[Bug]: Error: pool timed out while waiting for an open connection about atuin HOT 6 CLOSED

As I'm on wsl I can't be sure what IO is happening under the hood but the computer was pretty much idling. Writing helmcharts in emacs and a browser open.

But the error did occur when I did git mv on a directory and that can potentially be io-heavy for a brief moment.

I'll increase local_timeout and see how it goes.

You could turn this into a feature request to improve the error message.

from atuin.

Self hosted server, just now install the Atuin v18.2.0 - Build rev from Debian package

Executing via Supervisorctl
fronting it with Caddy
Postgresql 15.6

Using 4x servers and parallel terminal access, 3 works, and the fourth throws the error - "consistently"

IS that a RUST pool limit?

from atuin.

This occurs when we cannot write to your disk within a reasonable time

This could be due to IO pressure, or because you're using a filesystem like ZFS

essentially there are two options

1. Hang until the write is complete
2. Timeout

1. was the old behaviour

We've introduced a background daemon that's currently on main and keeps filesystem writes out of the hot path, but realistically if you're having IO problems writing to SQLite within 2 seconds you will probably have issues elsewhere too.

See this for config option: #1726

I'll be closing this because it's not really a bug

from atuin.

Hmmm...

and when I write to Postgresql on the server?

or is this a problem on the client only?

from atuin.

This issue is for a client error. If you're seeing something similar on the server you might want to check on your Postgres setup

from atuin.

This occurs when we cannot write to your disk within a reasonable time

This could be due to IO pressure, or because you're using a filesystem like ZFS

Strange, there just wasn't any io, so I'm hard press to believe the IO pressure as a cause.

essentially there are two options

1. Hang until the write is complete

2. Timeout

3. was the old behaviour

We've introduced a background daemon that's currently on main and keeps filesystem writes out of the hot path, but realistically if you're having IO problems writing to SQLite within 2 seconds you will probably have issues elsewhere too.

Hmmmm... I here what you are saying, though I'm hard pressed to believe it on these servers, though I think there might be something else involved causing the delay, so will keep an eye out

I'll be closing this because it's not really a bug

from atuin.