[04:04] http://heh.fi/tmp/idlemeter [13:12] if I have "start on started foobar", will it start only after the post-start stanza of foobar ? [15:57] Can anyone help me make upstart restart a getty after logging out of a serial console? [15:58] I login through the cyclades and it works. After logging out, getty stops [15:59] And I've created /etc/event.d/serial to look just like: http://pastebin.com/d31e43fff [16:12] respawn? [16:14] Is that what it is missing? [16:14] Sorry there is no man page and I've been fighting with this [16:19] SEJeff: try adding a line "respawn" to the event.d file instead of the emit thing in your post-start script [16:19] err, post-stop [16:26] Just comment the initctl line and put, "respawn"? [17:43] Maybe someone can give me some insight/ideas on what to check on next. I am trying to implement the inital switchover to upstart event based init from my current BSD style init on my distro. Just trying to get the initial sysinit script to run before moving on to the next script, however I get the following error while trying to debug this: "init: failed to spawn sysinit main process: unable to set oom adjustment: No such file or [17:43] directory". I don't know what file or directory its trying to access. Any ideas before I dig into the mailing list archives? [17:44] /proc/1/oom_adj probably. Do you have a recent enough version of Linux? [17:45] Oh, wait. It probably doesn’t have /proc yet at that point. Why the error about file or directory then... I’m out of ideas. [18:36] :) [18:46] https://bugs.launchpad.net/upstart/+bug/259801/+viewstatus <-- looks like its a bug of some sort :( [18:48] You could just comment out the OOM adjustment code as a temporary workaround. [18:51] Trying that now [18:58] Keybuk: greetings [18:58] hullo! [18:59] Ah, Portland [18:59] heh [18:59] I head for the airport in 30 minutes [18:59] be in at 9pm [18:59] Hi Keybuk [18:59] keybuk: http://gitweb.heh.fi/?p=ion/idlemeter.git;a=blob;f=README;hb=HEAD :-) [19:00] Keybuk: that reminds me, we need to redo libnih :D [19:01] * sadmac2 spent much of the week trying to hammer locking primitives into it to make a new version of that list iteration race go away [19:02] yeah, libnih has some issues [19:02] Bill must be running some kind of Murphy's Law CPU. He always finds these first [19:02] I quite like the idea of moving to using ion_'s refcounting ideas in it too [19:02] Keybuk: I've been reading some papers on region-based alloc. I'll have to point them out to you [19:03] we could just write Upstart in Python ;) [19:03] Keybuk: actually, I'll be bringing that up whenever we get to our little design session :) [19:03] have you found anywhere to stay yet? [19:04] Keybuk: nope [19:04] bill's hotel room might have a bath tub though [19:04] I forwarded your mail on to one of the organiser's, so hopefully that'll help [19:05] why sleep anyway? [19:05] http://heh.fi/tmp/idlemeter [19:06] Also, “how many hours did my visit to the city take?” % ruby -I~/src/own/idlemeter/ruby -ridlemeter -e 'puts IdleMeter::Analyze.intervals.find_all {|i| i.mode == :idle and i.length >= 3600 }.last' [19:06] idle 2008-09-16 13:19:02 +0300 – 2008-09-16 16:49:01 +0300 ( 3.5 hours)