mod_wsgi: Reload issue after an exception
Hello,
It seems that the application processes can be stuck and cannot get reloaded via the touch method when some kind of exceptions occurs.
When it happens, the only way to restart the app is to reload Apache. It’s something which can be unexpected and annoying in production and even more in development.
I’ve tested and reproduced the problem with both 3.4 (Ubuntu 14.04) and 4.3.0 (Ubuntu 16.04), using daemon mode.
It seems this is the same issue as described here.
Do you know what could be wrong? Is it a known issue or something that I’m doing wrongly?
If you want to dig into this issue and cannot reproduce it, I could provide an example.
Thank you.
About this issue
- Original URL
- State: closed
- Created 7 years ago
- Comments: 16 (9 by maintainers)
@HoangJerry Create a separate issue, please don’t add your own problem as comment on existing issue, especially when looks different. When you create the separate issue, supply the full stack trace and details of the exception. You have left it off. If you are seeing issue about Django not being re-entrant, ensure you are getting the first error after the server start, not the last.
The
startup-timeoutoption was added a year ago so it isn’t really new. If you are using operating system packages for mod_wsgi you really should consider installing mod_wsgi yourself. The operating system packages for distros such as Debian and Ubuntu are usually very out of date, not supported and don’t get fixed.