Re: [Denyhosts-user] Why are there entries in hosts-restricted when no restricted-usernames

[Jesse S. <jes...@ya...>]

I think that problem was addressed in the DenyHost fork of the project
last year, though I could be mistaken. Perhaps you could try v3.0 from
the http://denyhost.sf.net website and report back on whether you still
encounter the problem?

Jesse

On 30/09/15 09:05 AM, Daniel Sutcliffe wrote:
> Firstly I'm a new subscriber so I must say I have been using DenyHosts
> for quite some time and have found it to be a really useful and simple
> to configure piece of software.
> 
> I realize it hasn't been updated in quite some time but this probably
> just means it does exactly what is needed of it with little fuss or
> problems.
> 
> Recently however I noticed a problem which caused me to look closer at
> what was going on - the initial problem is probably a config thing
> that I'll mention in a later email. The other issue I discovered that
> I mention in the subject line already seems to be in the bug database
> twice:
>   https://sourceforge.net/p/denyhosts/bugs/32/
>   https://sourceforge.net/p/denyhosts/bugs/43/
> Although, it seems the root of the problem has not been understood.
> 
> I also came across the problem here: http://serverfault.com/q/647153/310481
> 
> What appears to be happening is that during LoginAttempt.add() in
> DenyHosts/loginattempt.py if any of the AGE_RESET_XXX configs are set
> and there is a valid host then .agg_count() gets called for the host
> in the specific __abusive_hosts_XXX set regardless of whether that
> host exists in that list or not. The result being a zero count entry
> for that host in a list it possibly does not really belong in.
> 
> The downside to this seems only to be large hosts files and time
> wasted. Maybe there is a reason for this action, if so I'd love to
> have it explained before I attempt to fix this inefficiency and
> horribly break something else in a perfectly stable bit of code ;)
> 
> Cheers
> /dan
>