This is a discussion on Re: v3.2.4 scan times slow - SpamAssassin ; Sorry for replying to my own topic, but I've figured out what's causing it to go so slow. It's the rules in sa-blacklist.current.uri.cf from http://www.sa-blacklist.stearns.org/...current.uri.cf . This ruleset works fine in 3.1, I'm not sure why it doesn't in 3.2, ...
Sorry for replying to my own topic, but I've figured out what's causing
it to go so slow.
It's the rules in sa-blacklist.current.uri.cf from
This ruleset works fine in 3.1, I'm not sure why it doesn't in 3.2, any
Sean Kennedy wrote:
> I recently upgraded from v3.1.9 to v3.2.4 and I've noticed a substantial
> increase in scan times. The general average scantime with v3.1 was
> about 1.2s and now with v3.2 it's about 2.2s. It's enough of a slow
> down so that my mail queue backs quite easily now.
> So I'm trying to debug SA and figure out whats going on by doing -D
> --lint and I've got a couple questions about some of the output.
> 1) Why am I getting lines like the following and how do I correct it?
>  dbg: rules: SARE_HTML_ALT_WAIT1 merged duplicates:
> SARE_HTML_ALT_WAIT2 SARE_HTML_A_NULL SARE_HTML_BADOPEN
> SARE_HTML_BAD_FG_CLR SARE_HTML_COLOR_NWHT3 SARE_HTML_FONT_INVIS2
> SARE_HTML_FSIZE_1ALL SARE_HTML_GIF_DIM SARE_HTML_H2_CLK
> SARE_HTML_HTML_AFTER SARE_HTML_INV_TAGA SARE_HTML_JSCRIPT_ENC
> SARE_HTML_JVS_HREF SARE_HTML_MANY_BR10 SARE_HTML_NO_BODY
> SARE_HTML_NO_HTML1 SARE_HTML_P_JUSTIFY SARE_HTML_URI_2SLASH
> SARE_HTML_URI_AXEL SARE_HTML_URI_BADQRY SARE_HTML_URI_BUG
> SARE_HTML_URI_FORMPHP SARE_HTML_URI_HREF SARE_HTML_URI_MANYP2
> SARE_HTML_URI_MANYP3 SARE_HTML_URI_NUMPHP3 SARE_HTML_URI_OBFU4
> SARE_HTML_URI_OBFU4a SARE_HTML_URI_OPTPHP SARE_HTML_URI_REFID
> SARE_HTML_URI_RID SARE_HTML_URI_RM SARE_HTML_USL_MULT
> 2) It hangs for like 30 seconds on the following line, what exactly is
> it doing and is it necessary?
>  dbg: rules: running uri tests; score so far=1.5
> It takes about 5s to run -D --lint on my boxes running v3.1, but about
> 50s to 1m10s using v3.2 (same hardware on all boxes).
> Any info is greatly appreciated!