Re: all of a sudden all estimates are off...
2006-07-25 16:42:54
* Jon LaBadie <jon AT jgcomp DOT com> [20060725 16:22]:
> On Tue, Jul 25, 2006 at 04:13:08PM -0400, Jean-Francois Malouin wrote:
> > Hello,
> >
> > I'm experiencing estimates that are way off with the result that level
> > 0s and 1s use the same space on tapes and now obviously that
> > throws the balance completely and every amdump eats through a lot
> > of tapes.
> > I'm using calcsize in my DLEs but it's the first time that the
> > estimate are completely off the tracks. It affects DLEs using gnutar
> > and xfsdump so I'm quite baffled by this one.
> >
>
> Might your dumptype(s) have a "record no".
Just double checked and no, record is set to yes:
yorick /data/noel/noel1 {
xfs-calc-100
} 2
define dumptype xfs-calc-100 {
global
program "DUMP"
priority high
compress none
maxdumps 4
estimate CALCSIZE
}
define dumptype global {
comment "Global definitions"
index yes
record yes
}
There is even stranger stuff that I just noticed:
sendsize[1930073]: estimate size for /data/mril/mril5/bojana/images
level 0: 35142619 KB
sendsize[1930073]: estimate size for /data/mril/mril5/bojana/images
level 1: 13554 KB
sendsize[1930073]: estimate size for /data/mril/mril5/bojana/images
level 2: 13554 KB
But the amdump log reports that :
driver: result time 31411.774 from dumper0: DONE 01-00040 35153370 35153370
6180 [sec 6179.877 kb 35153370 kps 5688.4 orig-kb 35153370]
driver: finished-cmd time 31412.169 dumper0 dumped
yorick:/data/mril/mril5/bojana/images
driver: send-cmd time 31412.170 to taper: FILE-WRITE 00-00041
/holddisk/conf2/stk_40/20060725031507/yorick._data_mril_mril5_bojana_images.1
yorick fffffeff9ffe7f /data/mril/mril5/bojana/images 1 20060725
D0 I see correctly that 35153370kb made it to tape which a tad bit
more than the level 0 estimate? (and at a crawling speed too...)
weird. jf
>
> --
> Jon H. LaBadie jon AT jgcomp DOT com
> JG Computing
> 4455 Province Line Road (609) 252-0159
> Princeton, NJ 08540-4322 (609) 683-7220 (fax)
--
<° ><
|
|
|