True, but systemd here at work communicates to and logs to a central logging system off-server and it is useful to see all the times it goes down, starts eating too much cpu, etc… etc… 
That is what my systemd service script does. ^.^
OvermindDL1
1
ijunaidfarooq
Just because there is a SystemD job written in the question , So anyone might help me as well.
I have this Service Unit.
[Unit]
Description=Evercam Media
After=network.target
[Service]
User=root
Group=root
WorkingDirectory=/
ExecStart=/opt/evercam_media/bin/evercam_media start
ExecStop=/opt/evercam_media/bin/evercam_media stop
#Restart=always
[Install]
WantedBy=multi-user.target
And it only giving me
this error always
Aug 23 08:28:30 Ubuntu-1604-xenial-64-minimal systemd[1]: Started Evercam Media.
Aug 23 08:28:31 Ubuntu-1604-xenial-64-minimal evercam_media[30064]: Starting up
Aug 23 08:28:31 Ubuntu-1604-xenial-64-minimal evercam_media[30009]: Node 'evercam_media@Ubuntu-1604-xenial-64-minimal' not responding to pings.
Aug 23 08:28:31 Ubuntu-1604-xenial-64-minimal systemd[1]: evercam_server.service: Control process exited, code=exited status=1
Aug 23 08:28:32 Ubuntu-1604-xenial-64-minimal systemd[1]: evercam_server.service: Unit entered failed state.
Aug 23 08:28:32 Ubuntu-1604-xenial-64-minimal systemd[1]: evercam_server.service: Failed with result 'exit-code'.
Can you point me the issue in this?
0
OvermindDL1
That indicates to me that the evercam_media start command is failing. Have you tried launching it manually? Does it require to be started after something else other than After=network.target like the USB system or something?
EDIT: Oh wait, this is a BEAM VM?
Since you do not have a Type specified then it defaults to Simple as I recall, meaning it expects the program to persist, not fork. You need to use Type=forking as in my example. You really should copy and use my example in that case.
0
ijunaidfarooq
Actually , I converted the Upstart Job to SystemD Unit, So thats why I added that After=network.target thing..
Yes Its starts manually..
Okay I am going to try with Type=forking and with your example,
1
ijunaidfarooq
Okay your snippet worked and in this way
LimitNOFILE 1000000 1000000
[Unit]
Description=Evercam Media
After=network.target
[Service]
Type=forking
User=root
Group=root
WorkingDirectory=/opt/evercam_media
ExecStart=/opt/evercam_media/bin/evercam_media start
ExecStop=/opt/evercam_media/bin/evercam_media stop
Restart=always
RestartSec=5
Environment=LANG=en_US.UTF-8
SyslogIdentifier=evercam_media
[Install]
WantedBy=multi-user.target
I have few concerns which I want to share with you, In case you see few things wrong here or may you can help in any case..
We were on Ubuntu 14.04 and was using upstart job, which was like
description "evercam_media"
start on filesystem or runlevel [2345]
stop on runlevel [!2345]
limit nofile 1000000 1000000
respawn
chdir /
setuid {{user_name}}
setgid {{user_name}}
env HOME=/home/{{user_name}}
env LANG=en_US.UTF-8
env LANGUAGE=en_US:en
env LS_ALL=en_US.UTF-8
env ERL_MAX_PORTS=10240
env ERL_MAX_ETS_TABLES=5000
{% for key, value in env_vars.items() %}
env {{key.upper()}}={{value}}
{% endfor %}
exec watch -n1 '/usr/local/bin/run_evercam_media.sh'
post-stop exec sudo pkill beam
The former developer who created the upstart job, He used watch with seconds values as 1 , just to watch the script every one second. and in SH script, he was doing something
if ! (ps aux | grep evercam_media/bin/evercam_media | grep -v grep > /dev/null); then
/opt/evercam_media/bin/evercam_media start
fi
for example, the evercam media have stopped due to some reasons it will start it again. and watching it every one second.
What if the above service unit get failed due to some reason? Will it restart it self? I have seen Restart=on-failure thing as well, do you think it will help?
Also big question: ENV variables, I dont see anywhere you pointed about ENV variables?
0
OvermindDL1
That looks horrifying! That is quite an extra load on the server for no reason… He’s spawning an entire BEAM VM instance every second just to ping it?! o.O
If you use Restart=always as in my script then it will always restart, including on failure. Only way it will not be restarted is it you systemctl stop ... it, and even then it will be restarted on next reboot unless you systemctl disable ... it too. ^.^
I had two in my script above:
Environment=PORT=3000
Environment=LANG=en_US.UTF-8
You can of course define your own, or you can import an external file too (see the systemd service file spec).
1
ijunaidfarooq
That looks horrifying! That is quite an extra load on the server for no reason… He’s spawning an entire BEAM VM instance every second just to ping it?! o.O
Nope, He is not doing that, new BEAM VM instance only get initiated if evercam_media process is not found through sh script.
Okay thanks for your help.
1
OvermindDL1
happysalada
I can start the application without problem, but when I try to launch it with systemd it keeps failing and restarting for some reason.
Here are the logs
un 08 10:00:15 union-staging systemd[1]: Starting Union Servers...
-- Subject: Unit union.service has begun start-up
-- Defined-By: systemd
-- Support: http://lists.freedesktop.org/mailman/listinfo/systemd-devel
--
-- Unit union.service has begun starting up.
Jun 08 10:00:16 union-staging union[8741]: ==> Generated sys.config in /test/union/var
Jun 08 10:00:19 union-staging union[8998]: ==> Generated sys.config in /test/union/var
Jun 08 10:00:19 union-staging union[9530]: Starting up
Jun 08 10:00:20 union-staging systemd[1]: union.service: control process exited, code=exited status=1
Jun 08 10:00:20 union-staging union[8998]: Node union@127.0.0.1 is not running!
Jun 08 10:00:20 union-staging systemd[1]: Unit union.service entered failed state.
Jun 08 10:00:20 union-staging systemd[1]: union.service failed.
Jun 08 10:00:25 union-staging systemd[1]: union.service holdoff time over, scheduling restart.
Jun 08 10:00:25 union-staging systemd[1]: Started Union Servers.
-- Subject: Unit union.service has finished start-up
-- Defined-By: systemd
-- Support: http://lists.freedesktop.org/mailman/listinfo/systemd-devel
--
-- Unit union.service has finished starting up.
--
lines 1190-1211/1247 98%
-- The start-up result is done.
Jun 08 10:00:15 union-staging systemd[1]: Starting Union Servers...
-- Subject: Unit union.service has begun start-up
-- Defined-By: systemd
-- Support: http://lists.freedesktop.org/mailman/listinfo/systemd-devel
--
-- Unit union.service has begun starting up.
Jun 08 10:00:16 union-staging union[8741]: ==> Generated sys.config in /test/union/var
Jun 08 10:00:19 union-staging union[8998]: ==> Generated sys.config in /test/union/var
Jun 08 10:00:19 union-staging union[9530]: Starting up
Jun 08 10:00:20 union-staging systemd[1]: union.service: control process exited, code=exited status=1
Jun 08 10:00:20 union-staging union[8998]: Node union@127.0.0.1 is not running!
Jun 08 10:00:20 union-staging systemd[1]: Unit union.service entered failed state.
Jun 08 10:00:20 union-staging systemd[1]: union.service failed.
Jun 08 10:00:25 union-staging systemd[1]: union.service holdoff time over, scheduling restart.
Jun 08 10:00:25 union-staging systemd[1]: Started Union Servers.
-- Subject: Unit union.service has finished start-up
-- Defined-By: systemd
-- Support: http://lists.freedesktop.org/mailman/listinfo/systemd-devel
--
-- Unit union.service has finished starting up.
Here is my systemd config
[Unit]
Description=Union Servers
After=network-online.target
[Service]
Type=forked
User=root
Group=root
WorkingDirectory=/test/union
ExecStart=/test/union/bin/union start
ExecReload=/test/union/bin/union reload_config
ExecStop=/test/union/bin/union stop
PIDFile=/test/union/union.prod.pid
Restart=always
RestartSec=5
Environment=PORT=4000
Environment=LANG=en_US.UTF-8
SyslogIdentifier=union
[Install]
WantedBy=multi-user.target
The system variable I use is
PIDFILE="./union.prod.pid"
Here is my mix file
[
{:phoenix, "~> 1.3.0"},
{:phoenix_pubsub, "~> 1.0"},
{:phoenix_ecto, "~> 3.2"},
{:postgrex, ">= 0.0.0"},
{:phoenix_html, "~> 2.10"},
{:gettext, "~> 0.13.1"},
{:cowboy, "~> 1.0"},
{:ex_admin, github: "sublimecoder/ex_admin"},
{:timex, "~> 3.2", override: true},
{:coherence, "~> 0.5"},
{:cloudex, "~> 1.0"},
{:number, "~> 0.5.4"},
{:turbolinks, "~> 0.3.2"},
{:ecto_enum, "~> 1.1"},
{:jason, "~> 1.0"},
{:absinthe, "~> 1.4.2"},
{:absinthe_plug, "~> 1.4.0"},
# Instrumentation
{:observer_cli, "~> 1.3.1"},
{:exprof, "~> 0.2.1"},
{:eflame, "~> 1.0"},
{:sentry, "~> 6.2.0"},
{:prometheus_ex, "~> 1.0"},
{:prometheus_ecto, "~> 1.0"},
{:prometheus_phoenix, "~> 1.2"},
{:prometheus_plugs, "~> 1.0"},
{:prometheus_process_collector, "~> 1.1"},
# dev dependencies
{:phoenix_live_reload, "~> 1.0", only: :dev},
{:ex_doc, "~> 0.18", only: :dev, runtime: false},
{:credo, "~> 0.8", only: [:dev, :test], runtime: false},
{:dialyxir, "~> 0.5.0", only: [:dev], runtime: false},
# release
{:edeliver, "~> 1.5.0"},
{:distillery, "~> 1.5", runtime: false},
{:pid_file, "~> 0.1.0"},
{:conform, "~> 2.2"}
]
I’m not really sure what I’m doing wrong.
0
NobbZ
Use an absolute path, just to be sure.
Also you haven’t shown your configuration of the pid_file-application.
1