> On 20 Nov 2015, at 10:26 PM, 'Lukasz Szajkowski' via modwsgi
> wrote:
>
> For now I have workaround with following mod_wsgi configuration
> processes=12 threads=1 inactivity-timeout=300 maximum-requests=1000
>
> and it does addresses the problem.
>
> But there is
below is Python code level stack traces and one more pstack
for the following configuration
ErrorLog "/opt/logs/pyws/vhost_pyws_prod_error.log"
CustomLog "/opt/logs/pyws/vhost_pyws_prod_access.log" combined
WSGIDaemonProcess pyws.prod processes=1 threads=25
Thanks, for suggestion.
Could you please suggest somebody in the London UK area
who could advise on mod_wsgi (ideally with Oracle)
Thanks,
Lukasz
--
You received this message because you are subscribed to the Google Groups
"modwsgi" group.
To unsubscribe from this group and stop receiving
> On 30 Sep 2015, at 5:55 pm, 'Lukasz Szajkowski' via modwsgi
> wrote:
>
> Thanks, for suggestion.
> Could you please suggest somebody in the London UK area
> who could advise on mod_wsgi (ideally with Oracle)
Unfortunately not. If you can get those Python code level
>
> Thank you Graham for answering while on vacations !!
>
Unfortunately this didn't help same problem with same stack trace.
Does it make any difference that I have in Apache conf
WSGISocketPrefix ../../var/run/wsgi
WSGIPythonHome /opt/python_envs/pyframe
and in virtual host configuration
> On 29 Sep 2015, at 6:25 pm, 'Lukasz Szajkowski' via modwsgi
> wrote:
>
> Thank you Graham for answering while on vacations !!
>
> Unfortunately this didn't help same problem with same stack trace.
>
> Does it make any difference that I have in Apache conf
>
>
> On 29 Sep 2015, at 5:10 am, 'Lukasz Szajkowski' via modwsgi
> wrote:
>
> The actual stack trace while hanging is
> #0 0x7f047ff099b0 in sem_wait () from /lib64/libpthread.so.0
> #1 0x7f0474b4c6d8 in PyThread_acquire_lock_timed () from
>
Thanks, Interesting hypothesis but:
- we use dedicated connection with persistent connection in Django
- the problem is observed in performance testing under constant heavy load
On Monday, 28 September 2015 17:08:28 UTC+1, Bill, KE1G wrote:
>
> One possibility:
>
> If cx_oracle is using
One possibility:
If cx_oracle is using connection pooling and it is not round robin (or
you just have large inactivity periods)
AND
There is a firewall between cx_oracle and oracle that times out unused
connections.
Then the firewall could be locking traffic for connections that
Hi ,
I have a problem with mod_wsgi (3.4-14) / Apache 2.4.12 / cx_oracle 5.2 /
Oracle 12 / Django 1.8.2 hanging under bigger load.
The configuration was used for the last several month in dev / testing.
Now when deployed to production environment Apache hangs after couple of
minutes under
10 matches
Mail list logo