https://bz.apache.org/bugzilla/show_bug.cgi?id=61616
--- Comment #28 from Carsten Wolff <[email protected]> --- Thanks Yann for your patience and all your work! I have some last observations to share: - When combining https://bz.apache.org/bugzilla/attachment.cgi?id=35427 and https://bz.apache.org/bugzilla/attachment.cgi?id=35446, I still get frequent stalls at the end of the transfer. Also, I observed a higher CPU usage by apache compared to e.g. apache-2.4.27 release. During transfer, the apache process used ~50% of one core in my test setup. For reference, I also ran the same test with squid3 as a connect proxy and it used ~9% of one core. - When combining https://bz.apache.org/bugzilla/attachment.cgi?id=35447 and https://bz.apache.org/bugzilla/attachment.cgi?id=35446, I can no longer produce any stall during or at the end of the transfer. Also, the CPU usage by apache is reduced to ~14% - When combining https://bz.apache.org/bugzilla/attachment.cgi?id=35447 and https://bz.apache.org/bugzilla/attachment.cgi?id=35446 and applying the settings you suggested in comment#24, everything still works and the CPU usage drops to ~8%. I will now have to run tests on my target platform, which is a bit more complicated, so it will take a few days, but I suppose we can assume this bug is fixed now. Thanks again! -- You are receiving this mail because: You are the assignee for the bug. --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
