Reap expired connections on drop #224
Closed
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
djc
reviewed
Oct 15, 2024
bb8/tests/test.rs
Outdated
| let conn = pool.get().await; | ||
|
|
||
| // And wait. | ||
| tokio::time::sleep(Duration::from_secs(2)).await; |
There was a problem hiding this comment.
Nit: prefer to just import sleep() for this.
| @@ -149,12 +149,19 @@ where | |||
| "handled in caller" | |||
| ); | |||
|
|
|||
| let max_lifetime = self.inner.statics.max_lifetime; | |||
| let is_expired = max_lifetime.map_or(false, |lt| conn.is_expired(Instant::now() - lt)); | |||
There was a problem hiding this comment.
Suggest we bind let is_broken = self.inner.manager.has_broken(&mut conn.conn); here and then use is_broken || is_expired in the match scrutinee.
tneely
force-pushed
the
tneely/reap-on-drop
branch
from
b57440a
to
5494c10
Compare
|
I'm using bb8 inside of PgCat to do transaction pooling. The servers I'm connecting to get cycled periodically for security reasons, so if the reaper lags too far behind then we end up querying a dead server and reporting the connection as broken. I'd rather reap it before we get to that point. |
tneely
force-pushed
the
tneely/reap-on-drop
branch
from
5494c10
to
4d7cab2
Compare
djc
reviewed
Oct 16, 2024
bb8/src/inner.rs
Outdated
| (ConnectionState::Present, false) => locked.put(conn, None, self.inner.clone()), | ||
| (_, is_broken) => { | ||
| match (state, is_broken || is_expired) { | ||
| (ConnectionState::Present, false) if !is_expired => { |
There was a problem hiding this comment.
So we don't need if !is_expired anymore, right?
There was a problem hiding this comment.
Good catch, I totally missed that
tneely
force-pushed
the
tneely/reap-on-drop
branch
from
4d7cab2
to
eba52a3
Compare
The reaper only runs against the connections in its idle pool. This is fine for reaping idle connections, but for hotly contested connections beyond their maximum lifetime this can prove problematic. Consider an active connection beyond its lifetime and a reaper that runs every 3 seconds: - [t0] Connection is idle - [t1] Connection is active - [t2] Reaper runs, does not see connection - [t3] Connection is idle This pattern can repeat infinitely with the connection never being reaped. By checking the max lifetime on drop, we can ensure that expired connections are reaped in a reason amount of time (assuming they eventually do get dropped).
tneely
force-pushed
the
tneely/reap-on-drop
branch
from
eba52a3
to
341432d
Compare
|
See follow-up in #226. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
The reaper only runs against the connections in its idle pool. This is fine for reaping idle connections, but for hotly contested connections beyond their maximum lifetime this can prove problematic.
Consider an active connection beyond its lifetime and a reaper that runs every 3 seconds:
This pattern can repeat infinitely with the connection never being reaped.
By checking the max lifetime on drop, we can ensure that expired connections are reaped in a reason amount of time (assuming they eventually do get dropped).