[rear-users] PXE tftp no space failure

Johannes Meixner jsmeix at suse.de
Thu Mar 23 09:55:44 CET 2017


On Mar 22 18:25 Styer, Darryl wrote (excerpt):
> 2017-03-21 12:14:07.288509428 Including output/PXE/default/80_copy_to_tftp.sh
> cp: error writing '/var/lib/rear/output/ead7002head7002h.initrd.cgz': No space left on device
> Apparently there's lots of left over junk in "/var/lib/rear/output"
> (including several kernels and ramdisks) and in my case not excluded
> and propagating to each subsequent backup.

On my test system where I did run many "rear mkbackup" I have only
# find var/lib/rear/output
so it depends on your particular ReaR setup that things
may pile up in var/lib/rear/output (or elsewhere).

In general "rear mkrescue/mkbackup" must not change the existing
system - in particular it should never ever delete whatever files.

The only place is ReaR's temporary working area $TMPDIR/rear.XXXX
that gets by default removed when "rear mkrescue/mkbackup" exits
(except KEEP_BUILD_DIR="yes" is set).

All other files in normal system directories should be sacrosanct
for ReaR (perhaps except some very very special exceptions).

This means that in general the admin needs to manually clean up
what is no longer needed in normal system directories.

I think ReaR should never ever make its own decision what files
in normal system directories seem to be no longer needed and
could therefore automatically be removed by ReaR,
cf. the comment regarding BACKUP_RESTORE_MOVE_AWAY
in usr/share/rear/conf/default.conf.

Because ReaR is meant to be adapted and enhanced as you need it,
you can automate your cleanup as you need it in your particular
case, see in particular things like PRE_BACKUP_SCRIPT and
POST_BACKUP_SCRIPT in usr/share/rear/conf/default.conf.

Kind Regards
Johannes Meixner
SUSE LINUX GmbH - GF: Felix Imendoerffer, Jane Smithard,
Graham Norton - HRB 21284 (AG Nuernberg)

More information about the rear-users mailing list