setup/start/systemd/box.service

[Unit]
Description=Cloudron Admin
OnFailure=crashnotifier@%n.service
After=mysql.service nginx.service
; As cloudron-resize-fs is a one-shot, the Wants= automatically ensures that the service *finishes*
Wants=cloudron-resize-fs.service

[Install]
WantedBy=multi-user.target

[Service]
Type=idle
WorkingDirectory=/home/yellowtent/box
Restart=always
ExecStart=/home/yellowtent/box/box.js
; we run commands like df which will parse properly only with correct locale
Environment="HOME=/home/yellowtent" "USER=yellowtent" "DEBUG=box:*,connect-lastmile,-box:ldap" "BOX_ENV=cloudron" "NODE_ENV=production" "LC_ALL=C"
; kill apptask processes as well
KillMode=control-group
; Do not kill this process on OOM. Children inherit this score. Do not set it to -1000 so that MemoryMax can keep working
OOMScoreAdjust=-999
User=yellowtent
Group=yellowtent
; OOM killer is invoked in this unit beyond this. The start script replaces this with MemoryLimit for Ubuntu 16
MemoryMax=400M
TimeoutStopSec=5s
StartLimitInterval=1
StartLimitBurst=60
Move from supervisor to systemd This removes logrotate as well since we use systemd logging 2015-09-07 11:18:44 -07:00			`[Unit]`
			`Description=Cloudron Admin`
			`OnFailure=crashnotifier@%n.service`
box.service: start after nginx 2017-01-25 11:28:27 -08:00			`After=mysql.service nginx.service`
rename the service to cloudron-resize-fs 2016-12-30 11:27:00 -08:00			`; As cloudron-resize-fs is a one-shot, the Wants= automatically ensures that the service finishes`
			`Wants=cloudron-resize-fs.service`
Move from supervisor to systemd This removes logrotate as well since we use systemd logging 2015-09-07 11:18:44 -07:00
cloudron.target is not needed 2020-08-01 20:00:20 -07:00			`[Install]`
			`WantedBy=multi-user.target`

Move from supervisor to systemd This removes logrotate as well since we use systemd logging 2015-09-07 11:18:44 -07:00			`[Service]`
			`Type=idle`
			`WorkingDirectory=/home/yellowtent/box`
			`Restart=always`
hardcode logging of box code to box.log 2020-08-04 09:34:03 -07:00			`ExecStart=/home/yellowtent/box/box.js`
Fix issue where df output is not parsed correctly LANG is the default locale i.e when LC_* are not specificall LC_ALL will override them all https://forum.cloudron.io/topic/4681/going-to-system-info-triggers-assertion-error 2021-03-17 11:10:47 -07:00			`; we run commands like df which will parse properly only with correct locale`
			`Environment="HOME=/home/yellowtent" "USER=yellowtent" "DEBUG=box:*,connect-lastmile,-box:ldap" "BOX_ENV=cloudron" "NODE_ENV=production" "LC_ALL=C"`
Kill child processes On Unix, child processes are not killed when parent dies. Each process is part of a process group (pgid). When pgid == pid, it is the process group leader. node creates child processes with the parent as the group leader (detached = false). You can send a signal to entire group using kill(-pgid), as in, negative value in argument. Systemd can be made to do this by setting the KillMode=control-group. Unrelated: Process groups reside inside session groups. Each session group has a controlling terminal. Only one process in the session group has access to the terminal. Process group is basically like a bash pipeline. A session group is the entire login session with only one process having terminal access at a time. Fixes #543 2016-01-21 17:40:41 -08:00			`; kill apptask processes as well`
			`KillMode=control-group`
Set OOMScoreAdjust to stop box code from being killed OOMScoreAdjust can be set between -1000 and +1000. This value is inherited and systemd has no easy way to control this for children (box code also runs as non-root, so it cannot easily set it for the children using /proc/<pid>/oom_score_adj. When set to -1000 and the process reaches the MemoryMax, it seems the kernel does not kill any process in the cgroup and it spins up in high memory. In fact, 'systemctl status <service>' stops displaying child process (but ps does), not sure what is happenning. Keeping it -999 means that if a child process consumed a lot of memory, the kernel will kill something in the group. If the main box itself is killed, systemd will kill it at all because of KillMode=control-group. Keeping it -999 also saves box service group being killed relative to other docker processes (apps and addons). Fixes #605 2019-01-06 19:16:53 -08:00			`; Do not kill this process on OOM. Children inherit this score. Do not set it to -1000 so that MemoryMax can keep working`
			`OOMScoreAdjust=-999`
Move from supervisor to systemd This removes logrotate as well since we use systemd logging 2015-09-07 11:18:44 -07:00			`User=yellowtent`
			`Group=yellowtent`
Ubuntu 16 needs MemoryLimit systemd[1]: [/etc/systemd/system/box.service:25] Unknown lvalue 'MemoryMax' in section 'Service' 2019-01-17 09:20:31 -08:00			`; OOM killer is invoked in this unit beyond this. The start script replaces this with MemoryLimit for Ubuntu 16`
Set OOMScoreAdjust to stop box code from being killed OOMScoreAdjust can be set between -1000 and +1000. This value is inherited and systemd has no easy way to control this for children (box code also runs as non-root, so it cannot easily set it for the children using /proc/<pid>/oom_score_adj. When set to -1000 and the process reaches the MemoryMax, it seems the kernel does not kill any process in the cgroup and it spins up in high memory. In fact, 'systemctl status <service>' stops displaying child process (but ps does), not sure what is happenning. Keeping it -999 means that if a child process consumed a lot of memory, the kernel will kill something in the group. If the main box itself is killed, systemd will kill it at all because of KillMode=control-group. Keeping it -999 also saves box service group being killed relative to other docker processes (apps and addons). Fixes #605 2019-01-06 19:16:53 -08:00			`MemoryMax=400M`
systemd: just wait for 5 seconds for box to die 2015-09-07 20:58:14 -07:00			`TimeoutStopSec=5s`
Try alternative configuration for systemd restart rate limit 2015-12-12 13:14:50 -08:00			`StartLimitInterval=1`
			`StartLimitBurst=60`
Move from supervisor to systemd This removes logrotate as well since we use systemd logging 2015-09-07 11:18:44 -07:00