Home / ubuntu / Questions / 1264764
In Process
Marc
Asked: 2020-08-06 05:53:36 +0800 CST

Is this systemd service running or not?

  • 772

When we re-deploy an application we use systemd restart to restart the application. It would seem that either systemd restart or systemd status (or both) are unreliable.

  1. Sometimes systemd restart DOESN'T restart the application and doesn't throw an error. PID is the same as before and the old version of the app is still running. Is restart effective immediately?
  2. Sometimes systemd restart does kill the old process and start the new process but systemd status shows the service as dead. We see the process running as a child of systemd.

As an example of the 2nd problem: systemctl status is says the service inactive(dead) but it's definitely running (new pids 25548 and 25551)!

$systemctl status gateway.service
gateway.service - The API Gateway - Nginx
   Loaded: loaded (/etc/systemd/system/gateway.service; enabled; vendor preset: disabled)
   Active: inactive (dead) since Wed 2020-08-05 14:00:16 CEST; 1h 43min ago
  Process: 25545 ExecStart=/usr/sbin/nginx -c /prg/nginx/conf/gateway.conf (code=exited, status=0/SUCCESS)
  Process: 25542 ExecStartPre=/usr/sbin/nginx -t -c /prg/nginx/conf/gateway.conf (code=exited, status=0/SUCCESS)
  Process: 25540 ExecStartPre=/usr/bin/rm -f /daten/nginx.pid (code=exited, status=0/SUCCESS)
 Main PID: 25545 (code=exited, status=0/SUCCESS)
    Tasks: 2
   Memory: 130.0M
   CGroup: /system.slice/gateway.service
           ├─25548 nginx: master process /usr/sbin/nginx -c /prg/nginx/conf/gateway.conf
           └─25551 nginx: worker process

ps | grep nginx shows the 2 nginx processes:

nginx      25548              root    5u  IPv4 58124367      0t0  TCP *:8080 (LISTEN)
nginx      25551             nginx    5u  IPv4 58124367      0t0  TCP *:8080 (LISTEN)

pstree shows these 2 processes as children of systemd:

systemd─┬...
        ├─nginx───nginx

systemd Config

[Service]
Type=forking
PIDFile=/daten/nginx.pid

# Nginx will fail to start if /run/nginx.pid already exists but has the wrong
# SELinux context. This might happen when running `nginx -t` from the cmdline.
# https://bugzilla.redhat.com/show_bug.cgi?id=1268621
ExecStartPre=/usr/bin/rm -f /daten/nginx.pid
ExecStartPre=/usr/sbin/nginx -t -c /prg/nginx/conf/gateway.conf
ExecStart=/usr/sbin/nginx  -c /prg/nginx/conf/gateway.conf

KillSignal=QUIT
TimeoutStopSec=5
KillMode=process
PrivateTmp=true
systemd

0 Answers