On 8/20/2012 8:07 AM, Axb wrote:
I'm questioning the point of running the weekend net masschecks.
In my case, where my corpus isn't huge tho larger than others being
massschecked, took nearly 24 hours to run.
status: starting scan stage now: 2012-08-18 11:30:11
status: completed run stage now: 2012-08-19 07:10:27
I run local mirrors for spamhaus/surbl/uribl list zones so NS lag on
those is non existing.
I assume Pyzor and a few of the smaller WL/BLs have either rating or
maybe even not so good connectivity over the pond.
I know Razor can hog processing but their pipes are huge and have
never been a problem in production.
Afaik, BL/WLs all have static scores. What's the point of hammering
them week after week ?
I'll be bailing out of weekly checks as I can't any longer afford to
block 8 cores for nearly a whole day.
Reducing the spam corpus size would affect daily results so that's not
a smart option.
FYI that I don't think the system is designed to have one without the
other. They intertwine.
I also keep meaning to compare the minimal masscheck script to the original.
But it sounds like we need to try and figure out where the delay is on
the masscheck net. It could be legit but if not, it's adding a lot of time.
Regards,
KAM
