If you want to take it a little deeper,  has anyone else noticed
that selective backups do not honor the exclude.dir statement?
Whether this is a feature or a bug I don't know.

James Thompson

From: Serge Gaudet/EIG <Serge.Gaudet AT ECONOMICALINSURANCE DOT COM>
Reply-To: "ADSM: Dist Stor Manager" <ADSM-L AT VM.MARIST DOT EDU>
To: ADSM-L AT VM.MARIST DOT EDU
Subject: Re: how is a weekly FULL bkup scheduled?
Date: Wed, 1 Nov 2000 08:40:03 -0500

...that's right.... selective backups do process the include-exclude
list...  I neglected to mention making reference to a separate dsm.opt file
when I perform selective backups.... (aliases do wonders in AIX, and when I
call for a selective backup, the behind-the-scenes details are taken care
of on my behalf.  After a while, you forget those details).  My apology.
The last thing I wanted, was mislead somebody.

Regards,
Serge Gaudet




                    "France, Don G
                    (Pace)"               To:     ADSM-L AT VM.MARIST DOT EDU
                    <don.france-ed        cc:
                    s AT EDS DOT COM>            Subject:     Re: how is a 
weekly
FULL bkup scheduled?
                    Sent by:
                    "ADSM: Dist
                    Stor Manager"
                    <ADSM-L AT VM DOT MAR
                    IST.EDU>


                    01/11/2000
                    03:46 AM
                    Please respond
                    to "ADSM: Dist
                    Stor Manager"





You need to realize that the include/exclude list is used for SEL as well
as
INCR;  if you must have varying kinds of overlapping/conflicting inclexcl
lists, you must use different node names --- all backup data is subject to
the inclexcl list.  (I understand this is not the case with NetBackup, but
that's another conversation;  I'm told it allows the equivalent of
different
management classes for each backup version/copy of a file --- no implicit
re-binding on subsequent backup sessions.)

A couple alternatives (a) archive (this was how we did things until v3.7
came along), or (b) generate backup set which consolidates all the active
(only) copies of backup files.

All the choices discussed have their various downsides, especially when
considering Oracle archive logs that are "gone" (ie, marked inactive due to
being deleted from the client file system), yet you want to be able to
reach
back, for point-in-time restore, beyond the most recent full backup to one
from a week or three ago, replaying logs from the most recent month-end
full
backup.

ARCHIVE can quickly become the most practical;  for databases, you can
repeatedly add to an archive package (eg, 1st daily full, plus all
subsequent archive logs, of a given month), changing the package
description
as part of each month-end archive process.

Regards,
Don


-----Original Message-----
From: Serge Gaudet/EIG [mailto:Serge.Gaudet AT ECONOMICALINSURANCE DOT COM]
Sent: Tuesday, October 31, 2000 10:43 AM
To: ADSM-L AT VM.MARIST DOT EDU
Subject: Re: how is a weekly FULL bkup scheduled?


I am of the opinion the "selective" makes sense to backup files that are in
an area normally excluded by the incremental process (via your
include-exclude definitions).  Have you ever considered Collocation?  At
least for your most critical server?  Just a suggestion.  Furthermore, a
server executed/generated archive or backupset would probably be better.
(avoiding the bottleneck... the network - hence executable during the day
as opposed to smack in the middle of the backup window).  I would just
perform incrementals otherwise.

What can I say:  Everywhere I went, network was always the primary issue.
I quickly learned the cost of using it, and learned to become effective
considering the constraint.

Serge Gaudet
Consultant, CMI




                    Tim Melly
                    <Tim.Melly.B@        To:     ADSM-L AT VM.MARIST DOT EDU
                    BAYER.COM>           cc:
                    Sent by:             Subject:     Re: how is a weekly
FULL bkup scheduled?
                    "ADSM: Dist
                    Stor Manager"
                    <ADSM-L AT VM DOT MA
                    RIST.EDU>


                    31/10/2000
                    12:59 PM
                    Please
                    respond to
                    "ADSM: Dist
                    Stor Manager"





We recently had a "minor" Data Center fire and I had to restore several
Novell
volumes. The Novell servers had been incrementally backed up for two years,
the
restoration times were horrendous (12 GB in 18 hours). The tapepool was 270
DLT
tapes and almost all of the tapes were mounted during the restore. We've
since
implemented every other weekend "selective/full" backups of our Novell
servers.
There are situations were "selective" backups make sense.

FYI *SMs, the Novell client does not always restore the file with the "long
filename". We had a significant amount of files which we restored to the
DOS
file name not the "long filename". We had to manually move the files to the
correct location when that location could be determined. Tivoli is looking
into
the issue.

Regards, Tim Melly




"DDA.RFC-822=ADSM-L AT VM.MARIST DOT EDU/P=Internet/A= /C=us" on 10/31/2000
12:00:55 PM
Please respond to "DDA.RFC-822=ADSM-L AT VM.MARIST DOT EDU/P=Internet/A= /C=us" 
@
X400
To: "DDA.RFC-822=ADSM-L AT VM.MARIST DOT EDU/P=Internet/A= /C=us" @ X400
cc:

Subject: Re: how is a weekly FULL bkup scheduled?

Why is this so strange?  I need the data for many servers for a months
worth
of data.  I have to be able to restore to any date they (user) says.  We
keep the tapes offsite and exercise our DR plan yearly.  I will be
restoring
whole servers.  Not a file here and there.  I would think a full backup and
a year of incrementals could get a little ugly.
Matt

-----Original Message-----
From: Serge Gaudet/EIG [mailto:Serge.Gaudet AT ECONOMICALINSURANCE DOT COM]
Sent: Tuesday, October 31, 2000 11:00 AM
To: ADSM-L AT VM.MARIST DOT EDU
Subject: Re: how is a weekly FULL bkup scheduled?


It's gotta be a tape stress test in progress.....  or a validation of the
reclamation process....

I don't know anyone in their right mind who would want to do such a
thing...
...but then again, it could be just to satisfy a paranoid ex-arcserv like
client user!

Serge Gaudet, CMI
"Couldn't help the wise crack - Have a great day"




                    Larry Way
                    <Larry.Way@TR        To:     ADSM-L AT VM.MARIST DOT EDU
                    W.COM>               cc:
                    Sent by:             Subject:     Re: how is a weekly
FULL bkup scheduled?
                    "ADSM: Dist
                    Stor Manager"
                    <ADSM-L AT VM DOT MA
                    RIST.EDU>


                    31/10/2000
                    10:49 AM
                    Please
                    respond to
                    "ADSM: Dist
                    Stor Manager"





My question to you is why would you want to ?

>>> Tim.Melly.B AT BAYER DOT COM 10/31/00 07:37AM >>>
Check "selective" backups.




"DDA.RFC-822=ADSM-L AT VM.MARIST DOT EDU/P=Internet/A= /C=us" on 10/31/2000
10:32:55 AM
Please respond to "DDA.RFC-822=ADSM-L AT VM.MARIST DOT EDU/P=Internet/A= /C=us" 
@
X400
To: "DDA.RFC-822=ADSM-L AT VM.MARIST DOT EDU/P=Internet/A= /C=us" @ X400
cc:

Subject: how is a weekly FULL bkup scheduled?

Hello,
        I am trying to figure out how to set up a scheduled job that will
do
a Full backup every weekend of a client.  I will be running daily
incrementals and will have the POLICY set to keep 4 copies of the backup.
I
just don't see how to get the client to do a full backup after the intial
has been done.   What am I missing?
Matt

_________________________________________________________________________
Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com.

Share information about yourself, create your own public profile at
http://profiles.msn.com.