> I'll look out for that throttle value if it happens again then. > > Been having a play with the new uwsgi in dev. The problem is that I > don't have any kind of repro for the problem, so it's hard to know what > to look for when testing. > > But here's one minimal repro, which is related but not the same, which > you can try if you want: > > - set up a couple of vassals with simple flask apps > - check they both work > - now set them to do the flask `app.run(port=8001)` (or 8002). Notice > that this breaks the vassals. They get stuck in a reload loop. sites > 502 (if using nginx). > - now try and un-break one of them, remove the app.run. On my setup, > the app never comes back, and if you try and hit them, the server now > never responds to request -- they just hang, you don't even get a 502... > > And that happens with the old uwsgi and with the new one.... >
Ok, this is enough to work on. For the new one you means the one from github ? > > -- > Harry Percival > Developer > [email protected] > > PythonAnywhere - a fully browser-based Python development and hosting > environment > <http://www.pythonanywhere.com/> > > PythonAnywhere LLP > 17a Clerkenwell Road, London EC1M 5RD, UK > VAT No.: GB 893 5643 79 > Registered in England and Wales as company number OC378414. > Registered address: 28 Ely Place, 3rd Floor, London EC1N 6TD, UK > > On 11/03/14 18:01, Roberto De Ioris wrote: >>> throttle_level looks to be set to 2. Can't see any OOMs in the dmesg >>> logs, no... >>> >> hi, i mean during the "cascade problem", if that value is high it could >> explain the problem. >> >> By the way, have you tried latest code ? >> >>> -- >>> Harry Percival >>> Developer >>> [email protected] >>> >>> PythonAnywhere - a fully browser-based Python development and hosting >>> environment >>> <http://www.pythonanywhere.com/> >>> >>> PythonAnywhere LLP >>> 17a Clerkenwell Road, London EC1M 5RD, UK >>> VAT No.: GB 893 5643 79 >>> Registered in England and Wales as company number OC378414. >>> Registered address: 28 Ely Place, 3rd Floor, London EC1N 6TD, UK >>> >>> On 10/03/14 16:22, Roberto De Ioris wrote: >>>>> Can you think of any reason why the cascade is happening? why all >>>>> new >>>>> vassals are affected? >>>> The only thing popping-up in my mind is that for some reason you have >>>> a >>>> very-high throttle level (it is the throttle_level attribute in the >>>> emperor stats). >>>> >>>> So you are effectively blocked from spawning new processes (for a >>>> pretty >>>> long time) >>>> >>>> During the "problem" you see the vassal list in the stats server, so >>>> the >>>> case of a corruption in the linked list is not a possibility. >>>> >>>> Can you check if there is some OOM in the dmesg ? >>>> >>> _______________________________________________ >>> uWSGI mailing list >>> [email protected] >>> http://lists.unbit.it/cgi-bin/mailman/listinfo/uwsgi >>> >> > > _______________________________________________ > uWSGI mailing list > [email protected] > http://lists.unbit.it/cgi-bin/mailman/listinfo/uwsgi > -- Roberto De Ioris http://unbit.it _______________________________________________ uWSGI mailing list [email protected] http://lists.unbit.it/cgi-bin/mailman/listinfo/uwsgi
