This is a small intoduction into common chromebot admin tasks.
To restart chromebot run:
systemctl --user restart chromebot-irc
The dashboard can be restarted with:
systemctl --user restart chromebot-irc-dashboard
If you want to retrieve the status of a service like chromebot-irc, replace “restart” with “status”.
Usually only jobs that ran for more than a few hours should be killed. Get a list of browser processes using:
ps -f -C google-chrome-stable
Then check the column stime (start time) and verify the process has been running a long time. Use the PID (not PPID) column to figure out the process ID, which you can then feed into:
kill 12345
Replace 12345 with the actual number. If you now run the first command the process should be gone and – if any – the next job will run.
Crashing jobs leave their temporary data behind in the directory crocoite-data/temp, thus it is necessary to clean those up from time to time.
Most jobs that crashed Google Chrome can be uploaded, because they usually fail during screenshot capturing. These jobs can be found and moved using:
cd crocoite-data/temp/ find . -mtime +0 -size +2M -ls
Examine list of files, check the WARCs look alright and they finished with status 2, then:
find . -mtime +0 -size +2M | parallel 'mv -vi {} ../finished/'