[05:04] <ZBandit> Since upstart apparently is now default in CentOS 6 (RHEL6) how does one make the autologin with only on one TTY?
[05:58] <ZBandit> Never mind, figured it out on me own.
[13:35] <lucs> I'm trying to get something working with upstart and I currently have a situation where "initctl status foo" responds with "foo stop/killed, process 3574"
[13:35] <lucs> Yet, there is no process 3574, and I seem to be stuck there.
[13:37] <lucs> (I think the situation was caused by a mistaken "expect" stanza.)
[13:38] <jhunt> lucs: more than likely. Take a read of the newly updated section in the cookbook - http://upstart.ubuntu.com/cookbook/#expect.
[13:38] <jhunt> lucs: also note the golden rule - don't use "respawn" until you are sure you have correctly specified "expect".
[13:39] <lucs> jhunt: Yes, thanks, I'm there already (re. the cookbook).
[13:39] <lucs> Right now, I'm wondering how I can get back to a "clean" state.
[13:40] <jhunt> lucs: if initctl / status is hanging, control-C it. If it's not actually hanging, but is reporting incorrect information, I'm afraid you'll have to reboot to get Upstart to clear its knowledge of that job. What you *can* do if you're on a dev box is to copy the .conf file to a new name and "try again".
[13:41] <jhunt> lucs: we're going to be looking into this issue for this (Ubuntu) cycle. It's quite a hard problem to solve though unfortunately.
[13:41] <lucs> jhunt: Ah, I was kind of afraid I'd need to reboot, but there it is :)
[13:42] <lucs> Yeah, it's a dev machine, so I can experiment all I want.
[13:42] <jhunt> lucs: well, unfortunately, Upstart is powerless in this scenario right now as if the expect stanza is mis-specified there is no reasonable error recovery it can perform.
[13:42] <lucs> Oh, copying the .conf and retrying is a good idea, thanks.
[13:43] <lucs> jhunt: Ok, thanks for the info.
[13:43] <jhunt> lucs: np.
[15:11] <lucs> (Note to self: in "script" sections, don't declare environment variables with "env", unless you want to waste _a lot_ of time :-/ )
[15:14] <jhunt> lucs: ah! :) http://upstart.ubuntu.com/cookbook/#init-checkconf is your friend
[15:15] <jhunt> lucs: actually, that wouldn't have helped as of course 'env' *is* valid in shell, but not in the context you were using it in.
[15:32] <lucs> jhunt: By the way, all my problems went away when I got rid of the faulty 'env's.
[15:33] <jhunt> lucs: I'd be interested to see your .conf. Certainly sounds like a script process was aborting but Upstart was attempting to "follow forks" still.
[15:35] <lucs> jhunt: I have many failing variations, basically all with "script \n env FOO=some_command args... \n exec $FOO \n end script".
[15:52] <lucs> (sorry, was talking to a cow-orker).
[15:53] <lucs> So FOO was empty, and a plain 'exec' was being done (I suppose).
[16:02] <lucs> Time for a reboot.
[16:02] <lucs> jhunt: Thanks again.
[16:03] <jhunt> lucs: thanks for the info. I'll ponder this for future cookbook updates.
[16:09] <lucs> upstart++
[16:09] <lucs> jhunt: Nice cookbook too.
[16:10] <jhunt> lucs: still lots to add when time allows :)
[16:10] <lucs> :)