--MimeMultipartBoundary
Content-Type: text/plain; charset=us-ascii
> The most common query I get with the caches is the large number of
> 'connection reset by peer' messages when accessing large numbers of
> sites.
>
> I don't know if this is only under linux?
Don't think so. I'm pretty sure I've seen this under Slowaris and also
FreeBSD.
> Has anyone investigated this?
>
> (this is again, almost always, on loaded IIS servers)
Its seems the IIS servers when under load reset connections. This happen a
lot with .asp it seems, probably because .asp is so horrible slow and used
in situations when it shouldn't be used.
I'm not sure if this is a valid thing for them to do (I looked at the HTTP
specs and can't find the bit that says if your a shit server running on a
shit OS your allowed to reset the connection where you loaded).
I have _never_ seen this with a non IIS server, but I haven't looked hard.
What I did do one day was capture all the time this happened and then go and
probe which WWW server they were using using a perl script.
Anyhow, an option to wait N*rand() where N E [1..5) seconds might be useful
for these cases.
-Chris
--MimeMultipartBoundary--
Received on Tue Jul 29 2003 - 13:15:49 MDT
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:11:46 MST