TSM server was a Win2k with TSM 4.1.1 (at the time) restoring to a Win2k
Server.  Gig network cards on both sides.  STK 9840FC tape drives.  Client
was collocated by client running TSM 4.1.2.  I ran 3 concurrent restores
starting them about 30 min. apart from each other so that each session would
hopefully not be waiting on the same tape that other session was using (for
the most part this worked but there were a couple of times that a session
had to wait for the tape).

In your case, I have seen slow response from a unix client to a W2k server,
just like our previous Backup software was Unix based and I saw the slow
response from a Unix server to a W2k client.  There are some documented
changes for the Unix client that we applied to speed things up some, but
again my case was W2k to W2k.  I think that you would look at it along the
lines of Richards post in that a lot of it is in the planning.  I had the
9840's, the disk to create large storage pools and installed W2k Servers
(since the majority of our clients are Intel based machines) that had the
horsepower to drive things.  Like just about every application/environments,
if you through the hardware at it it will work.  Alot of post have been made
about AS/400 backups/restores but I can tell you I have 1 AS/400 with a GB
card that backs up 15GB in 45 min and another that has a 100MB card and
older processor in it and backups up at 15GB in 5 hours.

-----Original Message-----
From: Ruomiao Wang [mailto:[EMAIL PROTECTED]]
Sent: Tuesday, May 08, 2001 10:26 AM
To: [EMAIL PROTECTED]
Subject: Re: Newbie Concern on Restore Times


Ron,

Can you please elaborate more on how you were able to restore 320 GB within
36 hours?

I did a partial server restore (about 50GB) in our environment. The average
that I got was a littler bit over 3GB/hour. The client is collocated by
file space, NT 4 with TSM 3.7.2 client. TSM Server is TSM v.4.1.1 on
Solaris 2.6 box. I started 4 concurrent restore sessions with
Backup/Archive GUI.
----------------------------------------------------------------------
Ruomiao Wang
Information Technology
People's Bank
(203) 338-4413
[EMAIL PROTECTED]



                    Ron Pavan
                    <ronpavan@MGMM        To:     [EMAIL PROTECTED]
                    IRAGE.NET>            cc:
                    Sent by:              Subject:     Re: Newbie Concern on
Restore Times
                    "ADSM: Dist
                    Stor Manager"
                    <[EMAIL PROTECTED]
                    IST.EDU>


                    05/08/01 12:57
                    PM
                    Please respond
                    to "ADSM: Dist
                    Stor Manager"





I can tell you that I have restore 320GB in 36 hours.  This 320GB consisted
of 2+ million files that were spread across 17 tapes over a 6 month backup
span  and was performed while there were about 3000 users accessing the
same
Win2k Cluster node.  I my opinion, even though it took 36 hours, this was
not a bad performance on the restore for a complete disaster recovery.  We
compared this to a server we had to restore under our previous backup
software and it would have taken about 3 weeks (GB for GB).  I have restore
other servers that are in the 40-60 GB range with many, many files and a
backup span of 2-3 weeks and the restore time was less that 3 hours.

The major thing you want to do is set your priorities on which restore need
to be completed first and set you users expectation to the correct
timeframe.  I though that would be hard to do with 45,000 employee's but it
is not really that bad when you take into consideration that most of you
restores will be from previous day and would be in DISKPOOL cache if you
can
configure it correctly.  I have had a request to restore a directory that
contained 1000 files (about 20MB total) and it restored about 10 seconds
since it was still in cache.

-----Original Message-----
From: Bernard Ruelas [mailto:[EMAIL PROTECTED]]
Sent: Tuesday, May 08, 2001 9:18 AM
To: [EMAIL PROTECTED]
Subject: Newbie Concern on Restore Times


My company is looking at replacing Legato's NetWorker with
Tivoli's TSM, and I've been told by another competitor's
marketing rep that TSM restore time takes forever.
The problem occurs with the backup of a volume of 30GB+
over a week, then trying to restore it many changes later.

We plan to test this case, but does anyone have any horror
stories/previous experience or docs that would refute this?
I'm a little nervous about trusing what the competitor says....

Thanks in advance,

-Bern

                                                                 _
Bern Ruelas          | <[EMAIL PROTECTED]>    ()                 |_>
Sr Systems Engineer  |                       /\              ___|__
Wk:  (408) 428-5246  |                       \ \         ___/   |  \__
Pgr: (888) 500-4043  | Are you excited?....  /_ \o) ____/   \______/  \____

Reply via email to