On 8 October 2011 04:16, Lisper <[email protected]> wrote:
> I've dug into this some more, and this issue seems to be quite a rabbit
> hole.  The TL;DR version seems to be that calling fork() from a process that
> has threads is a Really Bad Idea (tm) and cannot be reasonably expected to
> work reliably.

If all the process is then doing is subsequent exec then it is
generally not a problem. It is when it wants to do mored complicate
stuff that it becomes an issue. So to blanket say it is a really bad
idea is plain wrong.

Sorry that haven't responded in this issue properly. Been tied up
doing other changes in mod_wsgi related to work that have been trying
to get on top of first. I will try and investigate it soon.

If someone wants to see if they can come up with a test program that
doesn't involve subprocess module and uses just fork/exec directly
then that would help.

The subprocess module has been a source of problems in the past with
assumptions it makes, such as flushing stdout/stderr after the fork,
ignoring the fact that there may be buffered data that parent would
also flush. End result is duplicated content in logs.

Graham

> If anyone wants details let me know.
>
> --
> You received this message because you are subscribed to the Google Groups
> "modwsgi" group.
> To view this discussion on the web visit
> https://groups.google.com/d/msg/modwsgi/-/ocxE7791-QgJ.
> To post to this group, send email to [email protected].
> To unsubscribe from this group, send email to
> [email protected].
> For more options, visit this group at
> http://groups.google.com/group/modwsgi?hl=en.
>

-- 
You received this message because you are subscribed to the Google Groups 
"modwsgi" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/modwsgi?hl=en.

Reply via email to