On Wed, Apr 10, 2019 at 5:15 PM Jonathan Vanasco wrote:
>
>
>
> On Wednesday, April 10, 2019 at 3:29:47 PM UTC-4, Mike Bayer wrote:
>>
>> The connection pool is safe in single process multiple threads. It breaks
>> on fork() because the socket travels across the process boundary and is
>>
On Wednesday, April 10, 2019 at 3:29:47 PM UTC-4, Mike Bayer wrote:
>
> The connection pool is safe in single process multiple threads. It breaks
> on fork() because the socket travels across the process boundary and is
> essentially copied between two or more processes that are unaware of
The connection pool is safe in single process multiple threads. It breaks
on fork() because the socket travels across the process boundary and is
essentially copied between two or more processes that are unaware of each
other.
On Wed, Apr 10, 2019, 2:55 PM Jonathan Vanasco wrote:
>
>
> On
On Wednesday, April 10, 2019 at 1:17:01 PM UTC-4, Mike Bayer wrote:
>
> op, you're right, that is a thing.Not sure if mod_wsgi makes it
> easy to make that mistake though, that is, I thought all the Python
> happens after the fork. Well, if they are using daemon mode which
> you
On Wed, Apr 10, 2019 at 1:06 PM Jonathan Vanasco wrote:
>
>
>
> On Wednesday, April 10, 2019 at 9:23:02 AM UTC-4, tonthon wrote:
>>
>> Our services are served through apache and mod_wsgi (1 process, 10 threads).
>>
>> ...
>
>
>>
>> I can't reproduce this problem that seems to happen randomly.
>
>
On Wednesday, April 10, 2019 at 9:23:02 AM UTC-4, tonthon wrote:
>
> Our services are served through apache and mod_wsgi (1 process, 10
> threads).
>
> ...
>
> I can't reproduce this problem that seems to happen randomly.
>
>
Considering this is multi-process/thread system - does any of your