Hi,
performed the test on 2.9.11-stable+timestamp.2014.09.15.23.35.11
(Running on Apache/2.4.7 (Ubuntu), Python 2.7.6):
I utilised apache benchmark (ab), making 1130 request within a minute twice
with and without log.close().
I did not notice any memory leak, (neither with the log nor with the
python 2.7.8 64bit, Linux.
My rig is heavily armored but it shouldn't matter for this case: just for
completeness, it's a 6 core xeon x5660 , 24GB RAM, running *buntu 14.10.
test scenario1:
console1) python web2py.py -a mypwd
console2) ab -n 1000 http://127.0.0.1:8000/welcome/default/index
appointment app where having memory leaks
https://github.com/mdipierro/web2py-appliances/tree/master/AppointmentManager
https://groups.google.com/d/msg/web2py/u-eU-2vhims/1u3fOnqPoUMJ
https://groups.google.com/d/msg/web2py/QbR7pY3xHuM/rOyuHhVmbjYJ
Richard
On Mon, Feb 2, 2015 at 2:26 PM,
@niplhod do you know if is there a way to integrate memory profiling in
travis-ci??
I don't see it in any of the docs. if you have something in mind that will
print out something as a value, we could print an alarm on a crafted build
if some value passes a threshold, but that's pretty
I have been using this sometimes ago, if there is a cyclic reference or a
stale reference of some sort, it help to spot it.
https://groups.google.com/d/msg/web2py-developers/Rd8hC5aFRZo/UTxZHjAwWcUJ
http://mg.pov.lt/objgraph/
Il giorno lunedì 2 febbraio 2015 22:48:17 UTC+1, Louis Amon ha
@Leonel you should make a PR of this. We have a program profiler, a memory
profiler would be great!
@niplhod do you know if is there a way to integrate memory profiling in
travis-ci??
Il giorno lunedì 2 febbraio 2015 00:35:50 UTC+1, Leonel Câmara ha scritto:
If you have pool_size defined
What application is this? Are you using cache.ram or a cache decorator?
That would cause this.
On Sunday, 1 February 2015 03:18:05 UTC-6, Louis Amon wrote:
I have a web2py server running on heroku.
So far my server had little traffic, and Heroku being a cloud PaaS service
means they put
Update:
I've tried running 2 dynos to see if splitting traffic in two would reduce
the slope of my memory leak, and also to see if the garbage collector would
somehow work better.
It seems not :
2015-02-01T09:23:50.049734+00:00 heroku[web.1]: *source**=web.1* dyno=heroku
gluon.main.wsgibase calls gluon.main.serve_controller, which calls
gluon.compileapp.run_models_in, which executes the models.
Anthony
On Sunday, February 1, 2015 at 4:18:05 AM UTC-5, Louis Amon wrote:
I have a web2py server running on heroku.
So far my server had little traffic, and Heroku
Udate 2:
I've tried switching from the Rocket server to Gunicorn : the memory's
still increasing and quotas are reached in a matter of hours.
BTW I'm running web2py version 2.9.11.
Has the new patch fixed anything that might help with memory leaks ? Should
I try updating my server ?
--
Do you have pool_size defined?
--
Resources:
- http://web2py.com
- http://web2py.com/book (Documentation)
- http://github.com/web2py/web2py (Source code)
- https://code.google.com/p/web2py/issues/list (Report Issues)
---
You received this message because you are subscribed to the Google Groups
If you have pool_size defined can you tell me if putting pool_size=0 in
your DAL solves this problem?
Anyway I got really interested in this, for some unknown to me reason, and
decided to create a *VERY CRUDE* decorator to memory check your controller
functions for leaks, I guess with time we
No I do not have pool_size defined. Can it help ?
From: Leonel Câmara
Sent: Sunday, 1 February 2015 16:26
To: web2py@googlegroups.com
Do you have pool_size defined?
--
Resources:
- http://web2py.com
- http://web2py.com/book (Documentation)
- http://github.com/web2py/web2py
I was just curious if there was some bug where if you had a pool size then
the DAL would leak memory it appears not. Try the decorator I posted in the
other thread.
--
Resources:
- http://web2py.com
- http://web2py.com/book (Documentation)
- http://github.com/web2py/web2py (Source code)
-
Hi,
the garbace collector is called once every 100 requests.
See
https://github.com/web2py/web2py/blob/5a0ee722605164f1eae4064d0d1dd0d72b5eb14b/gluon/main.py#L197
In addition try to pack a very basic app which can reproduce the issue,
we'll have more info about what is going on wrong in your
15 matches
Mail list logo
