A few more runs, now with Mojolicious-6.51; most of them seem not quite
as erratic, but I have still seen the perl processes stop producing load
after a few (or very few) seconds with 10 workers or more. And sometimes
with fewer.
It is really particular that the connect problems only ever happen with
16 and 18 workers.
Workers Run1 Run2 Run3 Run4 Run5 Run6 Run7
2 3263 3264 3324 2792 3323 3297 3116
4 6755 6898 6716 6594 6561 6972 6728
6 10567 9821 9816 10579 10410 10509 10153
8 6406 5901 4548 5899 4745 5197 4764
10 6389 1424 2785 2104 2189 4546 3535
12 3497 14446 9811 11146 9839 9666 11098
14 14334 4082 11642 10435 12228 9964 9641
16 5883 2705 4456 3556
18 14424 10190 1638 10466
20 6658 4647 2903 11885 2426 18637 10773
22 3653 6010 10079 2863 10111 4565 3689
24 14368 10516 11680 10238 12374 1681 10440
(Requests/s each)
To compare, the same "wrk" run against another page on the same machine,
that is served by Apache with a 301 and the standard Apache "Moved
Permanently" template (a redirect to the HTTPS site) is good for 110361
requests/s.
For the raw data see
https://github.com/kraih/mojo/issues/925#issuecomment-191348484
Regards, Juergen.
--
<[email protected]> Tel +49.30.838-50740 Fax -450740
Zentraleinrichtung fuer Datenverarbeitung, Special Intelligence
Freie Universitaet Berlin, Fabeckstrasse 32, 14195 Berlin, DE
--
You received this message because you are subscribed to the Google Groups
"Mojolicious" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
Visit this group at https://groups.google.com/group/mojolicious.
For more options, visit https://groups.google.com/d/optout.
