On 02.03.2016 16:19, 'sri' via Mojolicious wrote: > My first guess is an uneven distribution of connections among the > workers, perhaps setting -M 2 or -M 5 would fix that.
I tried it again with the -M2 and -M5 settings as suggested (see the file names). https://github.com/kraih/mojo/issues/925#issuecomment-191302254 With a single-digit number of workers, I can see (using the precision measurement program top(1)) that at first, the load distribution of the workers is quite even, each at 100% CPU, and after a while, it gets a little more and more uneven. With 10 or more workers, the load of all workers seems to cave in after a while, sometimes even after very few seconds, with *none* of the workers bearing a significant load any more. Except that 20 workers indeed looks like a sweet spot, with the load caving in only later (but it does) in some cases. Not in all. So, behaviour is quite erratic for whatever reason. Oh, and by the way, this is on Debian Jessie, with the Debian-provided Perl 5.14.2. Regards, Juergen. -- <[email protected]> Tel +49.30.838-50740 Fax -450740 Zentraleinrichtung fuer Datenverarbeitung, Central Systems (Unix) 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.
