You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
move the threadpool into something managed by the core or MPM and include extra scoreboard handles at creation time (this happens really early though)
ask the MPM's to schedule and run the h2 threads (probably easiest in event)
find a way for a module to ask for an over-allocation the scoreboard handles AND coordinate thread offsets.
This would allow the h2 threads to have their actual slots int he scoreboard. Currently they all share the original connection processing threads scoreboard handle and mockup some event internal structures
The text was updated successfully, but these errors were encountered:
I had a closer look at scoreboard and what it does. And looked at /server-status after some loadtesting. Seeing threads that go into 'G' and never leave that state. Does not look ok, afaik.
As first mitigation, added some ap_update_child_status() into my main I/O handling. But that does not solve the problem.
For the 2.4 release, I need a way for mod_h2 to co-exist with its own threads, since I think substantial changes in MPMs are not going to fly.
icing
changed the title
scoreboard handles for h2 threads
Feature: scoreboard handles for h2 threads
Sep 2, 2015
So, the scoreboard (mod-status) can be extended, as see by balancers and ssl. Use this extension mechanism to display the status of all active http2 workers. Show request being processed and progress.
It seems like we need a way to do one of:
This would allow the h2 threads to have their actual slots int he scoreboard. Currently they all share the original connection processing threads scoreboard handle and mockup some event internal structures
The text was updated successfully, but these errors were encountered: