On 07 Oct 2015, at 3:43 PM, Stefan Eissing <stefan.eiss...@greenbytes.de> wrote:
> Having just had time to look at which test cases fail: I see that static > resources via HTTP/2 seem to work fine, however my tests with a proxy and or > rewrite in between fail with high likelihood. > > Any hint at what exactly I might have to look for, any hint about what > actually has changed, would be appreciated. In terms of requests you seem to be doing the right thing. You create a request with h2_task_create_request(), which was in turn called from h2_task_process_request(), which then calls ap_process_request() which handles the request and then passes an EOR bucket down the end of the chain. I think the problems start once the above is done - are you performing any sort of manual cleanup of either requests or connections or other pools that are parents of connections or pools? If you do you’re probably destroying the request before it is finished going over the network. A request is started and you then forget about it, when the core processes the EOR bucket the request will disappear on it’s own sometime at a future date. Can you describe how cleanups occur in the http2 world? Regards, Graham —