In the console/terminal for an app, run
ps -ux | grep server:
app@tree-jelly:~ 11:48 $ ps -ux | grep server app 1573 0.3 0.1 1128360 46804 ? Sl 11:46 0:00 node server.js app 1644 0.0 0.0 12944 932 pts/0 S+ 11:48 0:00 grep server
Then, back in Glitch add a file to the project named
myfile.html, and then run
ps -ux | grep server again:
app@tree-jelly:~ 11:48 $ ps -ux | grep server app 1573 0.2 0.1 1128616 47296 ? Sl 11:46 0:00 node server.js app 1646 0.0 0.0 12944 976 pts/0 S+ 11:50 0:00 grep server
Okay, it’s the same PID for
node server.js. Which is good.
Now, back in Glitch add a file to the project named
myfile.js, and then run
ps -ux | grep server a third time:
app@tree-jelly:~ 11:50 $ ps -ux | grep server app 1667 6.3 0.1 1062584 45944 ? Sl 11:51 0:00 node server.js app 1705 0.0 0.0 12944 936 pts/0 S+ 11:51 0:00 grep server
You will see that the PID for
node server.js has changed. Which is bad.
This is a problem for me because in the
node server.js process is where I have a custom chokidar watcher running, and it gets restarted when a
.js file changes (or is added) before it gets a chance to kick off its build. I’m hoping I don’t have to implement any hacks to work around this, because I like having my watcher be part of
server.js so that it gets integrated seamlessly with all the other auto stuff that Glitch does.
Is there a solution to my problem though, whether hack-y or not?
One peculiar observation is that I actually only observe this process restarting behavior for a brand new Glitch. If I walk away for >5 mins and come back after the app container has had a chance to shut down, then start it back up it again, then this process restarting problem magically goes away?
Anyone have any insight into this behavior I’m observing, or more importantly have a fix?
I have tested this many times with new vanilla and import projects and it is consistently a reproducible problem.