At 10:49 04/11/96 +0100, hhoff@modern-video.de wrote:
>some time ago you posted on squid-users about squid generating lots
>of repeating error messages on Linux. I have the same problem on
>a Linux 2.x machine and was wondering if you've gotten any hints
>on this..
N.B. I'm putting this on the list.
I've now investigated it a little, and I know why it happens now, whenever
source_ping is defined in squid.conf, squid sends a test packet to the
source on udp port 7 (echo). However some sites presumably have acl'ed
routers/filtering that instead return an icmp port unreachable message, so
squid complains.
I thought I had a cunning plan of using the icmp error in other ways, but I
made a bad logical flaw. Not enough coffee...
Would it be possible for ICMP port unreachables to be dropped if it is not
from a neighbor/sibling? Also when this happens, a single ICMP port
unreachable causes two separate errors, so perhaps one of them should have
its log level toned down, probably the "icpHandleUdp error_receiving" one.
I'm not sure (and in fact doubt) if this also explains the "comm_udp_recv:
recvfrom failure: Operation not supported on transport endpoint" messages,
so if anyone knows where this does come from, could they post why?
P.S. As a total change of subject, would it be a good idea to give some of
the US, French and British hierarchies as examples in squid.conf? Then
people can just uncomment the relevant line(s) for their region, and play a
part in their hierarchy.
Jonathan L.
Origin IT Services Ltd., 323 Cambridge Science Park, Cambridge, England.
Tel: +44 (1223) 423355 Fax: +44 (1223) 420724 E-mail: guess...
-------[ Do not think that every sad-eyed woman has loved and lost... ]------
-----------------------[ she may have got him. -Anon ]-----------------------
These opinions are all my own fault.
Received on Tue Nov 05 1996 - 02:59:42 MST
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:33:29 MST