Problem on dcc 1.3.30 - Continue Not Asking DCC...

Breno Moiana breno@haxent.com.br
Thu Mar 9 15:00:40 UTC 2006


Greetings.

I am new to this list, even though I have been digging through the 
archives for quite some time now.

I am having a problem that is very similar to the one reported by Gary 
Mills[1] on January 26th 2006.

[1] http://www.rhyolite.com/pipermail/dcc/2006/003019.html

According to the CHANGES file on the source tree, two fixes were made on 
version 1.3.28 that could have solved Gary's problem. However, I am 
still going thorugh it. Here is my report:

We have a DCC server set up on an email provider, handling around 3 
million email messages a day.
The server is a dual Xeon 3.2 64 bits, with 3GB RAM, and it is running 
only the DCC and a small mySQL server, which we use only for graylisting 
(an in-house developed graylisting solution).

The operating system is CentOS4.1, in its 64-bit build.


Without any apparent reason, something happens to DCC that makes it stop 
responding. Here is the log from the beginning of the problem:

: Mar  9 09:29:38 dcc dccifd[4782]: no DCC answer from 127.0.0.1,6277 
after 18264 ms
: Mar  9 09:29:38 dcc dccifd[4782]: continue not asking DCC 64 seconds 
after failure
: Mar  9 09:29:38 dcc last message repeated 4 times
: Mar  9 09:29:39 dcc dccifd[4782]: continue not asking DCC 63 seconds 
after failure
: Mar  9 09:29:39 dcc last message repeated 8 times
: Mar  9 09:29:40 dcc dccifd[4782]: continue not asking DCC 62 seconds 
after failure
: Mar  9 09:29:40 dcc last message repeated 8 times
: Mar  9 09:29:41 dcc dccifd[4782]: continue not asking DCC 61 seconds 
after failure
: Mar  9 09:29:41 dcc last message repeated 4 times
: Mar  9 09:29:42 dcc dccifd[4782]: continue not asking DCC 60 seconds 
after failure
: Mar  9 09:29:42 dcc last message repeated 2 times
: Mar  9 09:29:42 dcc dccifd[4782]: no DCC answer from 127.0.0.1,6277 
after 22221 ms
: Mar  9 09:29:42 dcc dccifd[4782]: continue not asking DCC 128 seconds 
after failure
: Mar  9 09:29:42 dcc last message repeated 2 times
: Mar  9 09:29:43 dcc dccifd[4782]: continue not asking DCC 127 seconds 
after failure
: Mar  9 09:29:43 dcc last message repeated 2 times
: Mar  9 09:29:44 dcc dccifd[4782]: continue not asking DCC 126 seconds 
after failure
: Mar  9 09:29:44 dcc last message repeated 9 times
: Mar  9 09:29:45 dcc dccifd[4782]: continue not asking DCC 125 seconds 
after failure
: Mar  9 09:29:45 dcc last message repeated 8 times
: Mar  9 09:29:45 dcc dccifd[4782]: continue not asking DCC 124 seconds 
after failure
: Mar  9 09:29:46 dcc last message repeated 5 times
: Mar  9 09:29:46 dcc dccifd[4782]: no DCC answer from 127.0.0.1,6277 
after 26375 ms
: Mar  9 09:29:46 dcc dccifd[4782]: continue not asking DCC 256 seconds 
after failure
: Mar  9 09:29:46 dcc last message repeated 3 times
: Mar  9 09:29:47 dcc dccifd[4782]: continue not asking DCC 255 seconds 
after failure
(...)

As you can see, it doubles the waiting time until it reahes the maximum 
of 2048 seconds, which is then repeatedly counted down.

Please notice that the RTT to the server remains low all the time, at 
around 50ms.


Not always, when I manually run the cron-dccd script, the errors stop:

(...)
: Mar  8 17:54:20 dcc dccifd[4782]: continue not asking DCC 1994 seconds 
after failure
: Mar  8 17:54:20 dcc last message repeated 4 times
: Mar  8 17:54:21 dcc dccifd[4782]: continue not asking DCC 1993 seconds 
after failure
: Mar  8 17:54:22 dcc dccd[4748]: 1.3.30 database /var/dcc/dcc_db 
reopened with 2016 MByte window

Then no more entries related to this error are seen in the log.
This doesn't happen every time I run the cron script.
Also, on 1.3.30 it sometimes just stops giving errors and gets back to 
work, correctly calculating and counting the checksums. On 1.3.27/29 we 
didn't see the server getting back to work without human intervention. 
Even when we did get it back to work, we couldn't say for sure what made 
it happen.

We were originally running 1.3.27 when this error first showed up, then 
upgraded to 1.3.29, and now to 1.3.30.

Any help will be greatly appreciated, as we are falling into RBLs every 
other day, due to the eventual lack of DCC service (we allow email to 
pass when the DCC doesn't respond)

Thanks for the attention!

Best Regards,

Breno Moiana
================
Haxent Consulting







More information about the DCC mailing list

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