Failed to get unit file state for ghost_<url>.service

[aileen]

I uninstalled all existing installations and started new again with the following settings:

• mysqyl user: yes
• system user: no
• nginx: yes
• ssl: no

Trying to start Ghost within the setup process prompted first in the warning to run ghost setup systemd. After doing so, trying to start Ghost again results in this error:

root@dev-aileen-01:~/cli1# ghost setup systemd
? Ghost-CLI would like to generate a systemd service file file. Yes, write config file
Running sudo command: ln -sf /root/cli1/system/files/ghost_128-199-61-218.service /lib/systemd/system/ghost_128-199-61-218.service
✔ Setting up systemd
root@dev-aileen-01:~/cli1# ghost start
✔ Validating config
Running sudo command: systemctl start ghost_128-199-61-218
Warning: ghost_128-199-61-218.service changed on disk. Run 'systemctl daemon-reload' to reload units.
✔ Starting Ghost
An error occurred.
Message: 'Failed to get unit file state for ghost_128-199-61-218.service: No such file or directory
'


Debug Information:
    Node Version: v6.11.0
    Ghost-CLI Version: 1.0.0-beta.5
    Environment: production
    Command: 'ghost start'

Additional log info available in: /root/.ghost/logs/ghost-cli-debug-2017-07-06T08_44_21_865Z.log

[kirrg001]

What i did to get rid of the error:

rm /etc/systemd/system/ghost_localhost.service [not sure this step was needed]
sudo systemctl daemon-reload
ghost start

[aileen]

It worked for me without the first command

[kirrg001]

I am running into this again:

'Failed to get unit file state for ghost_kate-ghost-test-acburdine-me.service: No such file or directory

sudo systemctl daemon-reload doesn't help this time.

systemctrl status ... looks great.

   Loaded: loaded (/var/www/ghost/system/files/ghost_kate-ghost-test-acburdine-me.service; bad; vendor preset: enabled)
   Active: active (running) since Thu 2017-07-06 20:51:19 UTC; 1s ago
     Docs: https://docs.ghost.org
 Main PID: 20218 (ghost)
    Tasks: 17
   Memory: 45.5M
      CPU: 1.541s
   CGroup: /system.slice/ghost_kate-ghost-test-acburdine-me.service
           ├─20218 ghost
           ├─20229 /bin/sh -c npm root -g
           └─20230 npm

Jul 06 20:51:19 dev-kate-01 systemd[1]: Started Ghost systemd service for blog: kate-ghost-test-acburdine-me.

The error is thrown directly from the CLI and that's why Ghost doesn't really start.

Stack: Error: Failed to get unit file state for ghost_kate-ghost-test-acburdine-me.service: No such file or directory

    at module.exports.sync (/usr/local/nvm/versions/node/v6.10.3/lib/node_modules/ghost-cli/node_modules/execa/index.js:298:26)
    at handleShell (/usr/local/nvm/versions/node/v6.10.3/lib/node_modules/ghost-cli/node_modules/execa/index.js:100:9)
    at Function.module.exports.shellSync (/usr/local/nvm/versions/node/v6.10.3/lib/node_modules/ghost-cli/node_modules/execa/index.js:307:43)
    at SystemdProcessManager.isEnabled (/usr/local/nvm/versions/node/v6.10.3/lib/node_modules/ghost-cli/extensions/systemd/systemd.js:36:19)

From here: https://github.com/TryGhost/Ghost-CLI/blob/master/extensions/systemd/systemd.js#L36

But the file exists:

root@dev-kate-01:/var/www/ghost# ll /lib/systemd/system/ghost_kate-ghost-test-acburdine-me.service
lrwxrwxrwx 1 root root 70 Jul  6 20:37 /lib/systemd/system/ghost_kate-ghost-test-acburdine-me.service -> /var/www/ghost/system/files/ghost_kate-ghost-test-acburdine-me.service

root@dev-kate-01:/var/www/ghost# ll /var/www/ghost/system/files/ghost_kate-ghost-test-acburdine-me.service
-rw-r--r-- 1 root root 375 Jul  6 20:37 /var/www/ghost/system/files/ghost_kate-ghost-test-acburdine-me.service

SEE
https://serverfault.com/questions/728543/unable-to-query-services-with-systemctl-is-enabled

I've changed https://github.com/TryGhost/Ghost-CLI/blob/master/extensions/systemd/systemd.js#L36 to only check if the file exists, then it worked. But this obviously not a replacement for the is-enabled feature?

[acburdine]

I figured out what this was. Will submit a fix soon.