On Sat, 14 Jun 2003 15:19:47 +0200, Henrik wrote: >Hmm.. delay pools should not make a difference here.. > >Are you sure you have not changed any of the timeouts? Especially the >read_timeout option.
Hello Henrik, We had the read_timeout at 5minutes for a while but set it back to 15minutes (which is 10 minutes longer than the job takes to finish). Also from what I understood of read_timeout from the conf file, if that had been the issue, then a specific timeout error would have been logged whereas we only get: Fri Jun 13 12:20:15 2003 300028 192.168.5.4 TCP_MISS/504 1396 POST http://external.web.server/A_44000.cfm/QueryFacility.xlc? - in the access.log file and nothing in the cache.log. The timeouts that are uncommented in my conf file are: connect_timeout 60 seconds read_timeout 15 minutes The only other "interesting" things I just saw are: quick_abort_min 8 KB quick_abort_max 8 KB half_closed_clients off Is there anything else I can do to test/check? Monday I am having the user try it again but with a delay_pool that is wide open (i.e. defined but defined as large as the aggregate pipe). Other than that, I am not sure what to try/test next. thanks, Adam
