If these tapes are marked FULL, then the tapes ARE being filled efficiently.
It's just that, as you said, your data is expiring at different rates.
Because TSM doesn't expire files based on date, this is a natural side
effect; tapes don't all free up at the same time.
This is normal, and I've seen it with every TSM server I've worked on.
If you set your reclamation threshold to 60, you can, over time, end up with
an apparently infinite number of tapes sitting around with 59% free space.
Your choices are
(1) occasionally bite the bullet and do a lot of forced reclaims or MOVE
DATAs (which is a waste of your time, which is a lot more expensive than a
$45 LTO1 cartridge).
(2) upgrade your robot capacity (as you said, in your case the best way to
do that is to upgrade the drives & cartridges)
(3) GET RID OF SOME DATA.
Deleting unused filespaces WILL immediately free up space on your other
tapes. It may free some entirely, or send the %reclaimable up very high so
that you can reclaim them quickly.
The other thing you can do is review your management class retention
periods. Are you keeping ALL data for the same number of versions/length of
time? Can you change the management class of some directories (like the
winnt directory) to keep fewer versions?
It may help to issue "q occ", and look at the LOGICAL SPACE OCCUPIED for
each filespace. If it is noticeably larger than the physical drive, or if
one is way out of whack with the others, you may be keeping more copies of
the data on that drive than you need.
Also, make sure you have EXCLUDED the TSM server storage pool and data base
volumes (normally have a .dsm suffix) from the backups.
If you don't find anything else you can get rid of, your management has a
choice: Change the backup coverage requirements, or fork over the money for
new drives and media.
Hope that helps...
"I/O, I/O, It's all about I/O" -(me)
From: Copperfield Adams [mailto:Copperfield.Adams AT WWAVRC.CO DOT UK]
Sent: Wednesday, April 07, 2004 11:22 AM
To: ADSM-L AT VM.MARIST DOT EDU
Subject: Help needed to free slots in a 3583 library
I have a query regarding tape usage:
Analysis: We currently have TSM 5.2 installed on a WIN2k Server platform
backing up appx. 50 nodes to 2 disk pools and then migrating to LTO tapes in
a 3583 library, one copy for offsite vault and one copy to remain onsite
within the tape library. We have 60 available slots for LTO volumes as we
leave the I/O drawer free to take volumes in and out of the library. We are
using 100gb native LTO volumes. We have 3 nodes that are configured to
backup using collocation.
Problem: We are constantly running at full capacity within the library
(always 59 Private volumes) leaving only one slot for scratch. Because we
have 3 collocated nodes and produce a dbbackup everyday to go offsite we
need a minimum of 5 free slots (assuming all other backed up data for other
nodes does not exceed 100Gb). When I issue 'q eve * * begind=-1' to check
the previous evenings backups I am usually shown that most nodes have
completed but often have only 2 tapes to go offsite (inc. the dbbackup)
suggesting that collocation is not happening. Aside from this, when I issue
a script to check what space I am able to relaim from the onsite tapes (in
order to free up some slots for scratch volumes) I am presented with the
following (I have omitted the offsite copy LTOs):
STGPOOL_NAME VOLUME_NAME PCT_RECLAIM ACCESS
------------------ ------------------ -----------
BACKUP_TAPE 000411L1 43.8 READWRITE
BACKUP_TAPE 000466L1 43.5 READWRITE
BACKUP_TAPE 000362L1 43.0 READWRITE
BACKUP_TAPE 000382L1 42.6 READWRITE
BACKUP_TAPE 000380L1 39.6 READWRITE
BACKUP_TAPE 000529L1 39.1 READWRITE
BACKUP_TAPE 000632L1 38.7 READWRITE
BACKUP_TAPE 000563L1 38.4 READWRITE
BACKUP_TAPE 000418L1 35.4 READWRITE
BACKUP_TAPE 000403L1 34.6 READWRITE
BACKUP_TAPE 000572L1 33.7 READWRITE
BACKUP_TAPE 000364L1 31.4 READWRITE
BACKUP_TAPE 000564L1 28.8 READWRITE
BACKUP_TAPE 000402L1 28.5 READWRITE
BACKUP_TAPE 000495L1 25.8 READWRITE
BACKUP_TAPE 000391L1 25.6 READWRITE
BACKUP_TAPE 000412L1 22.9 READWRITE
BACKUP_TAPE 000453L1 22.8 READWRITE
BACKUP_TAPE 000501L1 21.3 READWRITE
BACKUP_TAPE 000413L1 16.6 READWRITE
BACKUP_TAPE 000612L1 16.1 READWRITE
BACKUP_TAPE 000470L1 14.8 READWRITE
BACKUP_TAPE 000618L1 11.4 READWRITE
BACKUP_TAPE 000395L1 11.1 READWRITE
BACKUP_TAPE 000607L1 6.3 READWRITE
BACKUP_TAPE 000388L1 3.2 READWRITE
BACKUP_TAPE 000422L1 2.4 READWRITE
BACKUP_TAPE 000570L1 2.2 READWRITE
BACKUP_TAPE 000516L1 0.0 READWRITE
BACKUP_TAPE_COL 000463L1 44.7 READWRITE
BACKUP_TAPE_COL 000600L1 42.9 READWRITE
BACKUP_TAPE_COL 000575L1 41.4 READWRITE
BACKUP_TAPE_COL 000530L1 40.3 READWRITE
BACKUP_TAPE_COL 000540L1 37.7 READWRITE
BACKUP_TAPE_COL 000605L1 35.7 READWRITE
BACKUP_TAPE_COL 000417L1 34.7 READWRITE
BACKUP_TAPE_COL 000457L1 33.0 READWRITE
BACKUP_TAPE_COL 000437L1 32.4 READWRITE
BACKUP_TAPE_COL 000644L1 29.5 READWRITE
BACKUP_TAPE_COL 000415L1 28.4 READWRITE
BACKUP_TAPE_COL 000581L1 28.3 READWRITE
BACKUP_TAPE_COL 000583L1 28.3 READWRITE
BACKUP_TAPE_COL 000601L1 27.7 READWRITE
BACKUP_TAPE_COL 000452L1 26.3 READWRITE
BACKUP_TAPE_COL 000414L1 24.9 READWRITE
BACKUP_TAPE_COL 000464L1 23.1 READWRITE
BACKUP_TAPE_COL 000477L1 22.4 READWRITE
BACKUP_TAPE_COL 000416L1 22.2 READWRITE
BACKUP_TAPE_COL 000502L1 20.6 READWRITE
BACKUP_TAPE_COL 000545L1 19.2 READWRITE
BACKUP_TAPE_COL 000360L1 15.0 READWRITE
BACKUP_TAPE_COL 000379L1 7.7 READWRITE
BACKUP_TAPE_COL 000515L1 4.6 READWRITE
BACKUP_TAPE_COL 000596L1 4.5 READWRITE
BACKUP_TAPE_COL 000559L1 1.9 READWRITE
BACKUP_TAPE_COL 000442L1 0.6 READWRITE
By looking at the 'PCT_Reclaim' I can see that 000516L1 has 0% of
reclaimable space going up to 43.8% for the least used volume (000411L1).
What I have tried to do to get some more usage out of my LTO volumes is
issue a 'MOVE DATA <VOL_NAME> STGP=<DISK_POOL>' to move all the data from
the least used volumes (making it READONLY before starting the process so
TSM does not reuse the volume when it is returned to 'Scratch' status when
empty) back to the disk pool to be migrated onto other onsite volumes. In
effect I am running manual reclamations but as you can see from the output
above it does not look like tape space is being managed as efficiently as
it could be. I'm not sure whether this is to do with the way TSM has been
configured or whether it is natural to see volumes used to differing
percentages because of expirations, etc.
Does anyone know how to get onsite LTO volumes to use closer to 100% of
their capacity or are we expecting too much of our setup - do we need to
start using 200Gb tapes (this will presumably mean we will need to replace
our 4 existing drives)? Also, there are 27 onsite volumes for collocated
nodes but only 3 collocated nodes - is there something I can check to
ensure this is correct?
Also, I am planning to delete some unused filespaces that will remove appx.
1.5Tb of data from our backed up data. Will deleting these filespaces
invoke TSM to automatically purge the associated data that resides within
the onsite/offsite volumes and will this mean that we will free up some
slots within the library for scratch media?
Hope someone can offer some suggestions.
Regards, C. Adams.
IT Support Analyst
WRC Holdings Limited
This email and any files transmitted with it are confidential and intended
solely for the individual(s) or entity(s) to whom it is addressed. Any
views or opinions presented or expressed herein, are those of the author(s),
and may not necessarily represent those of WWAV Rapp Collins Group or it's
affiliates and representatives. No representation is given nor any
liability accepted for the accuracy or completeness of any information
contained in this email unless expressly stated to the contrary. Neither is
any liability accepted for any damage or financial or personal loss
howsoever caused as a result of using the information or data contained
within this email.
If you are not the intended recipient(s) and/or have received this email in
error, you may not use, disseminate, store, forward, print or copy it, in
any method or form, either digital or otherwise. You must delete it
completely from your system(s) and notify the sender that you received it,
so we can ensure it does not happen again.
Registered Office: 1 Riverside, Manbre Road, London UK, W6 9WA. Registered
in England No. 1581935