My determination of what the LID (I presume you are talking about the dates
in the filespace table) is goes along with what Alex and Andy have said. I
use this date only as an indicator that the last incremental backup against
that filespace completed. That is what updates the date based on what I
have seen. The only time it has not updated is the session dropped through
the middle of the backup because of a server failure or a client code
failure. I actually run a select comparing the begin and end dates so I can
catch the problems.
I would be very careful to not make any other determinations from those
dates in the filespace records.
When I first started with TSM about a year ago, this area was the most
confusing mess of documentation I have every seen. It even led you to
believe that you could specify a partial incremental with a management class
that said absolute. I could never understand why the filespace backed up
entirely. It was because the documentation was wrong.
The INCREMENTAL command should have been obsoleted years ago with only a
compatibility and a new set of commands FULLBACKUP, UPDATEFULLBACKUP, and
PARTIALLYUPDATEFULLBACKUP put in its place so that customers would not be so
darn confused. The terminology incremental is just flat confusing. I spend
so much time explaining it to new users it is ridiculous. There have been
sales of TSM lost because of this confusion.
At least if they divided it up into these three commands they may be able to
reduce the confusion as to what happens during the process. Such as,
filespace dates updates, expiration processing invoked, etc. Right now it
is like reading a contract with ultra fine print. Maybe an act of Congress
to require the fine print be written in laymen's terms would be in order.
The best test to see if they have fixed the confusion is to take 20 college
graduates, give them the documentation and 2 hours to complete a script of
backups/restores that would normally take 1 hour. If they cannot figure it
out, in 2 hours the documentation is not documentation.
Even the GUI is confusing on what a partial or full incremental is. I have
seasoned 25 year professionals that cannot understand what the heck is being
talked about.
I hope this clears the confusion up. I am still confused.
Paul D. Seay, Jr.
Technical Specialist
Naptheon Inc.
757-688-8180
-----Original Message-----
From: Alex Paschal [mailto:AlexPaschal AT FREIGHTLINER DOT COM]
Sent: Monday, August 12, 2002 7:30 PM
To: ADSM-L AT VM.MARIST DOT EDU
Subject: Re: Incremental Backup (full/partial)
Ah, look, Mark, I really hate to do this, because it's making me look like a
horse's hind end for not being graceful about this subject, but just now I
did a...
dsmc inc -incrbydate /
and it updated the Last Backup Completion Date (LID) in a "query filespace
f=d" for the / filespace. LID doesn't tell you whether you're doing
Partials or Fulls. The fact of the matter is LID has nothing to do with
Partial backups.
Look, I'm not trying to attack you, I'm not trying to prove you're wrong or
I'm right, and, believe it or not, I'm not trying to nitpick semantics. I'm
just saying that misleading terminology or terminology used in a misleading
way can mislead those who are new to TSM, haven't taking your class, and are
looking to this mailing list for help and education. Specifically, in this
case, the thread starter, Ken Horacek, who at this point doesn't know
whether to use the book's terminology or yours to answer the question he has
about what he read in the book.
Alex Paschal
Storage Administrator
Freightliner, LLC
(503) 745-6850 phone/vmail
-----Original Message-----
From: Mark D. Rodriguez [mailto:mark AT MDRCONSULT DOT COM]
Sent: Monday, August 12, 2002 3:16 PM
To: ADSM-L AT VM.MARIST DOT EDU
Subject: Re: Incremental Backup (full/partial)
Alex,
I have read through your response and I can understand your position.
However, firat I would like to point out that the TSM documentation is not
always clear and is not always consistant. The text you quoted from the
Admin Guide is an example of them not being consistant. Clearly what they
are describing there is a "incremental -incrbydate" type backup. And as I
stated earlier you can consider it a "partial incremental", but that does
not mean that a "partial incremental is only acheived by "incremental
-incrbydate", that's like saying "I live in Texas therefore I am a Texan and
an American" based on that fact I can say all Texan's are Americans, but not
all Americans are Texans.
I will concede that there is places within the documentation that refers to
the "-incrbydate" option as beeing a "partial backup" and I can show you
IBM/Tivoli education material that describes a partial exactly as I did in
my note. But rather than nit picking the symantics I would like to
re-phrase my explanation, Instead of calling it "full incremental" and
"partial incremental" maybe we should use full and non-full. The key here
is what happens when you don't use "full incrementals", in particular the
Last Incremental Date ( what I refer to as the LID) does not get updated.
This is a critical peice of information. Much of the documentations
explanation for other processes are assumming that you are doing fulls since
it keys off of the LID. In addition, what other processing is being
effected by your "non-full incremental" (filespec limited or -incrbydate
option), i.e. file expiration, rebinding, missed files etc.
The point that I am really trying to make is you should always be doing full
incremental backups! The only time to consider anything else is if there is
a severe time constraint on the backup window.
I think this thread has been great. It has given people a look at how, what
and why TSM is doing what it does.
--
Regards,
Mark D. Rodriguez
President MDR Consulting, Inc.
============================================================================
===
MDR Consulting
The very best in Technical Training and Consulting.
IBM Advanced Business Partner
SAIR Linux and GNU Authorized Center for Education
IBM Certified Advanced Technical Expert, CATE
AIX Support and Performance Tuning, RS6000 SP, TSM/ADSM and Linux Red Hat
Certified Engineer, RHCE
============================================================================
===
|
