On Tue, May 31, 2016, at 10:44, Mick wrote: > On Tuesday 31 May 2016 16:30:27 James wrote: > > Here is an interesting read:: > > > > Security brief: CoreOS Linux Alpha remote SSH issue > > May 19, 2016 ยท By Matthew Garrett > > > > <snippets> > > > > Gentoo defaults to ending the PAM configuration with an optional pam_permit. > > > > This meant that failing both pam_unix and pam_sss on CoreOS systems would > > surprisingly result in authentication succeeding, and access being granted. > > > > The operator user was not used by CoreOS, but existed because it exists in > > the Gentoo Portage system from which CoreOS is derived. > > <end/snippets> > > > > Full read [1]. It kinda shows that CoreOS is derived from Gentoo > > and not ChromeOS; at least when time to blame a security lapse elsewhere.... > > > > > > enjoy, > > James > > > > [1] https://coreos.com/blog/ > > Does this mean we need to do anything to improve the security of our > systems?
The interaction seems to be problematic when SSSD is introduced to the mix, as the SSSD documentation seems to assume a fallthrough to pam_permit.so. With a merely empty user password, I cannot trigger this issue. I will have to try with an expired account and a locked account later. the pam_permit.so line was introduced here: https://gitweb.gentoo.org/proj/pambase.git/commit/?id=ac9023eecfe3c13d212c548bb9d5d1b42a4e044b At the very least, it does seem like it ought to be pam_deny.so instead. I wonder if this issue is still even relevant for Kerberos users? https://bugs.gentoo.org/show_bug.cgi?id=333393
