[Savannah-hackers-public] rsync process load on vcs

[Karl Berry]

(Ward, please see query at end.)

When I looked at vcs this morning, the load average was around 15.  I
just ran top, and it was because of this rsync job, which had already
been running for over an hour (the start time is UTC today):

USER       PID  PPID %CPU %MEM   RSS    VSZ STAT  NI  STARTED COMMAND
root     25624 25618  2.0  0.7 43936  73600 Ds     0 12:22:12 rsync
--server --sender -lHogDtpAXre.iLsf --timeout=3600 --ignore-errors
--numeric-ids . /

I reniced it and things got back to what passes for normal, but the
question is, where did it come from.  The parent was just an ssh:

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND           
25618 root      20   0 13140 6844 2572 S    2  0.1   0:09.86 sshd: 
address@hidden  

lsof -p showed it accessing a file under
/var/lib/cvs2git/mirror-git-to-cvs/.state/gettext/cvsrepo/...
Unfortunately I didn't have time to delve further as to the other end of
the network connection, etc.

I grepped for numeric-ids under /var/lib/cvs2git and /etc/xinetd.d and
/usr/local/bin to no avail.  The mirror-cvs-to-git script does invoke
rsync, but not with those options.  The rsync from xinet is always
niced (but this is clearly different than that anyway).

The final . / on the command makes me wonder if it is some kind of
system backup.  Ward, could that be it?  And if so, could you please
adjust it to run with nice -19 :)?

Thanks,
Karl