Nothing changed on this regard.

Writes are aways sent on executors.. (On AIO, the main thread never block on a write, unless it is waiting for a sync callback).

There is aways a file opened, and another file is immediately created when that file is returned.


Perhaps, the only thing that changed was... we can get better throughput on the disk. If your disk is maxed out (your laptop for instance) the file creation may be competing with the journal write process. If the disk throughput is maxed out, the time to create a new file may be higher than the ammount of time needed to fill up a file, on that case we will have eventual pauses waiting for new files to be opened.

I aways have seen this happening before.. and I've done a lot of testing around this.


But I will take a look, see if we can improve anything on the process of files-creation.