Is there any way to achieve UDP logging and use a round robin of IPs or DNS names? What are possible solutions to achieve this type of solution?
Thanks, Tony On Thu, Jul 25, 2013 at 3:00 AM, <[email protected]> wrote: > Send uWSGI mailing list submissions to > [email protected] > > To subscribe or unsubscribe via the World Wide Web, visit > http://lists.unbit.it/cgi-bin/mailman/listinfo/uwsgi > or, via email, send a message with subject or body 'help' to > [email protected] > > You can reach the person managing the list at > [email protected] > > When replying, please edit your Subject line so it is more specific > than "Re: Contents of uWSGI digest..." > > > Today's Topics: > > 1. Re: Help with nginx "prematurely closed connection" error > (Django app) (Cycle Belfast) > 2. Re: Possible memory leak on http plugin (Roberto De Ioris) > > > ---------------------------------------------------------------------- > > Message: 1 > Date: Wed, 24 Jul 2013 12:05:47 +0100 > From: Cycle Belfast <[email protected]> > To: [email protected] > Subject: Re: [uWSGI] Help with nginx "prematurely closed connection" > error (Django app) > Message-ID: <[email protected]> > Content-Type: text/plain; charset=UTF-8; format=flowed > > On 24/07/13 06:05, Roberto De Ioris wrote: > >>> > >>> On 23/07/13 14:54, ?ukasz Mierzwa wrote: > >>>> @Cycle switch from file sockets in fastrouter-subscription-server and > >>>> subscripe2 options to port based ones. > >>>> > >>>> > >>> With this I don't get the Protocol error in the uwsgi log, but I still > >>> get the same "prematurely closed connection" error in the nginx log. > >>> Puzzling. > >>> > >>> gmf > >>> > >>> > >> (note: be sure to reply to [email protected]) > >> > >> I think there is something missing, i will send you a test config > tomorrow > >> to check why your system is resolving unix paths as udp. > >> > >> > >> > >> > >> > > Well, open 2 terminals: > > > > on the first one) > > > > uwsgi --fastrouter /tmp/sock1 --fastrouter-subscription-server /tmp/sock2 > > --master > > > > on the second one) > > > > uwsgi --socket /tmp/sock3 --subscribe-to /tmp/sock2:foobar > > > > run exactly this commands. What you see in the logs ? > > > > > That seems to work: > > subscribing to /tmp/sock2:foobar > > in the subscriber, and: > > [uwsgi-subscription for pid 2251] new pool: foobar (hash key: 9533) > [uwsgi-subscription for pid 2251] foobar => new node: /tmp/sock3 > > in the master. Haven't seen that up to now. > > (I have family duties today, so I'll get back to this tomorrow. Thanks > for your attention up to now.) > > gmf > > > > ------------------------------ > > Message: 2 > Date: Thu, 25 Jul 2013 09:17:27 +0200 > From: "Roberto De Ioris" <[email protected]> > To: "uWSGI developers and users list" <[email protected]> > Subject: Re: [uWSGI] Possible memory leak on http plugin > Message-ID: > <[email protected]> > Content-Type: text/plain;charset=utf-8 > > > > On Jul 23, 2013, at 7:45 AM, Roberto De Ioris <[email protected]> wrote: > > > >> Ok, can you enable --http-stats ? it would be useful to have its output > >> after 24 and 48 hours. > > > > What numbers do you think are interesting from that output? > > mainly the number of active sessions > > > > >> Currently the problem could be in: > >> > >> - ssl sessions (hard to catch) > > > > You mean ssl session caching, or just plain ssl usage? > > plain ssl usage, openssl make a performance-oriented usage of memory but > your numbers are really huge so there must be some kind of problem (unless > you have a high number of concurrent sessions) > > > > >> - uWSGI cache remote synchronization (i can write a test suite for it) > > > > I'm not synchronizing caches between uwsgi nodes, if that's what you > mean. > > > > ok, we can remove it from the list :) > > >> - uWSGI http-router auto-chunked (you can eventually test it using the > >> chunked routing action and removing auto-chunked) > > > > I have disabled auto-chunking in frontend 1, disabled ssl session cache > in > > frontend 2, and left everything the same in frontend 3. After a day, the > > memory usage seems to have grown the same in all 3 frontends. It's > > probably too early to tell anyway... > > > it is a feature i use too (without https) so i am doubtful it is the guilty > > I will investigate on plain ssl usage > > Thanks > -- > Roberto De Ioris > http://unbit.it > > > ------------------------------ > > _______________________________________________ > uWSGI mailing list > [email protected] > http://lists.unbit.it/cgi-bin/mailman/listinfo/uwsgi > > > End of uWSGI Digest, Vol 46, Issue 29 > ************************************* >
_______________________________________________ uWSGI mailing list [email protected] http://lists.unbit.it/cgi-bin/mailman/listinfo/uwsgi
