fwiw, removing 70_sandbox.cf hasn't had too much effect on that, so it's a promoted rule in 72_active.cf.
--j. On Tue, Feb 22, 2011 at 11:07, Justin Mason <[email protected]> wrote: > Feb 22 11:07:10.729 [32583] dbg: timing: total 6020 ms - parse: 3 > (0.0%), extract_message_metadata: 43 (0.7%), get_uri_detail_list: 81 > (1.4%), tests_pri_-1000: 56 (0.9%), tests_pri_-950: 1.72 (0.0%), > tests_pri_-900: 38 (0.6%), tests_pri_-400: 1.68 (0.0%), tests_pri_0: > 5664 (94.1%), tests_pri_500: 94 (1.6%) > > so it's tests at pri 0. doesn't really narrow it down too much ;) > > --j. > > On Tue, Feb 22, 2011 at 00:16, Mark Martinec <[email protected]> wrote: >> Justin wrote: >>> unfortunately I didn't save them; just "svn up"d and restarted the >>> process to see if it fixed it (it didn't, but the messages are no >>> more). >>> some strace'ing isn't showing network traffic, so it could just be >>> poor local rule performance with the current trunk ruleset. >>> Warren: Ubuntu 10.04, default /usr/bin/perl. >> >> Try: mass-check --debug=timing >> to see if anything unusual stands out. >> >> (after first updating svn - I just added a call to timer_reset >> to mass-check before processing each message, like the spamd does; >> previously the timing report would show cumulative times, which are >> not that interesting) >> >> Mark >> >
