https://bz.apache.org/bugzilla/show_bug.cgi?id=61616
--- Comment #37 from Carsten Wolff <cars...@wolffcarsten.de> --- (In reply to Yann Ylavic from comment #36) > Sorry for the delay Carsten, just to clarify, you are running httpd with > attachment 35513 [details] [Handle POLLOUT in proxy_connect (v8)] and > attachment 35447 [details] [ Read buffer size and flush threshold tuning], > no other patch right? Almost. I added these two patches to packages from debian sid (https://packages.debian.org/sid/apache2) and to the SuSE package from obs (https://build.opensuse.org/package/show/Apache/apache2), so there are some distribution patches in there, but nothing essential at this time. > While attachment 35513 [details] alone should work for the proxy_connect > case, perfomances are better with tunings from attachment 35447 [details], > still right? Yes, the single-transfer test runs for about the same time, but cpu load is less with attachment 35447, which promises better performance with concurrency. > If this is the case, I'd suggest using attachment 35513 [details] only for > now, that's less performances (than expected?) but the same as the original > proxy_connect ones. I did a quick test without attachment 35447 and it seems you're right, the regression seems to be in that patch. I will do more thorough testing tomorrow. > I'm looking at what could cause regressions in attachment 35447 [details], > but this patch is work in progress anyway and not really related to this PR > I think. Running the client (wget) on localhost, I cannot reproduce the error. Running it from another machine, it's there 100% of the time. The logs of both cases look exactly the same, except for the last two lines: GOOD: [Mon Dec 18 13:36:59.042282 2017] [core:trace6] [pid 3507:tid 140042357954304] core_filters.c(449): [client 192.168.122.11:43236] core_output_filter: flushing because of max threshold [Mon Dec 18 13:36:59.042875 2017] [core:trace6] [pid 3507:tid 140042357954304] core_filters.c(857): [client 192.168.122.11:43236] writev_nonblocking: 512276/512276 bytes [Mon Dec 18 13:36:59.042929 2017] [core:trace6] [pid 3507:tid 140042357954304] core_filters.c(458): [client 192.168.122.11:43236] core_output_filter: flushing because of FLUSH bucket BAD: [Mon Dec 18 13:37:21.164185 2017] [core:trace6] [pid 3537:tid 140042248849152] core_filters.c(449): [client 192.168.122.12:51428] core_output_filter: flushing because of max threshold [Mon Dec 18 13:37:21.164499 2017] [core:trace6] [pid 3537:tid 140042248849152] core_filters.c(857): (11)Resource temporarily unavailable: [client 192.168.122.12:51428] writev_nonblocking: 86880/512276 bytes [Mon Dec 18 13:37:21.164548 2017] [core:trace6] [pid 3537:tid 140042248849152] core_filters.c(449): [client 192.168.122.12:51428] core_output_filter: flushing because of max threshold [Mon Dec 18 13:37:21.164566 2017] [core:trace1] [pid 3537:tid 140042248849152] core_filters.c(571): (70007)The timeout specified has expired: [client 192.168.122.12:51428] core_output_filter: writing data to the network -- You are receiving this mail because: You are the assignee for the bug. --------------------------------------------------------------------- To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org For additional commands, e-mail: bugs-h...@httpd.apache.org
