Please keep replies on list, so that everyone sees them and can chip in.

On Tue, 30 Aug 2005 at 10:21am, Guy Dallaire wrote

> 2005/8/30, Joshua Baker-LePain <jlb17 AT duke DOT edu>:
> > On Tue, 30 Aug 2005 at 9:39am, Guy Dallaire wrote
> > 
> > > sendbackup: time 6032.923:  53:    size(|): Total bytes written:
> > > 5467463680 (5.1GiB, 886KiB/s)
> > > sendbackup: time 6032.969: pid 12853 finish time Tue Aug 30 06:20
> > > :19 2005
> > >
> > > ---------------
> > >
> > > It started at 4h39 and finished at 6h20 ! It really took a LONG time.
> > > Why, I don't know.
> > 
> > We really need more info -- client OS, type of filesystem, etc.  Also, are
> > there a lot of small files?  Many FSs slow down when there are lots of
> > inodes in use.
> > 
> 
> A full dump of the same disk on august 27th resulted in 3.1 GiB
> (Compared to 5.1 GiB) on august 30. As far as I know, there should not
> be such a big difference.
> 
> This is a solaris 9 Box running amanda 2.4.5. 
> 
> There is not a lot of small files as far as I know. 
> 
> As I told you, the problem only seems to occur on level 1 backups.
> Level 0 backups always seem to complete normally.

I'd look at both tar and the filesystem closely -- why does tar think the 
level 1 should be so big?  Is there a problem with tar (what version are 
you running), or is something odd going on with the filesystem?

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