this post was submitted on 29 Nov 2023
6 points (87.5% liked)

Lemmy Support

4651 readers
4 users here now

Support / questions about Lemmy.

Matrix Space: #lemmy-space

founded 5 years ago
MODERATORS
 

So, I run my own instance (dandroid.app), and as of about 10 days ago, all outgoing federation stopped working. Incoming federation still seems to work fine. I'm not sure exactly what changed. It was around the time that I set up pihole on my network, but I have since shut down pihole and federation has not resumed. I have tried restarting all the services with docker-compose down and then up. I also tried restarting my server completely.

What debugging steps can I take to figure out where the failure is occuring?

top 6 comments
sorted by: hot top controversial new old
[–] seahorse@midwest.social 3 points 11 months ago (1 children)

Try select * from instance where domain = 'popular.instance'; and see the last time it was updated via the updated column. I've been having to update that column with update instance set updated = now(); because once that timestamp becomes older than 3 days your instance assumes it's "dead".

[–] dandroid@sh.itjust.works 1 points 11 months ago (1 children)

I checked a few instances and they were all last updated on November 18th. I ran that query and updated the updated field to now for all rows, but they aren't continuing to update after that, and no traffic is coming through.

[–] seahorse@midwest.social 3 points 11 months ago (1 children)

I believe anything prior to that timestamp will not federate. New stuff should.

[–] dandroid@dandroid.app 3 points 11 months ago (1 children)

It took a little while, but it eventually started working again! Thanks for the help.

[–] seahorse@midwest.social 3 points 11 months ago (1 children)

No problem! I had the same issue and was losing my mind trying to fix it. It still doesn't update the timestamp like it should automatically so I've been doing it myself.

[–] dandroid@dandroid.app 2 points 11 months ago

I might just write a script to do it if mine does the same. Hopefully this gets fixed in 0.19.0