ADSM-L

Re: multiple reclaim processes

2002-09-06 00:25:13
Subject: Re: multiple reclaim processes
From: "Seay, Paul" <seay_pd AT NAPTHEON DOT COM>
To: ADSM-L AT VM.MARIST DOT EDU
Date: Fri, 6 Sep 2002 00:20:34 -0400
Not sure what Wayne was saying but this is from TSM 4.2.1 and maybe lower,
just do not remember when reconstruct was introduced.

>>-MOVe Data--volume_name--+-----------------------+------------>
                           '-STGpool--=--pool_name-'

   .-RECONStruct--=--No------.  .-Wait--=--No------.
>--+-------------------------+--+------------------+-----------><
   '-RECONStruct--=--+-No--+-'  '-Wait--=--+-No--+-'
                     '-Yes-'               '-Yes-'

MOVE DATA supports reconstructing aggregates, but not a MAXPR parameter.
However, you can run as many simultaneous MOVE DATA commands as your system
can stand.

Paul D. Seay, Jr.
Technical Specialist
Naptheon Inc.
757-688-8180


-----Original Message-----
From: Wayne T. Smith [mailto:ADSM AT MAINE DOT EDU]
Sent: Wednesday, September 04, 2002 11:34 AM
To: ADSM-L AT VM.MARIST DOT EDU
Subject: Re: multiple reclaim processes


Two quick (and probably obvious) points:

- You can do multiple simultaneous reclamations if you have multiple storage
pools.  Long on tape drives?  Create another domain (& storage
pool(s)) and split your nodes between them.

- MOVE DATA (tape to tape, or tape back to disk) allows multiple
simultaneous processes on one storage pool and each is often faster than
  the corresponding reclamation.  The down-side to MOVE DATA is that
aggregates are not reconstructed, so your newly filled tape will immediately
have reclaimable space.  Perhaps a statistician could tell us the expected
wasted space remaining after a MOVE DATA.  Maybe with too many caveats to be
useful, ... but when I switched from exclusively performing MOVE DATAs to
exclusively performing reclamation, I "gained" almost 20% empty volumes.

cheers, wayne

Wayne T. Smith, ADSM AT Maine DOT edu, University of Maine System

<Prev in Thread] Current Thread [Next in Thread>