Amanda-Users

Re: forcing skipped incrementals into holding disk

2005-02-15 23:27:43
Subject: Re: forcing skipped incrementals into holding disk
From: Frank Smith <fsmith AT hoovers DOT com>
To: Jamie Wilkinson <jaq AT spacepants DOT org>, amanda-users AT amanda DOT org
Date: Tue, 15 Feb 2005 22:13:15 -0600
--On Wednesday, February 16, 2005 11:51:11 +1100 Jamie Wilkinson <jaq AT 
spacepants DOT org> wrote:

> This one time, at band camp, Frank Smith wrote:
>> After further thought I believe Alex was correct about the
>> scheduler limiting itself to what will fit, so increasing
>> runtapes to 2 should let you get incrementals of everthing
>> on the day the full of your big DLE occurs.
> 
> I'm not seeing a post by Alex in this thread; can you perhaps help out my
> poor brain with a URL?

I don't see it in the archive either.  Perhaps he has an X-no-archive
header set in his email, or perhaps I quoted the wrong person, or maybe
I was just conversing silently with one of my other personalities ;-).

Anyway Alex (if that was the correct name) made the point that the scheduler
would only schedule tapesize*runtapes  bytes of backups, so that since the
large DLE was due for a full, the scheduler picked that and just enough
incrementals to fill the one allowed tape, skipping completely the rest of
the DLEs, deferring them to a future run.

It made sense once I thought about, so I was retracting my statement that
the incrementals should have gone to available holdingdisk space, so there
was no need to dig in the debug logs.

Frank