Ondřej Kuzník authored Jun 27, 2017 and Ondřej Kuzník committed Nov 17, 2020

If we can't find an upstream, we keep the client around, so it needs to
be unblocked.

Ondřej Kuzník authored May 25, 2017 and Ondřej Kuzník committed Nov 17, 2020

Potentially for timeout detection purposes in the future.

Ondřej Kuzník authored May 18, 2017 and Ondřej Kuzník committed Nov 17, 2020

We reset the connection pointer on a destruction attempt, avoid the
spurious asserts.

Ondřej Kuzník authored May 18, 2017 and Ondřej Kuzník committed Nov 17, 2020

Do not log when receiving the last bytes on a connection. Log failed
writes.

Ondřej Kuzník authored May 10, 2017 and Ondřej Kuzník committed Nov 17, 2020

Several threads calling backend_select might reset current_backend to a
different place, there are two options to deal with that:
- just let the last rotation win (the current approach)
- detect whether first == current_backend and only replace then

Not sure which one is more useful, going with the simpler.

Ondřej Kuzník authored May 10, 2017 and Ondřej Kuzník committed Nov 17, 2020

This alone doesn't make the server do a round robin.

Ondřej Kuzník authored May 09, 2017 and Ondřej Kuzník committed Nov 17, 2020

There was still a race where the connection could be freed as the
operation was still being used.