Personally I would wipe your pi (unless you have anything you want to save) and restart from scratch, or simply start on a clean separate SD card.
Looks like you had two discourse installs - hence maybe the second Redis process.
Personally I would wipe your pi (unless you have anything you want to save) and restart from scratch, or simply start on a clean separate SD card.
Looks like you had two discourse installs - hence maybe the second Redis process.
rm -rf *
? Or is that too destructive?
It kinda works now. It takes very long to the load the page, and gives timeouts because it takes too long.
E, [2024-11-08T14:57:11.960623 #2213] ERROR -- : worker=0 PID:2323 running too long (65s), sending USR2 to dump thread backtraces
E, [2024-11-08T14:57:18.033374 #2213] ERROR -- : worker=0 PID:2323 timeout (65s > 60s), killing
E, [2024-11-08T14:57:20.607509 #2213] ERROR -- : reaped #<Process::Status: pid 2323 SIGKILL (signal 9)> worker=0
E, [2024-11-08T14:57:20.645299 #2213] ERROR -- : worker=1 PID:2324 running too long (63s), sending USR2 to dump thread backtraces
E, [2024-11-08T14:57:20.645461 #2213] ERROR -- : worker=1 PID:2324 timeout (63s > 60s), killing
E, [2024-11-08T14:57:20.681291 #2213] ERROR -- : reaped #<Process::Status: pid 2324 SIGKILL (signal 9)> worker=1
I, [2024-11-08T14:57:23.676422 #2555] INFO -- : worker=0 ready
I, [2024-11-08T14:57:23.681733 #2556] INFO -- : worker=1 ready
Which model pi are you using? I wouldn’t attempt this on less than a pi 4.
It will definitely take some time.
Once the JavaScript is built then issue another browser call. It should then work.
It was already (I think It was after the embroider>webpack and the discourse-plugins>apply-patches (something like that)). I’m on a Pi 4.
Those workers are unicorns (back end). They look ready. You have to be sure the front end is built too.
And remember to connect on http://localhost:4200
How will I know when it has finished building?
You get something like:
[ember-cli] Build successful (41838ms) – Serving on http://localhost:4200/
[ember-cli]
[ember-cli] Slowest Nodes (totalTime >= 5%) | Total (avg)
[ember-cli] -+-
[ember-cli] @embroider/webpack (1) | 23127ms
[ember-cli] Babel: admin (1) | 4067ms
[ember-cli] Funnel (422) | 2826ms (6 ms)
(those times are on my mega-PC!)
Raspberry Pi 5 (8GB):
[ember-cli] Build successful (148604ms) – Serving on http://localhost:4200/
[ember-cli]
[ember-cli] Slowest Nodes (totalTime >= 5%) | Total (avg)
[ember-cli] -+-
[ember-cli] @embroider/webpack (1) | 83461ms
[ember-cli] @embroider/compat/app (1) | 13488ms
[ember-cli] Babel: discourse-plugins (25) | 9313ms (372 ms)
[ember-cli] Babel: admin (1) | 8695ms
[ember-cli] Funnel (407) | 8557ms (21 ms)
Looking at my pi 5 memory, though, it’s already using 4.27GB just with Discourse running and a single Chromium tab open!
So if you are running into heavy swap usage (which itself is a bad idea on an SD card) it will be very slow to get to this point. Check with command htop
in another terminal.
You might find running this on a pi with less than 8GB a challenge!
I think I got that as well.
I’ll run that and get back to you.
When it said this:
The last amount recorded was this:
Then it froze (an still is). Likely went over the limit. I closed the browser tab but it’s still frozen. Can’t even CTRL+C.
Yep, I’m afraid your little pi needs more memory. Once it is maxed out like that it will freeze.
One option is to add an NVMe drive and increase swap.
You can then increase swap safely.
I think the only other solution is to source an 8GB 5 for this task and use your 4 for something else.