Patrik Nordwall

https://github.com/akka/akka/runs/4495316564?check_suite_focus=true

https://github.com/akka/akka/runs/4554855749?check_suite_focus=true

In Aeron multi-node: https://github.com/akka/akka/runs/5297795403?check_suite_focus=true ``` 2022-02-23T04:10:49.5699744Z [INFO] [02/23/2022 04:10:49.569] [pool-1-thread-1] [ArteryTransport(akka://OutboundIdleShutdownSpec)] Started embedded media driver in directory [/opt/volumes/media-driver/OutboundIdleShutdownSpec-6f396db2-27f9-4dda-a2fa-2e9bf75c76e4] 2022-02-23T04:10:49.5824043Z [INFO] [02/23/2022 04:10:49.581] [pool-1-thread-1] [ArteryTransport(akka://OutboundIdleShutdownSpec)] Remoting started with transport [Artery aeron-udp];...

If this fails again we can switch to retry as discussed in PR

As far as I know, `nanoTime` should never go backwards, so it's interesting to investigate how this can happen. Wonder if there could be some "thinko" in that `now -...

Was this observed on Windows? I trust this answer https://stackoverflow.com/a/8853738

Sounds like a good idea.

Note the description at the top, it's not only about improving raw performance but avoiding > As long as new callbacks arrive, the stream would be completely stalled running the...

not sure why pr validation is not running here. I'll close-open to see if that helps.

we could drop this test completely, or just for JDK 17. osgi is not a priority, #28304