Alex Xu
Alex Xu
@mattklein123 @ggreenway would you like to give a high-level review on this? I didn't add any tests yet but I did a manual test and passed all existing tests. And...
@mattklein123 we have three listener fields for three specific socket options, they are `transparent`, `freebind` and `tcp_fast_open_queue_length`, I think they should work with an update. Do we want to support...
> @mattklein123 we have three listener fields for three specific socket options, they are `transparent`, `freebind` and `tcp_fast_open_queue_length`, I think they should work with an update. Do we want to...
/wait the duplicate address check isn't right, working on a fix
/wait still have test need to fix
> This approach LGTM. > > Will this also work for hot-restart? I think socket options changes aren't applied for hot-restart either. Yes, we doesn't change socket options when hot-restart,...
> Thanks this is great. Flushing out some more comments. For all the new tests, please try to run them many times to look for flakes. https://github.com/envoyproxy/envoy/tree/main/test/integration#reproducing-test-flakes > > /wait...
> > Yes, we doesn't change socket options when hot-restart, the existing socket option already applied on the socket which passed from the parent, the OS suppose ensure there is...
> > The simple solution here is to suggest the user don't change the socket option when hot-restart, the user can do another update for socket option after hot-restart. Also...