|
Re: [ADSM-L] going to all random disk pool for tsm for ve
2015-08-24 18:00:02
You treat the filepool as if it was tape.
You don't want hundreds of nodes data on a tape cartridge because it causes
contention, and it causes massive amount of tape mounts. You can
get the same types of problems with filepool volumes even though it is
disk. When a filepool volume is being read it is mounted, just like a tape.
If you are using replication between two ProtecTIER VTL's, production and
DR, you can actually checkout a volume in one vtl and check it in
in the other vtl, just like tape.
Best Regards, _________________________________________________________
 
email: ron.delaware AT us.ibm DOT com
From:
Stefan Folkerts <stefan.folkerts AT GMAIL DOT COM>
To:
ADSM-L AT VM.MARIST DOT EDU
Date:
08/24/15 12:39
Subject:
Re: [ADSM-L]
going to all random disk pool for tsm for ve
Sent by:
"ADSM:
Dist Stor Manager" <ADSM-L AT VM.MARIST DOT EDU>
But if enough volumes are in the filling state this
is not a problem and
may I ask why use collocation on a filepool?
On Monday, August 24, 2015, Lee, Gary <glee AT bsu DOT edu> wrote:
> No deduplication.
>
> However, I have the trouble of a file class pool, collocated by filespace,
> showing full and not allowing backups even though the pool shows 75%
full,
> just used itx maxscratch allocation.
> This causes an inordinate amount of human interfvention to keep afew
> volumes below maxscratch.
>
>
> -----Original Message-----
> From: ADSM: Dist Stor Manager [mailto:ADSM-L AT VM.MARIST DOT EDU<_javascript_:;>]
> On Behalf Of Stefan Folkerts
> Sent: Monday, August 24, 2015 6:07 AM
> To: ADSM-L AT VM.MARIST DOT EDU <_javascript_:;>
> Subject: Re: [ADSM-L] going to all random disk pool for tsm for ve
>
> I believe a fileclass storagepool is better suited for this kind of
> "permanent" storage on disk.
> A diskpool should be used for things such as disk between the backup
client
> and tape for temporary random I/O storage that can handle an unlimited
> amount of sessions.
> A fileclass storagepool is better suited for long term storage.
> I create a diskpool for VE metadata when using dedup on the filepool
and
> use fileclass storagepools for permanent storage on disk with or without
> deduplication.
>
> Are you planning on using deduplication on disk?
>
>
> On Fri, Aug 21, 2015 at 10:12 PM, David Ehresman <
> david.ehresman AT louisville DOT edu <_javascript_:;>>
wrote:
>
> > Gary,
> >
> > TSMVE 7.1.2 makes file level restores a reality. You may
want to work
> > with them a bit before deciding if you want to go to all random
disk.
> >
> > David
> >
> > -----Original Message-----
> > From: ADSM: Dist Stor Manager [mailto:ADSM-L AT VM.MARIST DOT EDU
> <_javascript_:;>]
On Behalf Of
> > Lee, Gary
> > Sent: Friday, August 21, 2015 3:31 PM
> > To: ADSM-L AT VM.MARIST DOT EDU <_javascript_:;>
> > Subject: [ADSM-L] going to all random disk pool for tsm for ve
> >
> > Hoping to get upgraded to tsm server 7.1.x within the next month.
> > At that time, we are considering changing our storage strategy
to all
> > random disk pool.
> >
> > This because we cannot do a file level restore from a vmware
backup from
> > sequencial media, (be it tape or disk).
> >
> > Has anyone done this, or is there a better way?
> >
> > Looking for guideance and others' experience.
> >
>
|
|
|
