On Wed, Jun 14, 2006 at 11:45:42AM +0300, Nikolay Kichukov wrote:
> Hi all,
> I was follwoing the thread and if possible I would like someone to
> elaborate on a few more points.
>
> Which is the version of the utils in developemnt that can resolve that
> matter if dentry limits is applied? When was this fix applied?
it's no fix, it is an additional limit, the typical
linux system (mainline) is still open to such attacks
which can be executed by an arbitrary user ...
the feature was added shortly after vs2.1.1-rc18, so
it is in recent devel kernels, the tools do not need
to be modified to allow to set this limit
> How does the dentry limit work
very similar to the other limits, the dentries are
accounted and compared to a hard limit, which, when
hit, will return an error ...
> and how is it configured on the host side?
again, very similar to the other limits ...
> As herbert said "(e.g. three strikes and you're out)" how to
> configure that for example?
this one was referring to some 'hosting' policy,
and in this case it means something like:
if you 'accidentially' hog the system with something,
you'll get a warning from the hoster ...
if you 'accidentially' do it again, you'll get
kicked ...
HTH,
Herbert
> Thanks and regards,
> -Nikolay Kichukov
>
> On Wed, 2006-06-14 at 01:10 +0200, Herbert Poetzl wrote:
> > On Tue, Jun 13, 2006 at 09:16:48PM +1000, Russell Kliese wrote:
> > > I was just reading an article on kernel trap that raised some issues about
> > > linux-vservers: http://kerneltrap.org/node/6492 .
> > >
> > > In particular, the following denial of service attack from within a
> > > vserver seemed worrying because of it's simplicity.
> > >
> > > > run a program doing `mkdir("aaa"); chdir("aaa");' in a loop inside
> > > > Linux-VServer VPS and see what happens.
> > >
> > > Is there work being done to prevent such DoS attacks? I have to admit
> > > that I haven't tested this yet and it might just be FUD, but I thought
> > > that I might as well ask.
> >
> > dentry limits in devel prevent this specific DoS
> > attack, but IMHO there will always be some way to
> > 'hurt' a system which is based on resource sharing,
> > so the best approach is to apply some policy there
> > (e.g. three strikes and you're out)
> >
> > best,
> > Herbert
> >
> > > Russell
> > >
> > > _______________________________________________
> > > Vserver mailing list
> > > Vserver@list.linux-vserver.org
> > > http://list.linux-vserver.org/mailman/listinfo/vserver
> > _______________________________________________
> > Vserver mailing list
> > Vserver@list.linux-vserver.org
> > http://list.linux-vserver.org/mailman/listinfo/vserver
> --
> ?????? ??? ????????, ??? ?????.
> ?? ?????? ??? ?????, ?? ?????? ??? ????????...
> -????? ?????
>
> _______________________________________________
> Vserver mailing list
> Vserver@list.linux-vserver.org
> http://list.linux-vserver.org/mailman/listinfo/vserver
_______________________________________________
Vserver mailing list
Vserver@list.linux-vserver.org
http://list.linux-vserver.org/mailman/listinfo/vserver
Received on Wed Jun 14 12:58:31 2006