I think you guys misunderstand jhawk. The problems here are the
*default* concurrency setting and the apparently widespread belief
within the qmail community that there are no adverse effects
associated with maximizing concurrency.
The bottom line here is that right now you are killing your own users
by using qmail. Qmail is almost certainly using up the network so
thoroughly when it's running that other people trying to use the
network have to be pretty much scrogged. Running this number of
concurrent TCP connections effectively disables all TCP's fairness
algorithms, which are what keep the network usable for everybody.
It seems like a cinch that apps with people waiting for them (e.g.,
telnet, the web) are more important than offline stuff like email.
But qmail does the network equivalent of applying nice --20 to
sendmail.
A default maximum, at least for network connections, on the order of
two, would seem more appropriate. Or one could institute a policy of
only forking off processes when a name lookup or message transmission
hangs for a while. Yes, qmail will be slower that way. But that time
qmail loses will go towards letting *users'* packets through, keeping
them happy. Which is the important part.
Jon
*default* concurrency setting and the apparently widespread belief
within the qmail community that there are no adverse effects
associated with maximizing concurrency.
The bottom line here is that right now you are killing your own users
by using qmail. Qmail is almost certainly using up the network so
thoroughly when it's running that other people trying to use the
network have to be pretty much scrogged. Running this number of
concurrent TCP connections effectively disables all TCP's fairness
algorithms, which are what keep the network usable for everybody.
It seems like a cinch that apps with people waiting for them (e.g.,
telnet, the web) are more important than offline stuff like email.
But qmail does the network equivalent of applying nice --20 to
sendmail.
A default maximum, at least for network connections, on the order of
two, would seem more appropriate. Or one could institute a policy of
only forking off processes when a name lookup or message transmission
hangs for a while. Yes, qmail will be slower that way. But that time
qmail loses will go towards letting *users'* packets through, keeping
them happy. Which is the important part.
Jon