browser-sockpuppet-chrome is unusable #3682
nzbtuxnews
started this conversation in
General
Replies: 4 comments 8 replies
-
|
Hmm sometning is different in your setup remove the |
Beta Was this translation helpful? Give feedback.
4 replies
-
|
Hello dgtlmoon, removing CONCURRENT=4 and moving the containers to their own network seem to have helped. I could reconfigure and refresh a bunch of feeds without immediately crashing. However, after a while, things went south: |
Beta Was this translation helpful? Give feedback.
-
|
What CPU is it running on? |
Beta Was this translation helpful? Give feedback.
-
|
Its a Rocky Linux VPS running on AMD Opteron 4334 CPU's (older CPU's). Just now the changedetection seems to have issues as well: |
Beta Was this translation helpful? Give feedback.
-
this is NOT an error its just the browser client disconnecting from the realtime update. |
Beta Was this translation helpful? Give feedback.
-
|
Can confirm I'm hitting the same behaviour with the exact same timeout exceeded error message. All default, FETCH_WORKERS at 1 and MAX_CONCURRENT_CHROME_PROCESSES at 2. Debian 12 VM on Proxmox, 4 vCores, 8GB RAM, i5 8400T. |
Beta Was this translation helpful? Give feedback.
0 replies
-
|
For me, these arguments (max concurrent processes, concurrent=X, etc) have no influence on the issues I described above. As an example, without any exotic parameters and a very bland docker compose, I just tried to force-check 18 feeds using fetch method = playwright-chromium and got for ALL of them
I carefully picked the relevant container logs during that timeline and I see some errors: |
Beta Was this translation helpful? Give feedback.
1 reply
-
|
Same here, the only thing that temporarily fixes it for me is restarting the browser-sockpuppet-chrome container. It will then work for some time and then start failing again for every single check. |
Beta Was this translation helpful? Give feedback.
-
|
Same here. Clearly something seems to "saturate" the container so it becomes dysfunctional. Can you post your docker compose to compare with mine? |
Beta Was this translation helpful? Give feedback.
3 replies
-
|
Sure, here's mine: |
Beta Was this translation helpful? Give feedback.
-
|
Right off the bat I see some minor differences but IMO unless the host OS lacks some memory and somehow starts killing processes, or something else does the same... they should not make much of a difference: However, for me, the changedetection container is tied to TWO networks (one called "changetectionnetwork" so it can "talk" to the sockpuppet container directly without going through traefik (as rightfully recommended above by dgtlmoon), and another network called "frontend" because I want the WebUI to be going through traefik for FQDN instead of blablabla:5000... Unless MY network configuration is problematic, I cant see a clear root cause for our issue.. |
Beta Was this translation helpful? Give feedback.
-
|
Well, I just re-deployed the stack with TRACE log level to see if anything comes up and nothing ;) De-deploying also resolves all issues. Really strange. I have a feeling its related to resource allocation somehow. Until @dgtlmoon weigh in, I'll keep searching.. |
Beta Was this translation helpful? Give feedback.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
-
Running changedetection with browser-sockpuppet-chrome with 40+ feeds I get the following error:
Docker compose:
Docker is running on a Rocky Linux VPS with 12GB RAM and 2 Cores. Hardware is not maxed out (as indicated per Proxmox's WebUI)
Here's a snippet of the sockpuppet container logs.
_changedetection-browser-sockpuppet-chrome-1_logs(1).txt
Beta Was this translation helpful? Give feedback.
All reactions