Mark Thornton
2003-05-29 14:21:36 UTC
We have resolved the external tests problem by installing the djbdns caching
dns server (http://cr.yp.to/djbdns.html) on the amavisd/spamassassin relay
server. Prior to installation the log was showing TIMING reports of ranging
from 10000 to 50000 ms, with the average being about 18000 ms. If we enabled
any external test this was the result. Now all messages are processing with
TIMING results of 300 to 2600 ms with the average being about 1200 ms. Since
making the change our system has not fallen behind or been under any
significant strain at all.
The reason: apparently the dns load from the relay server was sufficient to
drive our primary dns server's cache into some form of corruption. It would
resolve local sites fine but not external sites until reset. As a result the
relay server was then failing over to the secondary dns server to complete
the external queries. Since the primary server was partially responding it
seemed to always get first crack but failed. The relay server could put the
dns server into this state in about 18 hours, but I had never seen this
before under normal usage. Looks like I need to work on the dns servers...
Mark Thornton
San Marcos Internet, Inc
512-393-5300
-------------------------------------------------------
This SF.net email is sponsored by: eBay
Get office equipment for less on eBay!
http://adfarm.mediaplex.com/ad/ck/711-11697-6916-5
_______________________________________________
AMaViS-user mailing list
AMaViS-***@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/amavis-user
AMaViS-FAQ:http://www.amavis.org/amavis-faq.php3
AMaViS-HowTos:http://www.amavis.org/howto/
dns server (http://cr.yp.to/djbdns.html) on the amavisd/spamassassin relay
server. Prior to installation the log was showing TIMING reports of ranging
from 10000 to 50000 ms, with the average being about 18000 ms. If we enabled
any external test this was the result. Now all messages are processing with
TIMING results of 300 to 2600 ms with the average being about 1200 ms. Since
making the change our system has not fallen behind or been under any
significant strain at all.
The reason: apparently the dns load from the relay server was sufficient to
drive our primary dns server's cache into some form of corruption. It would
resolve local sites fine but not external sites until reset. As a result the
relay server was then failing over to the secondary dns server to complete
the external queries. Since the primary server was partially responding it
seemed to always get first crack but failed. The relay server could put the
dns server into this state in about 18 hours, but I had never seen this
before under normal usage. Looks like I need to work on the dns servers...
Mark Thornton
San Marcos Internet, Inc
512-393-5300
-------------------------------------------------------
This SF.net email is sponsored by: eBay
Get office equipment for less on eBay!
http://adfarm.mediaplex.com/ad/ck/711-11697-6916-5
_______________________________________________
AMaViS-user mailing list
AMaViS-***@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/amavis-user
AMaViS-FAQ:http://www.amavis.org/amavis-faq.php3
AMaViS-HowTos:http://www.amavis.org/howto/