you might have noticed we lost a bunch of posts...
in short: we tried to fix a Lemmy issue that started on 9/13 and is causing or has caused problems with apps for the software. we're aware of basically all of them besides Sync exhibiting some sort of problematic behavior since this issue arose. however: attempting to fix this issue on our side of things did not go well.
i am not qualified to make a formal write-up of the exact mechanism of what went wrong here--and in this case the mechanism isn't really important for your informational purposes, although @Penguincoder@beehaw.org can elaborate as needed--but basically we made a fix that seemed fine in testing and was not fine when actually applied. when it became obvious something was wrong, we tried restoring literally everything but the database. that, unfortunately, did not work, so eventually we just pulled the trigger on restoring from a backup. this has lost at least some posts, but probably no more than 8 hours of them by our estimation.
hopefully this will not happen again, either upstream or on our side of things.
I just want to say that the downtime made me extra sure that I will follow beehaw wherever it goes. I'm definitely more lurky than chatty, but after browsing via other instances, I missed beehaw.
The more I look at Lemmy...the more I think this platform is just not ready for primetime. I suppose that's obvious since its version is still sub 1.0. Idk.
I'm hoping that Postmill makes the cut. It looks very easy to work on, but I haven't had the time to set up an instance and poke around. Maybe I'll do that today.
it seems pretty clear that we aren't staying on Lemmy or on federated services at this point due to beehaw's policies and lack of fediverse moderation.
Did the fix end up working in the end, or did restoring everything from backups mean that the fix didn't work out either? (I don't use Jerboa or any other lemmy apps, so mainly just curious.)
This was the initially issue I was trying to fix; apps not loading. Tracked it down to the timezone handling of lemmy. The 'fix' (upgrading to upstream lemmy code) didn't work, and broke a lot else.