Re: unable to complete huge backup
[email protected] (Joerg Schilling) Fri, 07 Dec 2007 19:22:42 +0100
| Newsgroups | gmane.comp.archivers.star.user |
|---|---|
| Message-ID | <47598f72.xFcjDgFQxzPqWfb+%[email protected]> |
Udo Rader <[email protected]> wrote: > Hi, > > I am quite lost with a somewhat huge backup (approx 120GB) that > completes "sometimes" but hangs most of the time at some unpredictable > point. > > The backup is done this way: > > % star b=4096 errctl="WARN|GROW|SHRINK|OPEN|READ|READLINK|MISSLINK *" -c > -acl artype=exustar -f /dev/st0 ${BACKUPDIR} >>${LOGFILE} 2>&1 > > When I say "hang" it means that nothing is written to the tape any more > but instead I see two star processes "hanging" around doing virtually > nothing, both are in S state (=sleep) and stracing them does not show > any activity. > > Adding -vv to the star command above I saw that it hung on one arbitrary > file. > > Some stats: > ???star 1.5a67 > debian etch, running on a 2.4.27-3-686-smp kernel > Quantum DLT-V4 on a Adaptec AIC7892 Ultra160 SCSI Controller This is an extremely old version (> 2 years). Please upgrade to the recent 1.5a87 it contains a bug fix for a mutex problem with multi core systems. ~A