Hi Gretchen (and Gerhard),
Hope this finds you doing well. We've sure missed you at SHARE!
Question: Have you done anything new to "protect" yourself from a "bad" maintenance
level? Was there something special in 5.1.5 or 5.1.6 that attracted you to upgrade so
soon?
I have a customer wanting v5,
Cannot comment on the drive "view" or schedule-mode problems -- sounds like
a TSM-cluster service definition discrepancy (else, a bug; need to check the
cluster services setups and client code level, then contact Support Line or
install latest client code -- there's been a bunch of client-code act
Your backup sizing is quite small; are you sure it's that small? There are
Redbooks on sizing for AIX; also, there is much material from SHARE
proceedings on performance and tuning.
A more typical arrangement (I've seen) for a single-client+server situation
might be 50 GB to start, up to 100 G
Also, there *may* be other messages of interest (if b/a-client, 4961, 4959,
and others in close numeric proximity are the session statistics messages).
Don France
Technical Architect - Tivoli Certified Consultant
Professional Association of Contract Employees (P.A.C.E.)
San Jose, CA
(408) 257-30
Reading thru this thread, no one has mentioned that backup will be slower
than archive -- for TWO significant reasons:
1. The "standard" progressive-incremental requires alot of work in comparing
the attributes of all files in the affected file systems, especially for a
LARGE number of files/direc
nsfield
> Senior Consultant
> Solution Technology, Inc
> 630 718 4238
>
>
>
>
> "Don France (TSMnews)" <[EMAIL PROTECTED]>
> Sent by: "ADSM: Dist Stor Manager" <[EMAIL PROTECTED]>
> 05/03/2002 05:00 PM
> Please respond to "ADSM:
You must realize the Linux directories (likely) fit nicely within the
available space in the TSM server,,, how do you know Linux clients ignore
your DIRMC? (The only simple way I would know how to test is to create a
path greater than 160 bytes -- I think the TSM db only has room for about
150 by
Try using tapeutil (or some 3575 library tool, like mtlib on 3494) to mount
a tape in a given drive -- you must choose a scratch tape and ensure TSM is
not using the drive; then, issue (1) make drive "offline" to TSM, (2) the
cmd to mount the tape, (3) the cmd to mksysb to that device, then (4) t
This is documented in the Admin Guide about directories stored in backup
storage (and numerous APAR's to "fix" over the years since v1).
When directory information is stored for backups, TSM attempts to keep all
the info in the TSM database; if there is more data than the data base can
hold (due
? Not powerpoint,
I know...
>
>
>
> _
> William Mansfield
> Senior Consultant
> Solution Technology, Inc
>
>
>
>
>
> "Don France (TSMnews)" <[EMAIL PROTECTED]>
> Sent by: "ADSM: Dist Stor Manager" <[EMAIL PROTECTED]>
> 05
499
> Mailto:[EMAIL PROTECTED]
> http://intraCTI.etat-ge.ch/services_complementaires/messagerie.html
> ______________
>
-Message d'origine-
De : Don France (TSMnews) [mailto:[EMAIL PROTECTED]]
Envoyi : vendredi, 3. mai 2002 03:39
Probably not... however, if you are using AIX, you could run the dsmulog
daemon to capture the console log to a file (much like the old SYSLOG
feature on MVS), then you could "monitor" that file with "more" or "tail".
See the AIX Admin Guide for details.
Don France
Technical Architect - Tivoli Ce
pending on what should be done that day, leaving the version count the
same?
> If I change it to absolute, what does that do the modified backups already
> taken, anything?
>
> Thank you for your help.
>
> Diana
>
>
>
> Quoting "Don France (TSMnews)" <[E
You are abit confused. The *ONLY* way to have TWO policies applicable to a
given file is to use TWO node-names for your backups; swapping policy sets
*may* work for your situation, if what you want (and set) is 30 versions of
a given file... that piece will work.
Files can be bound only to one
F Y I --- if you try using the SUMMARY table in the latest TWO maint. drops of the
server, you get unreliable, mostly zeroes in the bytes-transferred column!
Don't know if other columns are correct, but at least they are (mostly) non-zero;
also, not sure of all the data is getting to this tabl
Nope... I do have hardcopy (from SHARE); you might find what you want in
the books --- there's a good "summary of changes" in the preface of the
Admin. Guide, Using xxx Clients, and Admin. Ref.
Don France
Technical Architect - Tivoli Certified Consultant
Professional Association of Contract Emp
I may have missed a large part of this thread; seems that normal backup stg
works just fine (notwithstanding the courier damaging media in transit --
maybe need a "closed container with padding" contract, like Paul Seay is
doing). Your concern becomes (1) the recovery plan (DRM solves this) and
Sounds like a bug -- yes, there has been a level (or 3) that incorrectly
caused exclude list to be processed by "ar" cmd... it's the client code that
controls it -- try running the latest (4.2.x) client, unless you're hot to
use 5.1, then get the latest 5.1 download patch.
Don France
Technical Ar
The customers I've worked with used a shell script to determine -archmc for
daily/weekly/monthly; without TDP, the script manipulates the parameter
passed in for the -archmc value on the "dsmc ar" cmd... you could use a
presched command to do the same (for flip the profile name, causing TDP to
us
1. Turn OFF client and admin schedules;
2. Turn OFF any real-time virus scan (on the destination client);
3. If you're doing restore to Windows platform, use -DIRSONLY option, to
restore just the directories, first -- after first pass, then restore
the -FILESONLY -- using PIT restore options, in
There are a couple (new) white papers on the Tivoli site... one's pretty
good ("Achieving Cost Savings..."), and has no "do not duplicate" notices;
the other is pretty weasel-ly, was commissioned to a consulting group *and*
has "do not duplicate" notices on it --- AND it's not all that good, excep
Petur,
You will need v5 server, as well as client; there are other limitations --
see the post from Anthony Wong, and go RTFM... they are now posted at
http://www.tivoli.com/support/public/Prodman/public_manuals/td/StorageManage
rforWindows5.1.html
Have fun!
- Original Message -
From
You've just about gottit all. There is some concern about this whole thing;
currently, you are limited to image backups using NDMP; if you want
file-level granularity, you must use NFS mount to access the file system
from a supported b/a-client... not as pretty as you'd like, but that's the
way
Another huge contributor to this phenomenon is failure to use DIRMC on disk
(and FILE on disk for sequential migration/reclamation) storage pools.
A lot of servers and a lot of objects (more common with Windows these days)
need all the help they can get; keeping directories on disk (using DIRMC
If you truly want to snapshot your current data on a weekly/monthly/yearly
basis, the current-release recommended solution would be "generate
backupset" -- which, at least, avoids re-sending the data across the LAN.
Alternatively, Unix has image backup capability, and 5.1 will have it for
Win2K. E
If you stay on the same platform-OS, all you need to do is shutdown (old server) after
a final DB backup, move the HW connections to the new server, restore DB, and you're
finished. Remember to copy the dsmserv.dsk (filesystems for all TSM server files --
db, log, disk pool vols, logical volum
Yep... it's possible to arrange a single-drive environment; it's highly
dependent on the onsite person doing tape mounts on request, and organizing
the tape pool to satisfy your retention requirements -- if you're lucky, get
sufficient disk pool storage to allow either single-drive reclamation or
Ben,
It's highly recommended that (on Solaris) you use raw volumes for the TSM
file systems (ie, db & log)... due to severe performance penalties. Other
than that, all should be okay. There are other rules of thumb to consider,
especially on Sun boxes, regarding how much CPU power to support mul
Are you sure you're using CAT-5 cable?
You say the NIC's are forced at 100/full -- how about the switch ports?
Is there adjacent "noise" that might be emitting across the network?
Do you have old vs. current switch HW? Is it up to date, microcode?
VLAN's -- are you sure it's a point-to-point and
If replacement HW is truly the same stuff, just ensure microcode (in library
*and* the drives) is up to date; after you re-run cfgmgr, you just re-do
the TSM DEFines for LIBRary and DRives.
BTW, checkin will skip volumes with data if you specify scratch (assuming
your TSM db is intact -- volhist
Bill,
Your arguments are excellent, and I have one large client dealing with the
health
records issue --- they need to save patient records for 32 years; we are
liking
the export solution, once a year, for those nodes (in order to control TSM
db
size) -- and, continue storing their data in archi
Chris,
I have looked at the Gresham EDT-DT info, am trying to understand why a
customer
might want to use it in a non-LAN-free environment.
I have a client who's installing a StorageTek SN-6000, which virtualizes the
tape drives
in a Powderhorn silo; they will use two TSM servers, initially --
Try running your SELECT queries; last I heard, there were various problems
with Summary table ---
which is a key helper to sizing a current environment. There may be problems
with other tables as well.
Hope to hear more info on what works (eg, Win2000 online-image backups), and
what doesn't.
Re
Bill,
Just out of curiosity, in your restore situation -- was it an NT platform (I
see it must have been alot of tiny files)?!?
If you are using NT, Netware or AIX as file server platforms, the DIRMC
option pays for itself... BIG TIME. I recently responded to a customer,
clearly a smaller shop,
Thomas,
This sounds like a recurring "bug"; did you try adjusting IDLETIMEOUT in
dsmserv.opt?
Alternatively, could you identify the session holding the tape (via q se
f=d)??? If so, then you can at least cancel the offending session (should
be in idle-wait, anyway).
Did Level 2 give you any h
Actually, why not continue using the same method already in place? There's
nothing inherently different between the two platforms that would require a
change -- only your operations preferences (and control) for managing the
inventory. There is (currently) no feature to "protect" one storage poo
Curtis,
Try setting summaryretention to 1, stop the server, change system date to 2
days later than the bad date, start dsmserv, ACCEPT DATE -- all of this with
sessions and schedules disabled, to prevent further records; after purge
processing, stop the server, reset system date, start dsmserv,
This is an *excellent*, best practices, suggestion... works fine, runs a
long time!
- Original Message -
From: "Francisco Molero" <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Monday, April 01, 2002 2:42 AM
Subject: Re: Disk Availability
> Hi,
>
> the best solution is 4 loops of 8 d
38 matches
Mail list logo
