=== Md is now known as Guest1090 === Md_ is now known as Md [13:06] hi - I need help with controlling uwsgi which is installed with pip - how should the /etc/init/uwsgi.conf look like? http://pastebin.com/vMK3Wyrb - i have this now [13:08] agend: http://uwsgi-docs.readthedocs.org/en/latest/Upstart.html is fairly good and comprehensive guide. [13:08] agend: looks ok. [13:09] my problem is that it doesn't want to restart [13:09] sudo service uwsgi stop stop: Unknown instance: [13:10] agend: see the doc I linked to, and use die on term option. [13:10] before sudo service uwsgi start gave me that: uwsgi start/running, process 16272 [13:11] yeah I use die-on-term [13:11] agend: and ideally use emperor option as well. [13:11] not sure i need emperor [13:12] agend: check with ps, that the pid of uwsgi actually matches what status uwsgi gives you. [13:13] and when i do: pgrep uwsgi - i can't see any process with this nr [13:14] it does not match [13:15] in that case they way you are starting uwsgi it either forks or demonizes and expect is then wrong. [13:15] uwsgi start/running, process 16534 [13:15] vagrant@one:~$ pgrep uwsgi [13:15] 16536 [13:15] 16541 [13:15] 16542 [13:15] 16543 [13:15] 16544 [13:15] start with minimal sample, I'm not sure what all of your yaml options cause uwsgi to do. [13:15] looks like you want to add: expect daemon [13:15] to the upstart job. [13:16] it's best to simply run uwsgi in the foreground. and then logs will be collected in /var/log/upstart/uwsgi [13:16] agend: also at this point you'd have to manually kill that uwsgi. [13:18] xnox: man - you are awsome - works now :) [13:28] xnox: sorry - i meant to say awesome - i just checked that awsome might mean something opposite, so yeah u r great thanks a lot [19:00] is there a way to make upstart do some checks if I do a status $service? [21:56] erkules: why would you need that? [21:56] erkules: generally no. status purely reports status of the pid from upstart's point of view. If it's still running & present than it's all good. [21:56] erkules: maybe you should use monitoring e.g. nagios or check_mk ?