ADSM-L

Re: Effect of multiple migration processes to a collocated pool - serial or parallel node migration ?

2003-07-30 07:21:39
Subject: Re: Effect of multiple migration processes to a collocated pool - serial or parallel node migration ?
From: Zlatko Krastev <acit AT ATTGLOBAL DOT NET>
To: ADSM-L AT VM.MARIST DOT EDU
Date: Wed, 30 Jul 2003 14:20:26 +0300
Diskpools are always "collocated" :-)))
Option 1 is incorrect. It can happen only if next pool is
*non-collocated*! Look at list archives and you will see that the answer
to question "why my multi-session client is having all but one in MediaW"
is to disable the collocation.
Option 2 is somewhat correct. Each migration batch is processed by a
process and next process can work only if there is another batch. If the
next storage pool is collocated by node, a complete node constitutes a
batch. If the storage pool is collocated by filespace, each filespace is a
separate batch.
Example:
A node is having three filesystems /data1, /data2 and /data3, 100 GB each.
The diskpool is having migproc=3.
If the tapepool is having collocation=filespace, three separate processes
will be lauched and each will migrate 100 GB.
If the tapepool is having collocation=yes, only one process will be
lauched and it will migrate all 300 GB. Only if there is another node's
data in diskpool, a second process can start to migrate it.

Zlatko Krastev
IT Consultant






PAC Brion Arnaud <Arnaud.Brion AT PANALPINA DOT COM>
Sent by: "ADSM: Dist Stor Manager" <ADSM-L AT VM.MARIST DOT EDU>
30.07.2003 13:28
Please respond to "ADSM: Dist Stor Manager"


        To:     ADSM-L AT VM.MARIST DOT EDU
        cc:
        Subject:        Effect of multiple migration processes to a collocated 
pool - serial   or
parallel node migration ?


Hi list,

A quickie one : when migrating multiple nodes data from non-collocated
disk pool to collocated tape pool, using migproc=2, in which order will
data be migrated ?
1) node 1 data on 2 different tapes, then node 2 data on 2 different tapes
etc ....
2) node 1 data on one tape and node 2 data on another tape, then node 3
data on another tape etc ....

I would tend to say 1) is the right response, as the order of nodes
processed is per largest amount of data in the disk storage pool, but
wanted to make it sure ...
If my assumption is right, is there any way achieving  parallel mounts for
nodes migration (option 2)  instead of serial mounts (option 1) ? This
would avoid waste of tapes (2 tapes created for 1 node migrated, in a
collocated pool looks strange  to me), and also from time (tape
reclamation will probably recombine those tapes to a new one after a
while)
Thanks in advance !
Regards.

Arnaud

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
| Arnaud Brion, Panalpina Management Ltd., IT Group     |
| Viaduktstrasse 42, P.O. Box, 4002 Basel - Switzerland |
| Phone: +41 61 226 19 78 / Fax: +41 61 226 17 01       |
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=

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