Simone Giusso

> > 1. I don't understand what exactly you're changing with the `DRIVER` and `PROTOCOL` options. Could you elaborate on how is this fixing the issue? What is the actual...

> connections from the pool are licked if an error occurs (and the 'cancel' signal happens) in parallel executions. That's it basically what has been tested [here](https://github.com/r2dbc/r2dbc-pool/issues/198#issue-1885269011). There is no...

Maybe related to https://github.com/r2dbc/r2dbc-pool/issues/198?

> I actually was able to reproduce the issue, it ran for about 25 minutes and ran into the problem. btw if you change the delaySubscription in the `runSomethingBadPeriodically` method...

> You can add a bit more debugging logging via `Mono.log()`, ideally on the `Mono` producing `usingWhen` to detect whether a cancellation signal is being propagated into `usingWhen`. It seems...

> I was not aware of the internals in Spring, I just had a look at r2dbc sources. I'll have a look then! Btw @chemicL I revised the author example...

Hello, as soon as I find the time I'll try to provide an example. Just as additional info, I tried limit the `CommandHandler` request queue size and even if the...