On Sat, May 11, 2002 at 07:46:42PM +0100, Alan J. Flavell wrote:
>
>
> That's about where we're at. Does this ring any bells, please? I can
> supply some more bdiffs if wanted.
>
Yes, that rings bells with me. We had exactly that problem here.
Files downloaded using our ISP's squid cache as a parent would corrupt
the file but going direct would result in a correct download. I saw
exactly the same corruption you saw, random http headers inserted into
the file. The odd thing is that we have an internal hierarchy of
squid caches but the corruption either never happens or is random
enough not to be noticed - it was only when our "border" squid used
our ISP's squid as a parent would the problem happen. I tried working
on this with my ISP but I just started to get back a "it works ok for
me" response, lost patience and just made our squid go direct to the
internet. Since our link is a flat rate one we could do that without
penalty. That is probably not much help for you.
-- Brett LymnReceived on Sun May 12 2002 - 06:39:51 MDT
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 17:08:04 MST