Alex Rousskov <rousskov@ircache.net> writes:
> Except that all _disk_ operations/handlers are blocking. Select/poll(2)
> always return "ready" on a disk file descriptor. Blocking disk I/Os were
Absolutely (and yet someone spent a real long time writing the
original code as if they were able to be non-blocking.)
> the motivation behind introducing real threads into Squid (the now
> officially dead async-io stuff).
oh, this is a shame - I've obviously not been following closly enough,
recently. Is anyone looking at implementing in a different manner (eg
multiple processes to schedule the reads/writes in a way akin to the
dnsservers)?
Otherwise, it remains a pretty serious cap on the performance of a
single Squid-running system.
James.
Received on Sat Oct 02 1999 - 09:39:48 MDT
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:48:42 MST