Significance of all-zero checksum counts

Gary Mills mills@cc.UManitoba.CA
Mon Jan 5 17:49:56 UTC 2004


On Fri, Jan 02, 2004 at 09:24:47PM -0700, Vernon Schryver wrote:
> 
> I've encountered an example of this syndrome in my logs.  However, it
> did not involve a failure by a DCC server to answer, but a switch in
> DCC servers between an initial transmission and greylist embargo and
> the final transmission.   The DCC client reports a message to the
> current DCC server when it is first seen and greylist embargoed.  Each
> time the client sees it again, the client only does a query and not a
> report.  If the client switches to another server before the final
> greylist retransmission, and if the message is not known as bulk among
> the DCC servers, then it is likely that the second DCC server will say
> it has never heard of the message.  In other words, the final DCC
> counts will be zero.

Doesn't the flooding between DCC servers keep then synchronized?
Maybe it's the time lag between updates that contributes to this
problem?  In any case, if it only results in the counts being zero
instead of one, it's not a big deal.

> The best way to avoid such zero counts is to make DCC clients
> prefer a particular server with RTT adjustments in /var/dcc/map.

Won't this defeat the load balancing?  We sometimes get wide variations
in server loading.  Having multiple DCC servers is essential in those
situations.

How much RTT adjustment do you recommend?

-- 
-Gary Mills-    -Unix Support-    -U of M Academic Computing and Networking-



More information about the DCC mailing list

Contact vjs@rhyolite.com by mail or use the form.