Hi Wanda,

In a similar scenario ( 700 VMs ) but TSM Server Linux 32 cores/64 GB RAM, I 
have the dedup threshold setup to 10. I know this is very agressive but I am 
getting good results in the deduplication ratio. The file size of stg volumes 
is 100 GB and I now it moves a lot of info to inrease the dedup ratio, ( I am 
not sure which is an optimal value ) . Currently the deduplication in server is 
around of 70 % and it is increasing a good slow, but at least the stg pool is 
reducing the space occupied, at the beginning it was aroun 48 %. I also have 
client dedup as well and the most of VMs are showing a 96-99,99% de total 
reduction. Anyway,  I think the calculation of Dedup is not clear  ;-))

My main headache now it is monthly or annual backups. I am studing several 
alternartives, export - import and node replication or split the full backups 
in 4 weeks. I have doubts about how much storage I will need for this kind of 
backups, TSM severs, etc.. But at the moment I am in fase 0. 

I don't kwow if sb have open a RFE in order to have a Backup with two 
Management Class, because the inc backup for one day can be incremental and 
monthly for example.. .. But this is other war.... 

Regards, 

Fran



________________________________
 De: "Prather, Wanda" <wanda.prat...@icfi.com>
Para: ADSM-L@VM.MARIST.EDU 
Enviado: Viernes 7 de junio de 2013 1:49
Asunto: Sequential dedup pool doesn't seem to reclaim as it should
 

TSM 6.3.3 on Win2K8-64

I have a sequential pool on disk with DEDUP=yes.  (Happens to be for TSM-VE 
data, but I don't think that's relevant.)
Settings are below.  There is 1 identify duplicates process always active.
Reclaim threshold is set to 20.

Every night the clients back up.  At 4am we start the backup stgpool to a tape 
copy pool.
When that is in process, several reclaims kick in on their own.
But once those are finished, they don't ever crank up again later in the day.
Every day it leaves several volumes above the reclaim threshold.  Right now 
there are 5.


*        Identify Duplicates is finished and idle.

*        No client activity.

*        "Backup stgpool file-ve copypool"  returns no data to be copied.
It has been that way for the last 9 hours.

If I start the reclaim myself with "reclaim stgpool file-ve threshold=20", it 
runs just fine.
But it won't reclaim (and therefore dedup) on its own.  Shouldn't it?

I'd like to have those volumes empty (and deduped) before the next backup cycle.
I can force it by scheduling the extra reclaim command, but I don't understand 
why it doesn't kick off on its own more than once a day?


tsm: LFTSM>q stgpool file-ve f=d

                    Storage Pool Name: FILE-VE
                    Storage Pool Type: Primary
                    Device Class Name: ONLINEFILE
                   Estimated Capacity: 20,447 G
                   Space Trigger Util: 68.3
                             Pct Util: 68.3
                             Pct Migr: 68.3
                          Pct Logical: 89.3
                         High Mig Pct: 98
                          Low Mig Pct: 70
                      Migration Delay: 0
                   Migration Continue: Yes
                  Migration Processes: 1
                Reclamation Processes: 2
                    Next Storage Pool:
                 Reclaim Storage Pool:
               Maximum Size Threshold: No Limit
                               Access: Read/Write
                          Description: Dedup VE pool
                    Overflow Location:
                Cache Migrated Files?:
                           Collocate?: No
                Reclamation Threshold: 20
            Offsite Reclamation Limit:
      Maximum Scratch Volumes Allowed: 0
       Number of Scratch Volumes Used: 0
        Delay Period for Volume Reuse: 0 Day(s)
               Migration in Progress?: No
                 Amount Migrated (MB): 0.00
     Elapsed Migration Time (seconds): 0
             Reclamation in Progress?: Yes
       Last Update by (administrator): WANDA
                Last Update Date/Time: 05/29/2013 10:25:01
             Storage Pool Data Format: Native
                 Copy Storage Pool(s):
                  Active Data Pool(s):
              Continue Copy on Error?: Yes
                             CRC Data: No
                     Reclamation Type: Threshold
          Overwrite Data when Deleted:
                    Deduplicate Data?: Yes
Processes For Identifying Duplicates: 1
more...   (<ENTER> to continue, 'C' to cancel)

            Duplicate Data Not Stored: 20,588 G (60%)
                       Auto-copy Mode: Client
Contains Data Deduplicated by Client?: No

Wanda Prather  |  Senior Technical Specialist  | wanda.prat...@icfi.com  |  
www.icfi.com
ICF International  | 401 E. Pratt St, Suite 2214, Baltimore, MD 21202 | 
410.539.1135 (o)

Reply via email to