6.x recommended way of perfoming catalog backup is online/hot backup, which is policy based). Look into your catalog backup policy and schedule attributes.

Also, see if you have backups going to more volume pools.

[TerryD2]

We called Symantec again to ask about this problem.  This time we got a new tech who told me this was a known issue that started in 6.0.  It is a bug that apparently was not corrected in 6.5.  The problem is that when the robot performs a tape cleaning the NetBackup TapeAlert flag does not get reset properly.  This causes a second tape to be used when it shouldn't.

Here is the solution:

To clear the TapeAlert Flag, run /usr//openv/volmgr/bin/tpclean -M <drive_name> one drive at a time for each tape drive.

This worked for us.  Our catalog backup is now back to one tape.