We just ran across this error, because we hit the 2GB filesize limit on ext2
filesystems. After reading the man pages, we've cranked down the
holdingdisk chunksize to 2000MB, which should alleviate the problem in the
future.
However, the behavior after this error seems a bit odd. Amanda flushed the
other disk files on the holding disk to tape ok, but then also left them on
the holdingdisk. amverify confirms they're all on tape, but ls shows
they're all still on the holdingdisk, too. Is this normal? I would have
expected amanda to flush all the files to tape, remove them from the holding
disk, and then retry the level 0 backup that failed on the next run.
I've forced a level 0 of the failed disk for the next run (probably
unnecessary, but what the hey), but I guess I need to run amflush again on
the holding disk.
Regards,
Carey
