Me?
I estimate the capacity roughly as sum(all clients Filespaces)*2
and watch occasionaly
Q STG
for PctUtil.
Once the PcTUtil is higher than about 50%
I watch for
Q VOL stg=TapePool1 status=full
Too many full tapes means too many clients are apportioned
to different tapes, in this case I allow for more scratch volumes.
Do not ask me how much "many" is,
it depends solely on what you want and what you can afford.
regards
Juraj
-----Ursprüngliche Nachricht-----
Von: William Rosette [mailto:Bill_Rosette AT PAPAJOHNS DOT COM]
Gesendet: Freitag, 07. März 2003 13:07
An: ADSM-L AT VM.MARIST DOT EDU
Betreff: Re: AW: progressive backup vs. full + incremental
How do you "ensure the capacity of those 20 volumes is fairly larger then
sum of all versions of files backed up for all 50 clients plus space for
expired files?"
Thank You,
Bill Rosette
Data Center/IS/Papa Johns International
WWJD
Salak Juraj
<[email protected] To:
ADSM-L AT VM.MARIST DOT EDU
T> cc:
Sent by: "ADSM: Subject: AW: progressive
backup vs. full + incremental
Dist Stor
Manager"
<[email protected]
.EDU>
03/07/2003 05:17
AM
Please respond to
"ADSM: Dist Stor
Manager"
Hi Bill!
You are on right way.
Disk pool 1 has NextPool TapePool1,
which allows 20 scratch volumes max, and is collocated.
All 50 clienst from Diskpool1 have to fit on 20 volumes in TapePool1.
You have to ensure the capacity of those 20 volumes is
fairly larger than sum of all versions of
files backed up for all 50 clients
plus space for expired files on tapes with status either filling or full.
So not to "mess",
DiskPools 2 and 3 have to point to another TapePool2,
which can be uncollocated, if you wish so.
regards
Juraj Salak
-----Ursprüngliche Nachricht-----
Von: William Rosette [mailto:Bill_Rosette AT PAPAJOHNS DOT COM]
Gesendet: Donnerstag, 06. März 2003 20:57
An: ADSM-L AT VM.MARIST DOT EDU
Betreff: Re: progressive backup vs. full + incremental
Now back to my original question of:
How do you get a disk pool to go to a certain amount of tapes without the
other disk pools messing them up?
Let me try and explain. I currently have my
Maximum Scratch Volumes 1487
Allowed
in my Copypool. Are you saying to change this amount to 20 and then let
the 1 disk pool put 50 clients on the 20 tapes. I have 3 disk pools, what
about the other 2 nocollocated pools, what do they use for scratch tapes?
(messing up)
Thank You,
Bill Rosette
Data Center/IS/Papa Johns International
WWJD
Chris Murphy
<cmurphy AT IDL DOT STAT To:
ADSM-L AT VM.MARIST DOT EDU
E.ID.US> cc:
Sent by: "ADSM: Subject: Re: progressive
backup vs. full + incremental
Dist Stor
Manager"
<[email protected]
.EDU>
03/06/2003 01:53
PM
Please respond to
"ADSM: Dist Stor
Manager"
>I do not understand how you can get 50 clients on 1 disk pool to go to 20
>tapes.
Bill,
If I am wrong, someone will surely correct me, but what I think happens is
that when migrating the 50 clients from disk to tape, TSM will try to put
each one on its own tape (as collocate says it should). So the first 20
clients each go to their own tape. Then, when client 21's turn comes
around, TSM searches for a new scratch tape to hold this client. If one
does not exist (because of MAX SCRATCH or simply out of scratch in the
library) TSM will go back and re-use the least consumed tape in that pool
for client 21 and migrates 21's data to that tape. Meaning that tape now
holds 2 clients. Then for Client 22, the same thing happens: it searches
for a scratch, when none are found, it finds the next least consumed tape
in
the storage pool and uses it. And so on...
>I would think that 50 clients on 1 (collocated) disk pool will
>take 50 tapes.
The Disk pool would not be collocated (because it would normally be a
random
access pool, not sequential). But, in our example, the destination tape
storage pool would be.
Hope that helps a bit!
Chris Murphy
IT Network Analyst
Idaho Dept. of Lands
Office: (208) 334-0293
cmurphy AT idl.state.id DOT us
|
