We've seen an interesting problem lately while running daemon mode with a large number of Apache children..
The basic issue is that we have apache with MaxClients at 600 -which is high, I know - and 24 mod_wsgi children. (We have a lot of RAM, and this is a 8 CPU box) - so requests come in and apache queues them up. Under load, the mod_wsgi daemons are pretty busy, and the apache children start to pile up waiting to be served. This is all well and good and we figured this was a good way to absorb load spikes, but we've seen something strange. When the client (i.e. browser, or whatever) times out, and drops the socket on apache, the request seems to already be queued up for mod_wsgi to run. So even after the HTTP connection has dropped, mod_wsgi chews on the request anyway... Is this known behavior in mod_wsgi, or does it at least make sense for the way mod_wsgi is today, or does this sound like something else? We are likely going to drastically reduce MaxClients but I wanted to run this hypothesis by this list to see if we're evaluating this right. Alec -- 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.
