ADSM-L

Re: EXCLUDE.DIR and EXCLUDE.FS

2000-12-13 16:45:10
Subject: Re: EXCLUDE.DIR and EXCLUDE.FS
From: Scott McCambly <mccambly AT ATTCANADA DOT CA>
Date: Wed, 13 Dec 2000 17:01:47 -0500
Hi All,

I wanted to add a couple of comments to this also...

First - I'm not too sure what the EXCLUDE.ARCHIVE has bought us since we
always had the option of specifying a management class on an include
statement that doesn't have an archive copy group. Very effective in
preventing archives of those specific files. Anyway, I like flexibility, so
the more options the better (and a global 'exclude.all' would be welcome).

Second - Michel has a good point, but I think his problem would be more
easily solved by allowing the VIRTUALMOUNTPOINT option to be used on NT as
well as UNIX. I handle this problem on UNIX by defining the directories as
virtual mount points and then only include them in my DOMAIN list. It
doesn't prevent intentional backups of the other directories, but it works
great for scheduled backups.

Thanks for listening to my $0.02 (and thats Canadian, so USD $0.013333)

Scott

At 10:11 AM 12/13/00 +0100, you wrote:
>Hi Keith,
>
>Now that we are again discussing over includes and excludes I had an idea.
More
>than once my clients ask me to exclude all directories except one or two.
>Specially directories that are on the C-drive of a WindowsNT environment.
Would
>it be difficult to foreseen an option like "INCLUDE.DIR" then I could specify
>something like.
>
>Exclude.dir "c:\*"
>Include.dir "c:\TSM"
>
>What do other think of that.
>
>Michel
>TSM Consultant
>
>
>
>
>
>Keith Davey <kdavey AT GUS33.HOMEIP DOT NET> on 12/12/2000 11:09:59 PM
>
>Please respond to "ADSM: Dist Stor Manager" <ADSM-L AT VM.MARIST DOT EDU>
>
>To:   ADSM-L AT VM.MARIST DOT EDU
>cc:    (bcc: Michel Engels/BE/Devoteam)
>
>Subject:  Re: EXCLUDE.DIR and EXCLUDE.FS
>
>
>
>On Tue, 12 Dec 2000, Rushforth, Tim wrote:
>
>> And speaking of Excludes ...
>>
>> The Exclude.archive is a welcome feature, but we tend to have to duplicate
>> excludes,  one line to exlude from backups, another line to exclude from
>> archives.
>>
>> Is there a way to exclude a file from both backups and archives with one
>> statement?
>
>For both backup and archive operations?  No.  Not to the best of my
>knowlege.
>
>>
>> Also, what is the difference between exclude, exclude.backup and
>> exclude.file.backup?
>>
>
>exclude :  Excludes a file or group of files from backup services and
>           space management services (if theTSM HSM client is
>           installed). Any file in your client domain that is not
>           specifically excluded with this option is considered for
>           backup.
>
>exclude.backup :
>                Excludes a file or a group of files from normal backup
>                services, but not from HSM.
>
>exclude.file.backup:
>                Excludes a file from normal backup services.
>
>Hope this is helpfull.
>
>Keith Davey
>Tivoli Systems
>
>> Thanks,
>>
>> Tim Rushforth
>> City of Winnipeg
>> trushfor AT city.winnipeg.mb DOT ca
>>
>> -----Original Message-----
>> From: Richard Sims [mailto:rbs AT BU DOT EDU]
>> Sent: Tuesday, December 12, 2000 2:39 PM
>> To: ADSM-L AT VM.MARIST DOT EDU
>> Subject: EXCLUDE.DIR and EXCLUDE.FS
>>
>>
>> >  Now, if I put Exclude "?:*" at the top of this list, no other files on
>> >  the disk are backed up but the directory structure is. Not terrible from
>> >  a space standpoint but could be pretty confusing for a user trying to
>> >  restore something. If, however I make the top (last) statement
>> >  Exclude.dir "?:*" then nothing but files at the root of C: are backup up
>> >  !? In other words it overrides all the include statements below it,
>> >  what's up with that?
>>
>> Jim - Tivoli has failed us on this one.  I just stumbled upon
>>       a documentation APAR that clears up this mystery...
>>
>> ------------
>>  APAR NUMBER: IC25492           RESOLVED AS: DOCUMENTATION ERROR
>>
>>  ABSTRACT:
>>  IC25492: PRIORITY OF EXCLUDE.DIR IN INCLEXCL IS NOT WELL DESCRIBED IN
>>  DOCUMENTATION.
>>
>>  ORIGINATING DETAILS:
>>  If the EXCLUDE.DIR is specified in DSM.OPT, it has the highest
>>  priority. All other specified includes/excludes were processed
>>  later.
>>  This behaviour of the EXCLUDE.DIR does not depend on the
>>  position of this statement in the incl/excl list.
>>  The documentation says:
>>  These options (include, exclude and exclude.dir) are checked
>>  from the bottom of the include-exclude list up, until a match
>>  is found.
>>
>>  LOCAL FIX AS REPORTED BY ORIGINATOR:
>>
>>  RESPONDER SUMMARY:
>>  ****************************************************************
>>  * USERS AFFECTED: TSM Windows Client*
>>  ****************************************************************
>>  * PROBLEM DESCRIPTION: The processing priority of an *
>>  * exclude.dir statement within an include-exclude list is not*
>>  * documented in the book.*
>>  ****************************************************************
>>  * RECOMMENDATION: Add documentation describing the processing*
>>  * priority of an exclude.dir statement within an*
>>  * include-exclude list.*
>>  ****************************************************************
>>  *
>>
>>  RESPONDER CONCLUSION:
>>  The following text was added to TSM Windows
>>  Client book: "Options are checked from the bottom of the list
>>  up, until a match is found. However, an exclude.dir statement
>>  takes precedence over other include-exclude statements.
>> ------------
>>
>> The worst part is that despite this APAR being more than a year old, the
>> specified text is still not in the TSM Windows Client book - or any other
>> client book!
>>
>> The Include-Exclude options are the single, worst, nightmarish part of the
>> TSM
>> product, and have been for years - now made worse by inconsistencies in the
>> once-rigorous rule of "bottom-up" processing, as EXCLUDE.DIR and EXCLUDE.FS
>> take precedence over all other Include/Exclude statements regardless of
>> position.  This whole area is in serious need of a full, complete
overhaul -
>> and cohesive, understandable documentation, with comprehensive examples.
>> What
>> is inflicted on customers now is abominable in its lack of logical
>> consistency
>> and understandable interrelationships.  Tivoli needs to finally appreciate
>> that many, many hours of customer time are being wasted struggling with the
>> Include/Exclude options, thanks to design and documentation problems.
>>
>> Richard Sims, BU
>>
>
>
Scott McCambly
AIX/NetView/ADSM Specialist - Unopsys Inc.  Ottawa, Ontario, Canada
(613)799-9269
<Prev in Thread] Current Thread [Next in Thread>