[JDEV] Stress testing and connection limits

[Colin Putney]

-- Chuck Bohling <chuck at ultris.com> noticed the connection limit 
imposed by pth:

>> This works great up to about 1000 (1024?) clients
>> (threads). Above that number, I start seeing all kinds of
>> failures. I think
>> it may be caused by pth_accept rejecting the connection, but

-- Eric Bowersox mentioned a solution in the works:

> We (the jserver developers, that is) are currently developing a
> solution that uses additional processes as "multiplexers" to
> aggregate large amounts of connections.  With this, it would be
> theoretically possible to increase the number of server connections
> to as many network sockets as your system will handle (which is more
> like 64K).  This isn't ready yet, though; if someone knows different,
> please step in and correct me...

Being a voice crying in the wilderness, I have a an alternate solution 
ready, but in need of testing. Instead of select() (via pth) it uses 
AIO  to wait for input from the network. I've been working on a stress 
client as Chuck describes, but got hit by a landslide and haven't been 
able to finish it yet.

Chuck: wanna help me test it?

igneous

---------------------------------------------
Colin Putney         cputney at whistler.net
Whistler Networks    http://www.whistler.net/