was code added to detect or die on sighup recently?

L. A. Walsh rsync at tlinx.org
Tue Aug 5 00:16:14 MDT 2014

Kevin Korb wrote:
> Hash: SHA1
> Any non-daemon software is supposed to exit (and cleanup) when hit
> with a hangup signal.  Pretty sure it isn't new to rsync.
    But I didn't send it a hangup signal.  By calling snaprun
in a subshell (snaprun "$@" </dev/null &), it spins it off
into a separate process group that doesn't get signaled
unless I specifically address it.

    I did change the invoking script to force
'output_wanted' to true so it would do a tail-f of
the log file to the console so I could monitor it more
easily, but I wouldn't have thought that would have added
any hangup signal.


    FWIW, it is still creating all the empty dirs....

   Oldest Snapshot = Home-2014.08.03-02.05.53.
Rsync with 9 excludes from config file...
rsync took 61m, 9s
Empty-directory removal took 0m, 59s
Create vol. Home-2014.08.03-02.05.53, size 4.5G
Copying diffs to dated static snap...Time: 0m, 6s.

rsync took 61m, 9s to gen 4.5G of diffs to the "diff" volume, then
ran through and removed all empty dirs -- took almost another full
minute!   Then to copy the contents to the target dir took 6s.

(am sure it took longer than 6s, but copy exited after that long).

More information about the rsync mailing list