ADSM-L

Réf. : Move from 3590 to LTO or 9840/9940

2002-09-05 21:30:55
Subject: Réf. : Move from 3590 to LTO or 9840/9940
From: Guillaume Gilbert <guillaume.gilbert AT DESJARDINS DOT COM>
To: ADSM-L AT VM.MARIST DOT EDU
Date: Wed, 4 Sep 2002 10:20:43 -0400
Hi Joni

See answers inline






Joni Moyer <joni.moyer AT HIGHMARK DOT COM>@VM.MARIST.EDU> on 2002-09-04 
09:13:16

Veuillez répondre à "ADSM: Dist Stor Manager" <ADSM-L AT VM.MARIST DOT EDU>

Envoyé par :   "ADSM: Dist Stor Manager" <ADSM-L AT VM.MARIST DOT EDU>


Pour :    ADSM-L AT VM.MARIST DOT EDU
cc :
Objet :   Move from 3590 to LTO or 9840/9940

Hi everyone!

I know that I have asked questions about his before, but I am now looking
for individuals that have done this conversion and what
experiences/opinions you may have on this topic.  I was wondering if anyone
has gone from 3590 Magstar tape cartidges to LTO or 9840B or 9940A?  If so,
could you answer the following questions for me so that I can get an idea
of what your environment is like and if it is similar to mine?  Thanks for
any input!!!!  If you have gone from any tape device to LTO, 9840B or 9940A
please let me know of you experiences also.

We have 2 TSM servers (1 production, 1 test) that we will be moving from
the mainframe to SUN servers on the SAN.  We will be doing 8 LAN-free
backups.  We back up approximately 1 TB per night with backups and
archives.  We also have about 200 clients.  We currently have 3590 Magstars
and it takes about 1/2 - 1 hour to reclaim a tape that is 40% utilized.  We
don't mirror our DB/log volumes.  If you have any suggestions on a move
from the mainframe to a SUN server I would also appreciate this a lot!!!!
Thanks again!

First off we have 6 9840a drives fiberattached through 1 Brocade 2800.

1.  What are the sizes of the files that you back up?  Are they all small
files or do you also have large databases being backed up or archived?

We backup mostly small files. We have 4 domino servers that do fulls on friday 
morning and they go to disk. I put the hi at 50% and the stgpool is almost 
empty when they
start. It goes up to 75% in 2 hours. The pool is 166 GB. We migrate at 4 
processes.

2.  How much data are you backing up per night?  And also archiving?

We backup between 150 to 200Gb per day. I still see some errors in the summary 
table so this could be off a bit.

3.  What is you maximum tape mounts per hour?  How many tape drives do you
use concurrently?

There are between 25 and 50 tape mounts per day per drive.

4.  How long does it take you to reclaim a tape that is about 50% full?

We reclaim at 40% full and it takes from 30 to 40 minutes for the most full 
tapes. Heres an output of a summary table query listing yesterday's reclaims

   Start          End                    BYTES
--------     --------     --------------------
09:20:08     09:23:14                942799703
09:23:14     09:29:59               2546257328
09:29:59     09:51:26               6026298456
09:51:26     10:11:40               7782034840
10:11:40     10:34:19               8944638654
10:34:19     11:28:21               9554202669
11:28:21     12:23:34               9673609174
12:23:34     12:55:27               9725246693
12:55:27     13:23:07              10280498712
13:23:07     13:46:28              10606367221
13:46:28     14:08:42              10605442698
14:00:08     14:37:35              11875361905
14:08:42     14:31:11              10953419049
14:31:11     15:00:15              11695091886
15:00:15     15:30:36              11648181858
15:30:36     16:04:07              12001399846

We also have 9840b drives escon attached on Z/OS. The performance is similar :

   Start          End                    BYTES
--------     --------     --------------------
09:15:04     09:20:54                767038992
09:20:54     09:30:44               4559055220
09:30:44     09:50:39               7862147547
09:50:39     10:06:54               7297480954
10:06:54     10:24:34               9948108448
10:24:34     10:46:06              10215817845
10:46:06     11:09:00              11500366738
11:09:00     11:28:10              12342929618

Keep in mind that we compress the data at the client level since all our 
network is Fast-Ethernet (100mb/sec)

5.  Do you consider your type of media reliable?  How many bad tapes have
you had?

6.  How long have you had this type of media working with TSM?

We've used 9840a's for just over a year now and have had no failures.

7.  Do you share tape drives?  If so, what software do you use?

We share the drives with 2 TSM servers on the same box. We use Gresham's DTELM 
since we use Library Station on Z/OS to control our Powderhorn library.

8.  How many clients do you have?

select count(*) from nodes = 140

9.  Do you mirror your recovery log and DB volumes?

No we do not. They are on a Hitachi 7700E subsystem in RAID5. The drives are 
15GB 12k rpm.

10.  How many drives do you have for TSM?

6

11.  What is your data compression ratio?

None... we client compress the backups.

12.  Do you have high duty cycles?

Our drives are pretty much used all day...

13.  Is your TSM server on the mainframe, AIX, SUN, etc.?

We have 1 on mainframe and 4 on AIX. When I first started worling with TSM, I 
thought we should put everything on the mainframe. I've changed my mind since. 
With the
advent of the SAN and fiberchannel drives, UNIX is better positionned for 
backups.

14. Do you have any LAN-free backups on a SAN?

We backup 6 Oracle server with 2 or 3 distinct dbs each via lanfree. They go to 
LTO tape drives though. The throughput is pretty good, maxing out the drives.

Joni Moyer
Associate Systems Programmer
joni.moyer AT highmark DOT com
(717)975-8338


Guillaume Gilbert
CGI Canada

<Prev in Thread] Current Thread [Next in Thread>
  • Réf. : Move from 3590 to LTO or 9840/9940, Guillaume Gilbert <=