[Feature] Auto-clean-up synapse database

[Gredin67]

what about implementing a cron job to run such a clean-up script by default with synapse_ynh ? It would improve a major synapse drawback, and improve privacy.

https://foss-notes.blog.nomagic.uk/2021/03/matrix-database-house-cleaning/
https://gitea.prk.st/romain/synapse-tools/src/branch/master/synapse-purge.sh

As these guys are maintaining a production synapse server I guess they would keep their clean-up scripts up-to-date.

[Thatoo]

I feel that I have already read this somewhere...
I agree, synapse database becomes big quiet fast.
I would like a script that :

• delete all data from a room that all members of our server have left.
• delete all data older than two month from big rooms that was not originally created on our server (in our case, we didn't create big room in our server)
• but I want to keep all history of direct chats and all history of rooms members of our servers have initiated if at least one member is still in it.

Of course I have no idea if its doable. Just requesting :-)

[tio-trom]

That would be a great addition!

[tio-trom]

Can we merge this with #354 ?

[Gredin67]

If you open a PR maybe :) But we may rather rely on the message retention policy and synadm event cleanup API ? Le 19 septembre 2023 02:21:04 GMT+02:00, Tio ***@***.***> a écrit :

…

Can we merge this with #354 ? -- Reply to this email directly or view it on GitHub: #283 (comment) You are receiving this because you authored the thread. Message ID: ***@***.***>

[zyphlar]

Running this script as a CRON job has been useful for me, though it'd be nice if it were just part of YunoHost (or Synapse!)

#!/bin/bash

set -eu

source /usr/share/yunohost/helpers

app=synapse

db_name=$(ynh_app_setting_get --app=$app --key=db_name)
db_user=$(ynh_app_setting_get --app=$app --key=db_user)
db_pwd=$(ynh_app_setting_get --app=$app --key=db_pwd)
server_name=$(ynh_app_setting_get --app=$app --key=server_name)

iptables -A INPUT -s 172.17.0.0/16 -p tcp -m tcp --dport 5432 -j ACCEPT

docker run -it ungleich/rust-synapse-compress-state:2022-06-19 synapse_auto_compressor -p postgresql://$db_user:$db_pwd@172.17.0.1/$db_name -c 500 -n 100

iptables -D INPUT -s 172.17.0.0/16 -p tcp -m tcp --dport 5432 -j ACCEPT

ynh_psql_execute_as_root --database=$db_name --sql="vacuum full"

exit 0