DELAY_HACK

From: David Luyer <luyer@dont-contact.us>
Date: Thu, 14 Aug 1997 17:37:53 +0800 (WST)

I've updated the DELAY_HACK patches I've been using here since the 1.0.x
days again, this time to 1.1.15 (not tested yet, but the patch only
rejected in parts due to nearby code changes and not any changes to the
code being patched). The documentation is probably in a post I did to
squid-users a while back and in the patch to the squid.conf.pre (after
all, isn't squid.conf where all the documentation is?). These patches
also contain the other misc changes used at UWA and some other places
around here.

Put simply, these patches give the ability to control the amount of
traffic...

* for all items matching one ACL "delay_child_access" on the basis of a
  single combined pool

* for all items matching the ACL "delay_access" on the basis of
       - a combined pool
       - a per-8-bit-net pool
       - a per-host pool
  (note: these are currently based on the assumption that if the last
         16-bits of an IP are unique, it's the same pool. easy enough to
         fix.)

each neighbour can be tagged as "no-delay" if fetches from it shouldn't be
delayed. neighbours or parents without this tag are delayed/slowed based
on avaliable traffic in the "pool", which is configured with an upper
limit and a per-second feed. there's a cachemgr interface to examine all
the delay pools and so on.

Duane: could you possibly put a link to ftp://ftp.uwa.edu.au/pub/squid on
the "Patches for Squid-1.1" page? This is where I maintain my DEBUG_HACK
patches.

David.

PS: Who else is going to AUUG97? (Australians know what I mean)
Received on Tue Jul 29 2003 - 13:15:42 MDT

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:11:22 MST