On 8/03/2012 1:00 a.m., Jaime Gomez wrote:
Hi all,
First of all I want to apologyse if this question has been solved before but I haven't found anything related with this.
We have a very weird issue with some https web pages. Some of them are very, very slow. After doing some debugging we have this in our cache.log
2012/03/07 11:10:41.072| httpParseInit: Request buffer is CONNECT ebanking.rbcdexia-is.es:443 HTTP/1.1
Host: ebanking.rbcdexia-is.es
2012/03/07 11:10:41.072| HttpMsg.cc(445) parseRequestFirstLine: parsing possible request: CONNECT ebanking.rbcdexia-is.es:443 HTTP/1.1
Host: ebanking.rbcdexia-is.es
2012/03/07 11:10:41.072| urlParse: Split URL 'ebanking.rbcdexia-is.es:443' into proto='', host='ebanking.rbcdexia-is.es', port='443', path=''
Host: ebanking.rbcdexia-is.es
2012/03/07 11:10:41.073| aclMatchDomainList: checking 'ebanking.rbcdexia-is.es'
2012/03/07 11:10:41.073| aclMatchDomainList: 'ebanking.rbcdexia-is.es' found
2012/03/07 11:10:41.073| The request CONNECT ebanking.rbcdexia-is.es:443 is ALLOWED, because it matched 'allowBancos'
2012/03/07 11:10:41.073| The request CONNECT ebanking.rbcdexia-is.es:443 is ALLOWED, because it matched 'allowBancos'
2012/03/07 11:10:41.073| clientProcessRequest: CONNECT 'ebanking.rbcdexia-is.es:443'
2012/03/07 11:10:41.074| tunnelStart: 'CONNECT ebanking.rbcdexia-is.es:443'
2012/03/07 11:10:41.074| fd_open() FD 20 ebanking.rbcdexia-is.es:443
2012/03/07 11:10:41.074| peerSelectFoo: 'CONNECT ebanking.rbcdexia-is.es'
2012/03/07 11:10:41.075| commConnectStart: FD 20, data 0x5b5fd60, ebanking.rbcdexia-is.es:443
2012/03/07 11:10:41.075| commConnectStart: FD 20, cb 0x5ac9200*1, ebanking.rbcdexia-is.es:443
2012/03/07 11:10:41.075| ipcache_nbgethostbyname: Name 'ebanking.rbcdexia-is.es'.
2012/03/07 11:10:41.075| ipcacheRelease: Releasing entry for 'ebanking.rbcdexia-is.es'
2012/03/07 11:10:41.075| ipcache_nbgethostbyname: MISS for 'ebanking.rbcdexia-is.es'
2012/03/07 11:10:41.075| idnsALookup: buf is 41 bytes for ebanking.rbcdexia-is.es, id = 0x5535
The DNS entries TTL period has expired. Squid is looking up new ones
when the timeout happens.
Can your resolve that domain easily yourself? If not, that is the problem.
If you can resolve it easily please add debug section 78,5 and make a
new trace.
Amos