=== Demonho is now known as X-warrior` | ||
=== MrSmurf is now known as xSmurf | ||
cwillu_at_work | does upstart do its own permissions checking? | 06:34 |
---|---|---|
cwillu_at_work | I think I've got /etc/dbus-1/system.d/Upstart.conf set to permit one more message through as a particular user to a particular service, but I'm getting a com.ubuntu.Upstart0_6.Error.PermissionDenied | 06:35 |
cwillu_at_work | (without the Upstart.conf change, I get a generic dbus error instead) | 06:36 |
cwillu_at_work | hmm, this appears to be something related to the (new?) session support | 06:45 |
cwillu_at_work | --no-session on the kernel command line makes it work right | 06:45 |
xnox | cwillu_at_work: interesting, what upstart version is that? | 10:30 |
slava_dp | hey everyone. in rhel6, what's the best way to switch uid and gid for a daemon? | 11:57 |
slava_dp | I'm writing an upstart job and setuid and setgid are not supported in rhel6 | 11:57 |
maveas | Hi guys. Does upstart emit events when a sysvinit script successfully starts? I am asking because I am about to write an upstart script for openerp and I need it to start on running postgres but I am quite new with init, both sysvinit and upstart.. | 13:05 |
maveas | And I don't think it does but I'm not sure | 13:06 |
maveas | I am thinking upstart is watching for a exit 0 but that doesn't really tell if the job is running ? | 13:07 |
maveas | Been reading a lot of documentation and from what I can understand this is how it works. Am I right? So as I need to control for a running postgres instance I could either do some pre-start code in my upstart script for OpenERP or I could create an upstart script for postgres so my OpenERP upstart script start when upstart emit that postgres is running. | 13:24 |
maveas | Never mind guys. Just realised that I can do a initctl emit postgres-running from the postgres sysvinit script. :) | 13:42 |
jodh | maveas: yes, that is the best method. However, you'll need to ensure you only emit that event when postgresl is actually *ready* (not necessarily the same time as when it *starts*) | 13:49 |
maveas | Yeah, I'm still looking into that. :) | 13:55 |
maveas | Thanks | 13:55 |
SpamapS | jodh: Hey, just wanted to give you a "nicely done" on the way upstart handles logging. | 16:17 |
SpamapS | jodh: I contrived a "worst case" scenario yesterday and I couldn't break upstart. | 16:17 |
jodh | SpamapS: cool - good to hear ;-) | 16:18 |
SpamapS | jodh: I do believe it will break down on a *REALLY* big box with *lots* of processes (like, 64 CPU w/ 10000 processes writing) | 16:18 |
SpamapS | jodh: but with 1CPU, 50 processes writing as hard as they can, init was using about 20% of the CPU, with the other 80% going to the kernel handling buffer shuffling | 16:19 |
SpamapS | jodh: I was wondering if we could somehow turn that into a test actually | 16:19 |
pseudonymous | anyone feel like translating "service disable -my service-" and "service stop -my service-" to upstart-speak ? In ubuntu I have two services, whoopsie and cron that NEED to be dead for the purposes of a benchmark, but quick as I kill them, they reappear | 18:30 |
SpamapS | pseudonymous: disable is rough, but 'service stop -my-service-' works fine | 18:33 |
SpamapS | pseudonymous: don't kill them, 'service cron stop' them | 18:33 |
SpamapS | pseudonymous: disable would be for reboots. For that you can 'echo manual >> /etc/init/whoopsie.override' | 18:34 |
pseudonymous | SpamapS: thanks so very much, I was really losing my temper with this stuff :D | 18:36 |
xnox | pseudonymous: see http://upstart.ubuntu.com/cookbook/ it actually tells one how to do that ;-) | 18:47 |
maveas | You guys behind Upstart. I just want to say thank you and to hug you.. As you may have noticed I started exploring Upstart today and I love it! :) | 18:56 |
maveas | It really makes a lot of sence and I hope it will become de facto on most systems. <3 :b | 18:57 |
maveas | Kudos | 18:58 |
pseudonymous | maveas: I couldn't disagree more </3 systemD forever and ever and ever and ever and ever.. :) And with that, good night ! But thanks still for the effort | 19:16 |
SpamapS | pseudonymous: I invite you to read the code of systemd and of upstart, and then re-assert what you are saying. | 19:20 |
SpamapS | pseudonymous: I know systemd has a ton of user-sugar .. but.. buyer beware. Its got a lot of easy, not much simple. | 19:21 |
maveas | There's a lot of fuss around Upstart but I think it's because people just don't understand it yet (and won't try to).. | 19:21 |
maveas | He'll come around some day ^ | 19:22 |
SpamapS | maveas: unfortunately, the rest of the Linux world got tired of waiting for Upstart to fix their problems, and so they invented systemd. | 19:31 |
SpamapS | maveas: it may take years for that to fail.. or systemd may become "the one true thing" | 19:31 |
xnox | SpamapS: i wonder if you failed the double negative there. </3 is "do not love" where as <3 "do love", thus pseudonymous is on the upstart side. | 19:52 |
cwillu | xnox, whatever's in precise | 20:34 |
cwillu | 1.5-0ubuntu5 | 20:35 |
SpamapS | xnox: I'm not sure I understand what you're saying. | 20:49 |
cwillu | SpamapS, </3 is a broken heart | 20:50 |
SpamapS | I think pseudonymous meant <3 .. since pseudonymous was disagreeing with maveas hoping that it will become de facto on most systems | 20:52 |
magicrobotmonkey | I've got a stop/killed <pid> job I can't get rid of. What should I do? | 21:42 |
SpamapS | magicrobotmonkey: https://bugs.launchpad.net/upstart/+bug/406397 | 22:55 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!