On 08/27/2014 06:30 AM, Tiago Bortoletto Vaz wrote:
Hi Lorenzo,
Thanks for your detailed report, I was not aware of this issue.
However, I'm not sure if I understood your proposal. If we include the
old cron daily script we'll end with at least two issues which were
fixed by removing it. One is having two instances of apticron running
(potencially at the same time). Other is having the cron daily running
apticron at the same time, which can break network if you have many
systems using it in your network.
Sorry if I misunderstood, I'm open to deal with this systemd-cron thing,
but I can't think of a simple solution now.
I'm not a fan of systemd-cron(*); I was going to ask to switch the
dependency from cron to cron-daemon, but that wouldn't be feasible
because systemd-cron is a __broken__ replacement for cron (until bug
752376 is fixed), so I proposed running apticron in cron.daily to work
around that issue.
Since it's systemd-cron's fault:
- the "clean" solution would be to depend on cron-daemon with a note
(changelog, "Breaks:", whatever) that systemd-cron needs to be fixed
- the hackish solution, to add an entry in cron.daily and __only__ run
it if systemd-cron is detected (something like "test -e
/lib/systemd/system/cron-daily.service && (run apticron)")
- any better solution is of course welcome :)
Btw, the guy who filed the bug against bcron is actively interested in
systemd-cron, so I'm CC'ing him - hi Alexandre!
Regards
(*) I went to popcon.debian.org just out of curiosity, and it looks like
systemd-cron is ~2x more popular than bcron - dammit!
--
To UNSUBSCRIBE, email to [email protected]
with a subject of "unsubscribe". Trouble? Contact [email protected]
