question about taper/planner sequencing

To some extent, this is just a curiosity question. I was asked this,because the Amanda Mail reports go to all the department sysadmins sothat they can be aware and keep an eye on things if I'm gone. Anyway, onweekends, I use the -o configuration override to point amdump to anon-existent tape at /dev/rmt/0n. Amanda responds by doing incrementalsto the holding disk and then flushing that to tape when the mondaybackups run.

So, the output on weekends starts with "A Tape Error Occured", which Ihave told them to ignore (on Sunday morning and on Monday morning). Thenit gives the statistics and the notes. The statistics are exactly what Iwant and expect. The notes start with the taper error, and then go onwith a bunch of lines saying that planner is promoting some full dumps.

The question is, "why does the planner go on about promoting full dumps,when taper already knows there is no tape?"

I'm guessing the planner goes through a bunch of initial phaseestimation and planning before it applies some kind of final decisionmaking algorithm. But I'm just waving both arms in the air. Can anyonesay whether that's a reasonable description? Or whether there might besome shorter path the planner can take to the decision to runincrementals only, without making the rest of the code end up lookingmore kludgy? I recognize that this is a special case and doesn't warrantpriority over the more general case when they are weighed in the balance.



---------------

Chris Hoogendyk

-
  O__  ---- Systems Administrator
 c/ /'_ --- Biology & Geology Departments
(*) \(*) -- 140 Morrill Science Center

~~~~~~~~~~ - University of Massachusetts, Amherst

<hoogendyk AT bio.umass DOT edu>

---------------

Erdös 4