env_From incrementing

John Doherty jdoherty@gstype.com
Sat Dec 10 05:43:21 UTC 2005


On Friday 09 December 2005 07:21 pm, Vernon Schryver wrote:

>> X-DCC-NIET-Metrics: is1.gstype.com 1080; bulk env_From=14 Body=1
>>   Fuz1=1 Fuz2=1

> Could you check with `ps -wwwwvax | grep dccm` or `ps -lef | grep
> dccm` or some other way that dccm is not running with a -t threshold
> that would explain it?

I mistakenly said earlier that I was not using a -t option, but you're 
right and "ps -wwax" shows that I am using one. It's what you'd 
expect from DCCM_REJECT_AT and DCCM_LOG_AT in dcc_conf, though:
-t CMN,10,10.

I think I've misunderstood the meaning of DCCM_ARGS in dcc_conf: I 
thought those were the only arguments passed to dccm. Since my 
DCCM_ARGS doesn't have a -t option, I mistakenly thought it was 
running without one (and using DCCM_LOG_AT and DCCM_REJECT_AT without 
being passed those values on its command line). Sorry for that 
confusion.

>> I can't reproduce the problem. I tried sending a couple of dozen
>> messages from my home to the shop, and all were accepted just
>> fine, though they were checked against a different DCC server
>> (X-DCC-Misty-Metrics).

> Few if any public DCC servers other than dcc.niet.net run with a
> `dccd -K` value that saves env_From checksums.

> It should be easy to reproduce the syndrome by adding an entry to
> /var/dcc/map biased to dcc.niet.net with

> cdcc "add dcc.niet.net RTT-2000 ms"

I will try this. My understanding is that this gives a preference to 
that server, adjusted by -2 seconds to whatever its preference would 
be anyway, so that it will likely be consulted first.

> I've done more testing with something like that and still failed to
> get anything to pay attention to the env_From totals.

Thanks very much for looking into it. I'll try to make it happen again 
if I can.

-- John



More information about the DCC mailing list

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