Bug#780461: updates from #debian-systemd

2015-04-21 Thread Wolfgang Schweer
On Tue, Apr 21, 2015 at 03:18:18PM +0200, Petter Reinholdtsen wrote: > [Wolfgang Schweer] > > Seems to be /etc is writable at the time. Maybe the reason is that > > 'BOUND' isn't evaluated, only 'RENEWAL'? > > I tought both were handled in the script? Is there something wrong with > that part?

Bug#780461: updates from #debian-systemd

2015-04-21 Thread Petter Reinholdtsen
[Wolfgang Schweer] > Seems to be /etc is writable at the time. Maybe the reason is that > 'BOUND' isn't evaluated, only 'RENEWAL'? I tought both were handled in the script? Is there something wrong with that part? > I guess I'll file a separate bug about this issue to get it sorted. Yeah. >>

Re: Bug#780461: updates from #debian-systemd

2015-04-21 Thread Holger Levsen
Hi, (just for the avoidance of doubt..) On Dienstag, 21. April 2015, Wolfgang Schweer wrote: > > Also, why is the sethostname() function in the > > dhclient-exit-hooks.d/hostname script? It seem to be completely > > unused. Remove it? > Yes, seems to be leftover cruft. And the log() function co

Bug#780461: updates from #debian-systemd

2015-04-21 Thread Wolfgang Schweer
On Mon, Apr 20, 2015 at 11:00:17PM +0200, Petter Reinholdtsen wrote: > > Could this be caused by dhclient running at a point during boot where > /etc/ is read-only, causing /usr/sbin/update-hostname-from-ip to exit > with an error code instead of updating the hostname? It is my best > guess for w

Bug#780461: updates from #debian-systemd

2015-04-20 Thread Petter Reinholdtsen
[Wolfgang Schweer] > After looking at this bug once again, there seem to be two issues: > > (1) The reported delay of (up to) 15 minutes is due to the fact that > dhclient updates the hostname on the first lease renewal. (The > interval length between renewals seems to be a random value >

Bug#780461: updates from #debian-systemd

2015-04-20 Thread Wolfgang Schweer
On Wed, Apr 15, 2015 at 05:06:21PM +0200, Holger Levsen wrote: > > from #debian-system on IRC: > > 15min sounds a bit excessive and doesn't look like the > internal 90s timeout systemd uses for services to start After looking at this bug once again, there seem to be two issues: (1) The report

Bug#780461: updates from #debian-systemd

2015-04-15 Thread Holger Levsen
control: tags -1 moreinfo Hi, from #debian-system on IRC: hi. do you have any idea about what causes / how to fix #780461? (delayed hostname update via DHCP) h01ger: can you pastebin the hook somewhere? h01ger: is that behaviour specific to systemd or also happening under sysvinit? how exa