You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@spamassassin.apache.org by John Hardin <jh...@impsec.org> on 2010/03/25 23:06:37 UTC
ruleqa: overlaps broken again, SVN URIs still broken
Justin:
The ruleqa overlap analysis isn't working again, and the rule detail ->
SVN URIs are still pointing at the old rules/trunk sandbox path.
--
John Hardin KA7OHZ http://www.impsec.org/~jhardin/
jhardin@impsec.org FALaholic #11174 pgpk -a jhardin@impsec.org
key: 0xB8732E79 -- 2D8C 34F4 6411 F507 136C AF76 D822 E6E6 B873 2E79
-----------------------------------------------------------------------
Phobias should not be the basis for laws.
-----------------------------------------------------------------------
167 days since President Obama won the Nobel "Not George W. Bush" prize
Re: ruleqa: overlaps broken again, SVN URIs still broken
Posted by John Hardin <jh...@impsec.org>.
On Wed, 31 Mar 2010, Justin Mason wrote:
> that URI issue is now fixed.
Thanks!
> so it appears to be backlogged by 10 days. :(
I was wondering if that was the situation...
--
John Hardin KA7OHZ http://www.impsec.org/~jhardin/
jhardin@impsec.org FALaholic #11174 pgpk -a jhardin@impsec.org
key: 0xB8732E79 -- 2D8C 34F4 6411 F507 136C AF76 D822 E6E6 B873 2E79
-----------------------------------------------------------------------
Ignorance doesn't make stuff not exist. -- Bucky Katt
-----------------------------------------------------------------------
Tomorrow: April Fools' day
Re: ruleqa: overlaps broken again, SVN URIs still broken
Posted by Justin Mason <jm...@jmason.org>.
hey John --
that URI issue is now fixed.
regarding overlap: the log file on spamassassin2 , /home/automc/freqsd/log ,
has this for the most recent OVERLAP file generation:
creating: /export/home/ruleqadb/20100321/r925742-n/OVERLAP.new (OVERLAP)
started Wed Mar 31 12:26:07 2010...
so it appears to be backlogged by 10 days. :( It appears we don't
have the CPU
time to keep up with generating these on spamassassin2 this way.
I've stopped and restarted the freqsd anyway, in an effort to persuade
it to concentrate
on more recent logs. (I have no idea if this will work.)
To fix the issue in the long term we can either:
a. split off work to another host (unlikely to be viable with ASF
infrastructure)
b. be more selective about the logs we generate overlaps from (it's already only
reporting on nightly logs alone, which is the main one we want)
c. change 'report-from-logs' to use more parallelism/background processing
(higher load, defeats the purpose of freqsd's batch processing model, more
complex code)
d. rewrite the overlap-generation code in hit-frequencies to be more efficient
(I've already spent a lot of effort on this, it's FAST perl. C maybe? anyway,
will take dev effort)
e. rewrite 'report-from-logs', 'freqsd' to be more selective about the log sets
it generates OVERLAP files from, get the more recent stuff built first.
thoughts guys?
--j.
On Thu, Mar 25, 2010 at 22:06, John Hardin <jh...@impsec.org> wrote:
> Justin:
>
> The ruleqa overlap analysis isn't working again, and the rule detail -> SVN
> URIs are still pointing at the old rules/trunk sandbox path.
>
> --
> John Hardin KA7OHZ http://www.impsec.org/~jhardin/
> jhardin@impsec.org FALaholic #11174 pgpk -a jhardin@impsec.org
> key: 0xB8732E79 -- 2D8C 34F4 6411 F507 136C AF76 D822 E6E6 B873 2E79
> -----------------------------------------------------------------------
> Phobias should not be the basis for laws.
> -----------------------------------------------------------------------
> 167 days since President Obama won the Nobel "Not George W. Bush" prize
>
>
--
--j.