Veritas-bu

[Veritas-bu] Veritas Netbackup vs Tivoli (ADSM)

2000-11-02 15:21:26
Subject: [Veritas-bu] Veritas Netbackup vs Tivoli (ADSM)
From: Jonathan Meyer jmeyer AT ptc DOT com
Date: Thu, 2 Nov 2000 15:21:26 -0500
My experience with pre-emptive scheduling in netbackup is that it does
work.  However, at the time I deployed our installation, almost two
years ago, the documentation was very unclear.

If you have two schedules with EXACTLY the same windows then only one
of them will run.  If they are both due, then the one with the longer
frequency will run.  What I found unintuitive is that if the windows
have the same start time but different end times, then they may both
be run.  I haven't looked lately, but I don't remember ever seeing
this behavior clearly described in the documentation.

I also think there is room for debate if this is the correct behavior.
An obvious problem occurs if the window is changed for one schedule
and not another.  I have had to write a kludgy shell script which
copies windows from one schedule to another in order to correct these
configuration mistakes in an automated way.

--------------------------------------------------
Jonathan Meyer
(781)398-6594
UNIX Systems Administrator
Paramtric Technology Corporation
--------------------------------------------------



   On Thu, Nov 02, 2000 at 12:39:32PM -0600, John_Wang AT enron DOT net filled 
up my inbox with:
   > 
   > Inability to pre-emptive schedule.   If the scheduler happens to have an
   > incremental and a full backup scheduled for the same day, I would like to 
have
   > the incremental automatically pre-empted as unnecessary since the full 
would
   > cover everything that the incremental would've anyways.
   > 

    This is supposed to work the way you wish. In fact, the behavior
    that has been told to me by consulting, support, and other users that
    it is possible to have a full, incremental, and diff running at the
    same time, or on the same day, is a bug. While at Vision, I spoke to
    Mark Winger, who is principle developer of the scheduler. He had never
    heard of this behavior before and is very interested in any documented
    cases being opened as a support call so this bug can be worked out.