Re: squid3-ipv6 squid3/src cf.data.pre,1.68.2.40,1.68.2.41 dns_internal.cc,1.15.6.29,1.15.6.30 structs.h,1.66.2.32,1.66.2.33

From: Henrik Nordström <henrik@dont-contact.us>
Date: Wed, 09 Jan 2008 09:36:14 +0100

ons 2008-01-09 klockan 18:05 +1300 skrev Amos Jeffries:

> Nice in principle, the problem becomes how to do it cleanly. Given that
> timeout is known only to client/server side, and the whole ipcache
> stands between the resolver and the point of most desirable lookup.

Not really. Connect timeouts & failures is supposed to be handled by the
commConnect state machine, which also is the one performing the DNS
lookup.. But yes, it's not done very good and there is also a timeout
managed by the caller, and due to timing connect timeout is usually
handled there.. To get the connect timeout handled by commConnect you
need to either set connect_timeout larger than your OS default connect
timeout, or fix Squid to tune the os timeout..

But it can be done cleanly within ipcache alone while ignoring the whole
timeout mess, by performing the A lookup if ipcache finds an entry where
all addresses is marked bad and there has not yet been an A lookup
performed.

Regards
Henrik

Received on Wed Jan 09 2008 - 01:38:54 MST

This archive was generated by hypermail pre-2.1.9 : Wed Jan 30 2008 - 12:00:09 MST