Yes, with frequency based scheduling, keeping the frequency as small as
possible (yet larger than your window is expected to be) will let you do
a manual retry the next day without adversely affecting the backup next
week.
However, during the past year I have become a fan of calendar based
scheduling. The advantage is that you can setup the retry to occur
automatically!
To make this happen for your example, you simply set your calendar day
to every Tuesday, while making windows for the schedule on both Tuesday
and Wednesday. If the Tuesday backup "works", you'll have no Wednesday
backup; if the Tuesday backup fails, you'll have the Wednesday retry.
Slick.
I needed to spend less time with NetBackup and calendar based scheduling
helped a bit.
cheers, wayne
Martin, Jonathan (Contractor) wrote, in part, on 5/15/2006 9:46 AM:
> After auditing a few backup servers for a couple of weeks now, I've come
> to realize that if a certain weekly backup job fails on any given day,
> and I re-run it manually during the day/week following that it does not
> fire off again next "week"! Why? Because (according to Veritas) when
> the job schedule triggers, it finds the last successful backup (of that
> schedule) was within the last week, and does not "need" to fire. This
> is the way 99% of my full backups are scheduled.
>
> Example -
>
> Full Backup Scheduled Tuesday Night at 8:00PM - 3 month retention -
> Frequency Based Scheduling - 1 week. If it fails and I rerun it on
> Wednesday then it doesn't fire the next Tuesday, because its run in the
> past 7 days. =/
>
> Anyhow, how are people working around this "issue." I've come up with
> two possible solutions.
>
> 1 - Calendar based scheduling for "every week" on Tuesday
> 2 - Change my frequency to 2 or 3 days - assuming the window only opens
> that one day.
>
|