| lordievader | Good morning | 06:16 |
|---|---|---|
| === Osares102 is now known as Osares10 | ||
| === cpaelzer_ is now known as cpaelzer | ||
| === genii-core is now known as genii | ||
| === vlm_ is now known as vlm | ||
| === toabctl_ is now known as toabctl | ||
| === akaWolf1 is now known as akaWolf | ||
| === paulw2 is now known as PaulW2U | ||
| === ^wuseman is now known as Guest3499 | ||
| === lotuspsychje_ is now known as lotuspsychje | ||
| === Beret- is now known as Beret | ||
| === bahamat_ is now known as bahamat | ||
| === kinghat4 is now known as kinghat | ||
| === paulw2 is now known as PaulW2U | ||
| === dbungert1 is now known as dbungert | ||
| === JanC_ is now known as JanC | ||
| StoneMonarch | I am on Ubuntu Server 20.04 I have some docker-compose files that run on start up through systemd. When running in detached (`-d` or `--detached`) mode docker-compose immediately takes down the containers after they start, however this does not occur when not detached. Is it okay to run these service files without `-d`. I have not experienced any issues running multiple service files in this manner, but is there | 19:23 |
| StoneMonarch | problems that can occur? Thanks. | 19:23 |
| tar_xvf | on my vm ubuntu server I currently use a @reboot crontab line to run my server automatically when i reboot the machine. Personally I don't have any experience starting custom systemd jobs, are there any reasons I should consider moving to systemd, such as, idk, something like being less taxing on the CPU, or is either way fine? | 19:24 |
| sdeziel | StoneMonarch: dunno docker but with systemd units, it's recommended to not detach | 19:24 |
| sdeziel | tar_xvf: reliability maybe? systemd can revive a dead process if you ask it to | 19:24 |
| StoneMonarch | @sdeziel much appreciated. | 19:25 |
| tar_xvf | good point sdeziel , thanks | 19:25 |
| sarnold | tar_xvf: getting output from the command in the journal etc might be convenient, rather than only ever getting failures emailed to you | 19:29 |
| tar_xvf | i will look into that, thanks sarnold | 19:29 |
| sarnold | tar_xvf: you can also set ulimits or seccomp restrictions or use systemd's namespace shenanigans to hide portions of the filesystem | 19:29 |
| sarnold | tar_xvf: and the best reason I can think of is that if you want to restart the service for some reason, you'll get a consistent runtime environment to just systemctl stop the service, systemctl start the service, or systemctl restart, etc; to get the same execution environment a *second* time with cron would require adding a new cronjob entry to start the service, then go remove it again once it's | 19:31 |
| sarnold | running | 19:31 |
| tar_xvf | i like that reason. For me, usually i just restart the vm, but i have multiple services running on a machine so that would be good to restart one without disrupting the others too much. | 19:32 |
| tar_xvf | my last question between the two, is running a script at startup via cron or systemd generally more CPU intensive, for the same script? or is that negligible | 19:38 |
| sarnold | tar_xvf: it shouldn't matter | 19:40 |
| tar_xvf | ok, thank you very much sarnold. | 19:40 |
| teward | ummm... who's the apache2 expert? | 20:12 |
| teward | because I think 18.04 upgrade has a broken Apache2 problem - if prefork is enabled it still tries to enable mpm_event and breaks PHP and such | 20:13 |
| teward | and then triggers apt failures | 20:13 |
| teward | blah i fixed it by manually overwriting the postinst >.> | 20:16 |
| sdeziel | I've seen do-release-upgrade disabling mod_php and leaving at that but that was fixed months ago... | 20:17 |
| teward | sdeziel: it's possible then that was done, I think this is a 16.04 -> 18.04 upgrade bug because it's treating it as if it's a fresh install | 20:29 |
| teward | and enabling mpm_event even if mpm_prefork is enabled, leading to a hard upgrade failure | 20:29 |
| teward | i managed to *bypass* that for now, but... | 20:29 |
| teward | it's something that took down a server today and is on my radar of "infinitely annoying things" | 20:29 |
| === pizzaiolo is now known as pizza | ||
| === genii is now known as genii-core | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!