Closed spirillen closed 7 months ago
A little undocumented note:
It seems like the inbox stops working when I start deletes cat- and bot spams from the global timeline, I don't know if these are related... but the time for deleting vs the inbox stops working in almost impeccable
the issue persists :sob:
After restart of both redis & misskey...
Today I have restarted MK 7 times to get the msg out of the inbox...
Is there really no glue for what can be wrong here, or where to look for some error msg to help solve this issue?
Logs from the minutes up to the 8th failure within 30 minutes
Oct 26 13:07:47 ns3109441 "matrix.rocks"[752925]: INFO 1 [drive register] file with same hash is found: 9laripk3f2
Oct 26 13:07:47 ns3109441 "matrix.rocks"[752925]: DONE 1 [drive downloader] Got: 9laripk3f2
Oct 26 13:07:48 ns3109441 "matrix.rocks"[752925]: Error: Vapid public key should be 65 bytes long when decoded.
Oct 26 13:07:48 ns3109441 "matrix.rocks"[752925]: at Object.validatePublicKey (/home/misskey/misskey/node_modules/.pnpm/web-push@3.6.6/node_modules/web-push/src/vapid-helper.js:111:11)
Oct 26 13:07:48 ns3109441 "matrix.rocks"[752925]: at WebPushLib.setVapidDetails (/home/misskey/misskey/node_modules/.pnpm/web-push@3.6.6/node_modules/web-push/src/web-push-lib.js:60:17)
Oct 26 13:07:48 ns3109441 "matrix.rocks"[752925]: at PushNotificationService.pushNotification (file:///home/misskey/misskey/packages/backend/built/core/PushNotificationService.js:69:14)
Oct 26 13:07:48 ns3109441 "matrix.rocks"[752925]: at process.processTicksAndRejections (node:internal/process/task_queues:95:5)
Oct 26 13:08:15 ns3109441 "matrix.rocks"[752925]: WARN 1 [queue inbox] stalled id=3211855
Oct 26 13:10:00 ns3109441 "matrix.rocks"[752925]: INFO 1 [queue check-expired-mutings] Checking expired mutings...
Oct 26 13:10:00 ns3109441 "matrix.rocks"[752925]: DONE 1 [queue check-expired-mutings] All expired mutings checked.
Seems to have solved this issue, despite the missing pointers from the log, and this was by accident as I had to make changes to the DB "tuning" for other websites.
So for other, I solved this by half-fen my "max_wal_size" in /etc/postgresql/15/main/postgresql.conf
This should still be in the log that the DB have gone, and connection is lost|stalled
💡 Summary
read https://github.com/misskey-dev/misskey/issues/12025
🥰 Expected Behavior
msg in inbox are delivered
🤬 Actual Behavior
For about every 1 day, 20 - 30 hours, my inbox stops working for no reasons and no error msg. This have been going on for about a week
https://github.com/misskey-dev/misskey/assets/44526987/30effa48-ce4e-43dc-81a0-6b3f29971266
https://github.com/misskey-dev/misskey/assets/44526987/f66d41d1-1b49-42b5-b850-23984ef5bd97
https://github.com/misskey-dev/misskey/assets/44526987/09cfbf08-7ab1-49cb-8061-ef46afc69d35
📝 Steps to Reproduce
📌 Environment
2023.10.1
💻 Frontend
5.15.0-86-generic
🛰 Backend (for server admin)
5.15.0-86-generic