Conversation
Notices
-
Embed this notice
>cawfee still down
Even a single gochiusa post would've saved it..
-
Embed this notice
@r
>It's just that one step of the maintenance is being throttled pretty hard, and how to put it... My initial estimate was between hours and days, so I just let it run its course. But it's becoming increasingly obvious that if I were to let it run its course, at its own pace, it could take months or years. And I'm getting over letting it run at its own pace.
is he facing the pleroma database importing experience that's a rite of passage for all admins at this point
-
Embed this notice
@r @wowaname See you in a decade
-
Embed this notice
@wowaname He's running a script to delete the old remote activities from the database. The script is part of Pleroma. I haven't looked at the query, but apparently, it's not very efficient. That combined with the slow networked storage, what could go wrong?
-
Embed this notice
@r @wowaname > networked storage
Lmao, but even with hetzners shitty block storage it never took even an hour for me iirc
-
Embed this notice
@nyanide @r
-
Embed this notice
@r @wowaname hetzner one is so slow that it leads to constant errors as queues overflow. That's why this instance doesn't run on their hw anymore
-
Embed this notice
@snacks @wowaname BuyVM's block storage is really slow.