We are using subfile for pst files since about 2 years and are happy with it.
The fileserver has around 150 userhomes all with pst. Most user have pst
between 1 and 2 gb, some above.
We needed subfile because we did the daily backup over a 2 Mbit connection for
about all together of 500 GByte da
About deduplication, Mark Stapleton said:
> It's highly overrated with TSM, since TSM doesn't do absolute (full)
> backups unless such are forced. If you want maximum bang for your buck
> with dedupe, you need to run an app like IBM CommonStore for Email
that
> will enable SIS (single-instance st
1. Use the web gui, start a restore, navigate to a sample directory and
scroll right to see the management class that each file is bound to
2. Use the dsmc command on the client and run a q backup on a sample
directory
Regards
Steve
Steven Harris
TSM Admin, Sydney Australia
Ho
From: ADSM: Dist Stor Manager [mailto:[EMAIL PROTECTED] On Behalf Of
Johnny Lea
> I do have a question on something someone mentioned and that is de-
> deduplication. Are many of you using it and do you consider it an
extra
> level on potential problems? Just curious.
It's highly overrated with
Charles, since I'm the one who originally asked the question I don't consider
this topic hijacked at all! I'm enjoying reading all the responses. Thanks to
everyone who has thrown in their suggestions. Some of the libraries mentioned
are new to me so I'm compiling all these ideas and research
There is no one-to-one correspondence between filespaces (or nodes) and
management classes. There is a one-to-one correspondence between nodes
and policy domains, and a policy domain has one DEFAULT management
class, but can have many.
There is a one-to-one correspondence between individual files
I have a question, as I've been challenged to provide this information.
I need a way to query TSM and find out which Management Class each file
space is bound to. We have some domains that were setup with no less
than 5 Management Classes, each with copy groups that have vastly
differing retentio
I found out what a TBUNDO012 error means - It is IBM speak for "You are
Screwed" - your database is corrupt and you need to dump/load/audit.
This process should complete in about a week. Fortunately, we have just
built out several new TSM servers and have moved all of the nodes(250)
to different
Setting up NDMP is almost a trivial task when using TSM V5.4 and NetApp
filers.
There is no device driver set up on the filer - just enable NDMP.
Schedule snapshots if you want to backup a snapshot for a specific time.
Create two primary storage pools (one for data one for the table of
contents o
At 12:33 PM 2/14/2008, Del Hoobler wrote:
The earlier description on subfile backup is not correct.
Thanks Del, Andy, Bill.. I thought I recalled something about this from
one of the technical sessions I attended at SHARE or one of the Oxford
Symposiums. What you said matches what I had origi
Howard Coles said:
> Sounds like NDMP does things along the lines of the Oracle TDP. TSM
> doesn't know, and can't manage what's in the data sent to it. Is that
> about right? We're considering dumping the oracle TDP because of
that.
> I don't want my TSM system just being a dumping grounds fo
Hi Paul,
the subfile cache doesn't hold the whole file, it holds signatures of
some
sort, probably a checksum for each page. I use subfile for pst files.
My pst is
75 meg but the folder is 5 meg.
We went thru a conversion to outlook 2 years ago. First, you're lucky
to be involved
before the con
The earlier description on subfile backup is not correct.
The cache does NOT hold a copy of the original file for files > 3M.
It keeps a file that stores a MAP of the blocks of the original file.
The TSM Server stores the original file, and the subfile differences.
At restore time, the original fi
As has been noted already, subfile backup was originally targeted mainly
for remote workstation/laptop users coming in over slower network
connections. Most of these users (so we thought) would not be dealing with
such large files. While subfile can technically be used in higher
bandwidth networkin
That's the challenge we're not sure where the issue is, we do know with
STK SL8500 Microcode 3.96 / 3.97 there are internal library
communication issues, where over time the internal Ethernet network
slows down to the point where the ACSLS LM will ask for a tape mount and
the robot will pull a tape
I believe there is a mistake here.
I believe that the cache is only used to cache the ccw's for blocks. The reason
for the 2gig limit was given a long time ago on this list by a TSM dev, they
use a windows api to map the file in virtual memory for i/o. The api they use
has the 2 gig limit.
We
Top of message
>>--> 02-14-08 08:41 S.SHEPPARD (SHS)Re: Backing up PST files
Ditto on the 'bummer' comment. I had noticed the 1GB subfilecachesize
limitation and wondered about the implications.
I guess we'll continue to backup
We migrated from an L700/ACSLS/LTO3 to TS3500/TS1120. Since that
migration, we have been able to stop troubleshooting library/server/tape
issues and focus on client issues. The TS3500/TS1120 combination (with
encryption) just plain works.
Cheers,
Neil Strand
Storage Engineer - Legg Mason
Baltim
Charles,
You're statement about the SL8500 make me nervous. We are planning to
move to a configuration where we are sharing the SL8500 with our
mainframe. We will use ACSLS with a library manager and 3 clients, 16
LTO3 drives.
Are you having TSM issues? Or ACSLS?
Larry McNutt
-Original Messag
> But the 3584 we have remotely is
> bloody -fast-. Scary fast.
I flinch evertime I watch the robot in the 3584 and it moves to
the home frame. I keep thinking it's going to fail to stop and break
through the window.
-
The information contained in this me
"ADSM: Dist Stor Manager" wrote on 02/14/2008
10:44:34 AM:
> I miss my old 3494 12 Frame HA Lib with 3592 FC Tape drives, It never
> skipped a beat even during and HA Failover, in 2 years only one Service
> call on a 3592 Drive for a Slow Fan... Never lost its inventory like our
> current 3584
At 11:04 PM 2/13/2008, Wanda Prather wrote:
Subfile backup was designed to work on laptops, not servers.
Yes, that is what we are trying to do. Sam, on the other hand
The "base" or initial backup copy of a file is stored in a cache directory
on the client, which is also limited in size t
>> On Wed, 13 Feb 2008 10:38:24 -0600, Johnny Lea <[EMAIL PROTECTED]> said:
> I've thought about replacing the 3590 drives with 3592 but I'm
> thinking a new LTO4 library might be better.
Wanda said some balanced things about 3592 vs. LTO4. I share her
empirical observations, but conclude diff
Sounds like NDMP does things along the lines of the Oracle TDP. TSM
doesn't know, and can't manage what's in the data sent to it. Is that
about right? We're considering dumping the oracle TDP because of that.
I don't want my TSM system just being a dumping grounds for data, I can
use setups that
I miss my old 3494 12 Frame HA Lib with 3592 FC Tape drives, It never
skipped a beat even during and HA Failover, in 2 years only one Service
call on a 3592 Drive for a Slow Fan... Never lost its inventory like our
current 3584. And 2 very confused fully loaded with Pass-through
enabled STK SL8
>> On Wed, 13 Feb 2008 17:06:24 -0500, Jim Zajkowski <[EMAIL PROTECTED]> said:
> I'm disappointed with a lack of a real TSM MySQL client; I think the
> arguments that MySQL is a mediocre database are long in the tooth,
> and there are a lot of real commercial operations that are backed by
> MySQL
>> On Wed, 13 Feb 2008 21:23:25 -0500, "Strand, Neil B." <[EMAIL PROTECTED]>
>> said:
> Have you looked into ndmp from the filer to your tsm server? If you
> are at TSM V5.4 you can do this over ethernet and handle the data
> just like any other node - move, copy etc. It also minimizes impact
>
Forgive my ignorance about that number, but some numbers once set stay
that way until they are reset. Is this one of them? I'm thinking that
maybe at some point in the past there were 8 failures even though
recently there haven't been any.
Like I say, there's a lot I'm still picking up on here
28 matches
Mail list logo