=== JanC_ is now known as JanC [12:17] in an upstart job, if i use `respawn \n respawn limit 10 5 \n post-stop exec sleep 60`, does that mean that my respawn limit will never be reached because it should never respawn more than once every 60 seconds? [12:49] it looks like it's true that if you sleep between respawns you have to also increase the interval to encompass the sleep duration [12:49] [14:01] timvisher: the respawn limit tells the max times it can restart for a X time [14:01] So if it's broken or something and keeps restarting, it won't do it for ever [14:01] If I understood your question [14:47] supergonkas: right. so if you're sleeping between each invocation (say for 1 second) and your respawn limit is 10 5, then you'll by definition never hit that limit [14:47] because you can't respawn 10 times in 5 seconds if you're sleeping for a second every time you respawn [14:47] Yes [14:48] Mm well depends if the service crashes before the sleep is executed, I think [14:48] so you basically need to incorporate the sleep time and assume that you're failing immediately or some how otherwise determine the acceptable failure overhead [14:48] eh? sleep is being executed in post-stop [14:49] so no matter how the process dies it'll sleep, unless you're using `stop` explicitly iiuc [14:49] is that accurate? [14:49] Ignore me, just re-read your original message [14:49] lol. no problem :) [14:49] thanks for the help [14:50] Yes I would say the sleep is always executed [14:59] Your best bet is just increase the interval of the respawn limit, that is within the same frequency as the original [14:59] Or move the sleep [14:59] *I think your best bet.. [15:00] supergonkas: yep. that's what i did. i have a sleep of 60 seconds and a respawn limit of 30 ~1800 [15:57] timvisher: Check the cookbook. It has the flow diagram. [16:00] AnrDaemon: where? [16:01] oh nevermind i found it [16:08] google "upstart cookbook", first result ;) [16:15] not the cookbook itself, the flow diagram :) [16:24] The two biggest pics in there.