Best Regards
Fabio Buda
Web Developer/Designer @ netdesign
+39 338 9423302
fabi...@netd.it
http://www.netd.it/
If you need to restart a Tornado process (whether due to code changes,
hung process, or whatever), then having them all as discrete processes
makes this simpler and reduces the possibility of downtime (Nginx will
simply move on to the next backend if one is down). By contrast, if an
internal Tornado process needs to be restarted, you must restart all of
them at the same time.
For instance, I have a simple shell script that tells supervisor to
restart my instances in reverse order (in order to avoid racing with
Nginx's failover procedure), so I effectively get zero downtime even
with a full restart of my application.
Regards,
Cliff
--
Cliff Wells <cl...@develix.com>
-Ben
I'm not sure what would be the reliable way on monitoring children forks.
- Didip -
Best Regards,
Fabio Buda
thanks
Yes, I switched from Pound to Nginx about 5 years ago. Pound suffers
from the same problems as Apache: it is threaded and when the amount of
concurrency gets high it consumes massive amounts of RAM and then dies.
Pound is okay for smaller sites, but it's not really easier to setup
than Nginx and it has a smaller community, so I have never seen any
advantage to it.