Akkoma db taking up a lot of space (relays configured to save 7 days )

database for akkoma is thirty gb on disk, i am sure that is federated content. is there a way to clear past entries brought in from relays? thanks

Have a look here, maybe that’ll help a bit.

is there a way to keep my local posts then “prune” all federated posts? thanks.

so i am thinking to run " /bin/pleroma_ctl database prune_objects --keep-threads --prune-orphaned-activities" however i am not sure it will only remove federated posts not local ones. is it available on akkoma-3.5? thanks.

if you’re subscribing to relays, this is expected

you’re pulling in FAR most posts than most instances ever will

i attempted to stay with smaller relays, though it looks like the size of the db is greater than likely for the retention configuration.

i changed it to 1 day. how to purge old federated data?


please note that deliberately pulling in a great volume of posts will always balloon the size of your database, and reclaiming that space requires intense IO activity on the vacuum

i would not recommend being subscribed to any relays for most instances if you are that concerned about 30gb

/opt/akkoma/bin# ./pleroma_ctl database prune_objects

09:31:59.516 [info] Pruning objects older than 1 days

09:32:12.464 [debug] QUERY OK source=“objects” db=12948.2ms idle=317.2ms
DELETE FROM “objects” AS o0 WHERE (o0.“data”->‘to’ ? $1 OR o0.“data”->‘cc’ ? $2) AND (o0.“inserted_at” < $3) AND (split_part(o0.“data”->>‘actor’, ‘/’, 3) != $4) [“ActivityStreams 2.0 Terms”, “ActivityStreams 2.0 Terms”, ~N[2023-06-04 21:31:59], “fedi.lamer-ethos.site”]

09:32:12.973 [debug] QUERY OK db=508.1ms queue=0.4ms idle=744.9ms
DELETE FROM hashtags AS ht
SELECT 1 FROM hashtags_objects hto
WHERE ht.id = hto.hashtag_id)

i get this error running the previous command

there is no error in this log

ok, thank you, running /bin/pleroma_ctl database prune_objects --keep-threads --prune-orphaned-activities now

the command has run for over an hour without completing. help please?

i aborted the command now akkoma backend will not load showing blank pages with the error “Request failed with status code 404 - Not implemented”

this is expected, you have pulled in hundreds of thousands of posts and are now asking your server to analyze and delete them all

it will take a long time

the documentation says in a very large box that they take a while

the ssh connection timed out before this command completed. akkoma now tells me “Request failed with status code 404 - Not implemented” when i login to admin, and no new messages are coming through. help please?

i also upgraded from 3.5 to 3.9, could it be to do with that? thanks.

i am assuming you read all the patch notes and carried out updates to the frontends as appropriate

yes updated frontend, followed upgrade instructions on update page. not sure why there is an error.

i am placing all of my bets on you not having actually updated the admin frontend

1 Like