On Wed, 21 Jul 2004 at 4:35pm, CPH wrote

> On Wednesday 21 July 2004 16:12, Joshua Baker-LePain wrote:
> > On Wed, 21 Jul 2004 at 4:08pm, CPH wrote
> >
> > > Hi,
> > >   I have very simple needs for an amanda installation :
> > > - backup one server which also contains the tape drive;
> > > - incremental nightly backup - essentially 2 databases ;
> > > - full weekly backup;
> > > - I have a dump disk with 1G free
> >
> > full weekly backup -> dumpcycle 1 week
> > runspercycle 5 (assuming running only on weeknights)
> > tapecycle ? (How many tapes do you have -- preferably 10+).
> >
> > dumpcycle is the maximum amount of time between full dumps.  
> Ok, the above is now much clearer! 
> So : 
>   dumpcycle = duration between complete backups;
>   runspercycle = no of incremental backups between complete backups;
> but tapecycle ?
> I have 7 tapes, and I was thinking I would use 5 for the incremental backups
> and 2 for the full backups. Does that make sense ?

Amanda works differently from the standard backup scheduling model.  
Amanda's goal is to backup approximately the same amount of data each 
night.  To do this, it spreads the level 0s through the dumpcycle.  
dumpcycle specifies the max interval between level 0s, although amanda is 
free to do them "early" if it wants.

So no tape is dedicated to fulls or incrementals.  You just tell amanda 
how often you want level 0s (dumpcycle), how many times in that time 
period you'll run amdump (runspercycle), and how many tapes you're going 
to let amanda use (tapecycle), and amanda does the actual scheduling.  
With a runspercycle of 5, a tapecycle of 7 is OK, but 10 would be much 
better.  

> > As for  documentation, have you read "the chapter"?
> > <http://www.backupcentral.com/amanda.html>
> I tried, but it makes much more sense with your explanation.
> It's probably me, as I could find almost no other docs on the web to me a an 
> "amanda for dummies" type of explanation.

After the chapter, the best docs are probably docs/INSTALL and the list 
archives.

-- 
Joshua Baker-LePain
Department of Biomedical Engineering
Duke University