On Fri, 2020-11-20 at 22:30 +0000, [email protected] wrote: > On Fri, 2020-11-20 at 18:44 +0100, Oleg Kalnichevski wrote: > > On Thu, 2020-11-19 at 10:29 +0000, [email protected] wrote: > > > > ... > > > > > On Wed, 2020-11-18 at 23:39 +0100, Oleg Kalnichevski wrote: > > > > On Wed, 2020-11-18 at 21:20 +0000, [email protected] wrote: > > > Hi Oleg, > > > Sure here is a reproducer using the simple benchmark that I've > > > been > > > fixing. > > > > > > you will need to.... > > > > > > a) get the branch i pushed [1] > > > b) get the artifacts to serve using httpd [2] > > > c) configure you httpd with a volume to serve the artifacts > > > d) install the 5.1-beta dependencies to your local maven > > > repository > > > e) run the Maven package goal > > > f) run the benchmark [3] > > > > > > The field Main.LIMIT controls the number of artifacts to > > > download. > > > The argument to the java -jar command controls the number of > > > download > > > repititions. > > > The class eu.janinko.foobar.httpmetrics.Appache2TaskProvider is > > > the > > > class that uses the httpcomponents-core api. > > > > > > Jeremy > > > > > > [1] > > > https://github.com/whitingjr/http2performance/tree/reproducer-reset-error > > > [2] > > > tps:// > > > www.dropbox.com/sh/1twdn5hma67qh2l/AABcdDxPKR2EaJ4bJZ6Aut_2a?dl=0 > > > [3] java -jar httpMetrics/target/httpMetrics-1.0-SNAPSHOT-jar- > > > with- > > > dependencies.jar 1 > > > > > > > Jeremy > > > > I re-ran our own compatibility tests and performance benchmarks > > with > > the latest HttpCore 5.1-beta-SNAPSHOT and found no evidence of > > anything > > being wrong. > > > > I will now try to reproduce the issue using your code but I would > > appreciate if you could also try to reproduce the issue > > with org.apache.hc.core5.benchmark.HttpBenchmark > > > > Oleg > > --- > > > > /opt/adopt-openjdk-11/bin/java <...> > > org.apache.hc.core5.benchmark.HttpBenchmark -c 20 -n 100000 -k -2 > > http://localhost:8080/ > > > > Server Software: Apache/2.4.46 (Unix) > > Protocol version: HTTP/2.0 > > Server Hostname: localhost > > Server Port: 8080 > > Document Path: http://localhost:8080/ > > Document Length: 13145 bytes > > > > Concurrency Level: 20 > > Time taken for tests: 15.743000 seconds > > Complete requests: 100000 > > Failed requests: 0 > > Kept alive: 0 > > Total transferred: 1320385280 bytes > > Content transferred: 1314500000 bytes > > Requests per second: 6,352.03 [#/sec] (mean) > > Time per request: 3.149 [ms] (mean) > > Time per request: 0.157 [ms] (mean, across all concurrent > > requests) > > Transfer rate: 81,905.53 [Kbytes/sec] received > > > > Process finished with exit code 0 > > > > > > > ----------------------------------------------------------------- > > > ---- > > > To unsubscribe, e-mail: > > > [email protected] > > > For additional commands, e-mail: > > > [email protected] > > > > > > > ----------------------------------------------------------------- > > ---- > > To unsubscribe, e-mail: [email protected] > > For additional commands, e-mail: > > [email protected] > > > Hi Oleg, > Thanks for the guide on running the Benchmark suite. I'll give it a > go > next week and let you know if there are any issues. > There do look to be some options that were not in the example code I > had used as a reference. These extra options may solve the benchmark > I > provided. > > Many thanks, > Jeremy >
Hi Jeremy I ran your benchmark against a local Apache HTTPD instance running in a docker container with the following set of request uris and 10000 repetitions and could not reproduce the problem --- http://localhost:8080/ http://localhost:8080/news.html, http://localhost:8080/status.html --- --- Apache HTTP/2: 5203.08 / 10000 = 0.520 ± 0.424 ms Process finished with exit code 0 --- As far as I understand all 10000 have been successfully executed. Oleg --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
