On Feb 28, 2001, Gerhard den Hollander <[EMAIL PROTECTED]> wrote:
> * Alexandre Oliva <[EMAIL PROTECTED]> (Wed, Feb 28, 2001 at 09:22:10AM -0300)
> I did ^C a test amdump for this config during the estimate phase.
> Could it be that amandates is corrupt ?
Well, it might be, but it certainly wouldn't cause this behavior,
since the planner sends the estimate request before the client has a
chance to look at the file.
> How does planner figure out which levels to ask for ?
Basically, it asks for an estimate for a full, an incremental of the
same level as the last run and, sometimes, another incremental one
level above that, depending on bumpdays.
> command: 5
This means FORCE_FULL | FORCE_NO_BUMP. I suppose FORCE_NO_BUMP should
be ignored if FORCE_FULL is set. Ditto just after running a full. Or
does it make sense to prevent a bump to level 1 after a full backup?
> The 2 disks that give this weird problem both have a command: 5
> all other disks have a command: 0
> What is the impact if I remove the info file for the 2 offending disks ?
Amanda would consider these as new disks on the next run, and would
run full backups.
You can probably fix the problem with amadmin unforce-bump.
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicamp oliva@{lsd.ic.unicamp.br, gnu.org}
Free Software Evangelist *Please* write to mailing lists, not to me
