continue not asking DCC

Vernon Schryver vjs@calcite.rhyolite.com
Fri Dec 26 15:44:53 UTC 2003


> From: Valentin Chopov <valentin@valcho.net>

> Sorry for returning to this but I'm continuing getting a lot of this:
> Dec 26 08:09:07 milter dccm[41486]: continue not asking DCC 17 seconds
> after failure
> Dec 26 08:09:08 milter dccm[41486]: continue not asking DCC 16 seconds
> after failure
> Dec 26 08:09:09 milter dccm[41486]: continue not asking DCC 15 seconds
> after failure
>
> /var/dcc/map contains 3 DCC servers localhost (127.0.0.1) + 2 servers on
> local Ethernet network  (100baseTX <full-duplex>). So I'm sure that we don't
> have network problems here ;)

I've seen plenty of 100BaseTX networks that have problems.  Autonegotiation
added more fun to the things that can happen with 10BaseTX.

> It looks that some time the current DCC server with the lowest RTT stops
> responding and dccm starts "continue not asking..." complaints.
> Is it possible in  this case, dccm to start asking the next available DCC
> server (maybe based on the RTT) or it can be just a simple round robin.

Dccm has always switched to the next server based on RTT.  You can
test it by shutting down the DCC server that dccm is currently using.

All 3 of your servers seem to be not answering dccm at once.  That is
probably what is happening, perhaps because they are all running dbclean
at once (many not at 08:00) or because something has happened to your
network.  It could also be that dccm only thinks that all 3 DCC servers
are failing because the system running dccm is so busy that dccm does not
manage to run more than once every dozen seconds.


Vernon Schryver    vjs@rhyolite.com



More information about the DCC mailing list

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