What I would like to see is a way to use the cachemgr.cgi to go in and stop
runaway fetches...
Without having to down the entire cache itself.
(like when I forgot to filter out the 208.228.228.0 class-C network from
entering into our Squid 2.1.patch2 servers... Requests are made by the
McAfee virusscan update software itself, and I think it uses byterange
requests... which are also uncacheable cgi calls with the users's
registration code as part of the request). After a mere hour of attempting
to service those, our T1 was flooded with requests because of those (there
were 10 requests going to that network at the time). After restarting
Squid, the flow went right back to normal. (approx 800k/s)
What's worse, the request doesn't seem to stop when the client aborts the
connection. (And yes our abort settings are correct; my guess is that with
byterange requests then the first part is exempt from the speed limit, and
seems to be exempt from aborting)
--Jeff Barrow, Internet Connections, Inc.
>Just wondering if anyone knew of a WWW front end to Squid
>configuration - something like the CacheFlow admin system, for anyone
>who's familiar with that.
Received on Thu Feb 11 1999 - 23:41:07 MST
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:44:32 MST