We're seeing a strange thing occur on occasion where we can no longer reach a couple sites behind an ISA reverse proxy. Both their OWA (Exchange) webmail solution and RPC over HTTPS connections to the Exchange box suddenly start to hang during times of increased load. When there are (guessing) less than 40 or so people using the service everything is fine and there are no errors in the debug logs. When more people start to load up, access gets super slow or times out completely. During those times Squid starts to report these: 2010/12/15 11:55:16.981| errorConvert: %%U --> 'https://rpc.xxxxxxxx.com/*' 2010/12/15 11:55:16.981| errorConvert: %%U --> 'https://rpc.xxxxxxxx.com/*' 2010/12/15 11:55:16.981| errorConvert: %%W --> '?subject=CacheErrorInfo%20-%20ERR_CONNECT_FAIL&body=CacheHost%3A%20elvis.xx xxxxxxx. com%0D%0AErrPage%3A%20ERR_CONNECT_FAIL%0D%0AErr%3A%20(60)%20Operation%20time d%20out%0D%0ATimeStamp%3A%20Wed,%2015%20Dec%202010%2019% 3A55%3A16%20GMT%0D%0A%0D%0AClientIP%3A%2010.61.11.128%0D%0AServerIP%3A%20216 .24.136.185%0D%0A%0D%0AHTTP%20Request%3A%0D%0ACONNECT%20 %2F%20HTTP%2F1.0%0AUser-Agent%3A%20MSRPC%0D%0AHost%3A%20rpc.xxxxxxxx.com%0D% 0AContent-Length%3A%200%0D%0AProxy-Connection%3A%20Ke ep-Alive%0D%0A%0D%0A%0D%0A' 2010/12/15 11:55:16.983| errorConvert: %%U --> 'https://webmail.xxxxxxxx.com/*' 2010/12/15 11:55:16.983| errorConvert: %%U --> 'https://webmail.xxxxxxxx.com/*' 2010/12/15 11:55:16.983| errorConvert: %%W --> '?subject=CacheErrorInfo%20-%20ERR_CONNECT_FAIL&body=CacheHost%3A%20elvis.xx xxxxxxx. com%0D%0AErrPage%3A%20ERR_CONNECT_FAIL%0D%0AErr%3A%20(60)%20Operation%20time d%20out%0D%0ATimeStamp%3A%20Wed,%2015%20Dec%202010%2019% 3A55%3A16%20GMT%0D%0A%0D%0AClientIP%3A%2010.61.10.210%0D%0AServerIP%3A%20216 .24.136.184%0D%0A%0D%0AHTTP%20Request%3A%0D%0ACONNECT%20 %2F%20HTTP%2F1.0%0AUser-Agent%3A%20Mozilla%2F4.0%20(compatible%3B%20MSIE%208 .0%3B%20Windows%20NT%205.1%3B%20Trident%2F4.0%3B%20.NET% 20CLR%201.1.4322)%0D%0AProxy-Connection%3A%20Keep-Alive%0D%0AContent-Length% 3A%200%0D%0AHost%3A%20webmail.xxxxxxxx.com%0D%0APragm a%3A%20no-cache%0D%0AProxy-Authorization%3A%20Basic%20ZGZpbm46NTM2OTYzNTE%3D %0D%0A%0D%0A%0D%0A' 2010/12/15 11:55:17.587| fd_close FD 65 webmail.xxxxxxxx.com:443 2010/12/15 11:55:17.588| fd_close FD 394 webmail.xxxxxxxx.com:443 2010/12/15 11:55:17.590| fd_close FD 478 webmail.xxxxxxxx.com:443 2010/12/15 11:55:17.591| fd_close FD 530 webmail.xxxxxxxx.com:443 2010/12/15 11:55:17.592| fd_close FD 531 rpc.xxxxxxxx.com:443 2010/12/15 11:55:17.598| httpRequestFree: rpc.xxxxxxxx.com:443 2010/12/15 11:55:17.598| clientLogRequest: al.url='rpc.xxxxxxxx.com:443' 2010/12/15 11:55:17.599| fd_close FD 13 rpc.xxxxxxxx.com:443 2010/12/15 11:55:17.608| httpRequestFree: rpc.xxxxxxxx.com:443 2010/12/15 11:55:17.608| clientLogRequest: al.url='rpc.xxxxxxxx.com:443' 2010/12/15 11:55:17.609| fd_close FD 382 rpc.xxxxxxxx.com:443 2010/12/15 11:55:17.627| httpRequestFree: webmail.xxxxxxxx.com:443 We have the remote sites setup to not cache with the following: acl NeverCache dstdomain xxxxxxxxxxx.com always_direct allow NeverCache cache deny NeverCache I guess I'm confused by the "CacheError" instance if we have the sites set to not cache. Searching the lists on "errorConvert" has proven unsuccessful for our symptoms. We're running squid-3.1.8 on FreeBSD 6.2. Thanks! -Steffan
