the homeserver has a database outage (a 32-bit txn id wraparound) since 07:15 UTC - we're on the case. apologies for the downtime; folks on their own homeserver should be feeling smug right now.


...and it's back. sorry about that. looks like long-lived queries to fix the schema to support more than 2^32 rows stopped autovacuums from running, which meant we hit a 2^32 txn id wraparound instead. πŸ™ƒ


Can you write a blogpost about this that explains the issue, and how to mitigate it?

Sign in to participate in the conversation's Mastodon is one server in the network