Nigel, I did a couple of Google searches on "policyd block expiration", "policyd ratelimit expiration", "policyd block email", it seems the closest thread to my question was this:
http://www.mail-archive.com/[email protected]/msg00857.html Is that the thread you have in your mind? I did some tests and monitored mysql data changes, it seems that the v1 implementation calculates the block period by adding the timestamp of the first message and the time limit set in the configuration. In other words, if you trigger the threshold early on, you get longer block time, if close to the timer starting time + timelimit, then you get shorter block time. I am tempted to change the behavior by monitoring the policyd log for abuse flag and call an external program to set the the timer (_date field in throttle table) to the time the abuse occurred. Will that give me what I want? If that's reasonable workaround, can that be a feature request for future release? Apologize if these questions have been discussed before! Xueshan ----- Original Message ----- > On 09/01/10 16:09, Xueshan Feng wrote: > > I am still running v1. > > > > I wonder how the block expiration time is calculated. > > > > Suppose my message quota is 100, my time limit is 2 hours. > > > > I started sending email at 8am. If I reached my quota at 9:50am, > > will I be blocked until 10:00am or 11:50am? > > > > It seems that the expiration time is calculated by _date + > > _time_limit field in policyd database throttle table. The _date is > > the time of the first message. > > > > Then I would only be blocked for 10 minutes instead of two hours > > starting from when the abuse > > occured? So in theory I can time it so that I'd never be blocked if > > I hit the limit near the time limit mark? > > > > Or did I miss understand how does it work? I'd like to be able to > > block based on the time abuse started, not when the timer started. > > > > Xueshan > > > > As far as I know there is a detailed post to the mailing list with how > its done, check the archives, I you cannot find it I'll search for it. > > Starting to think we need a FAQ on the site to address some topics > like > this :) > > -- > Regards > Nigel Kukard > > > > _______________________________________________ > Users mailing list > [email protected] > http://lists.policyd.org/mailman/listinfo/users -- Xueshan Feng <[email protected]> Technical Lead, IT Services, Stanford University _______________________________________________ Users mailing list [email protected] http://lists.policyd.org/mailman/listinfo/users
