On 2 Jun 2008 at 19:41, Juan Miscaro wrote:
> 2008/6/2 Jason Dixon :
[snip]
> > In my case, I use Courier's authdaemond with MySQL, and Cyrus SASL's
I cannot (and do not) speak for Jason, but as I have discovered for
myself, PAM epitomizes the general linux approach: "if it is critical
system component, don't simplify -- over-design!" yes, it is highly
configurable and very flexible/powerful/blah/blah... and I even came
across one instance where some of this power and flexibility was being
put to good use. BUT in the vast majority of cases it is not warranted.
Do you _really_ need a multi-layered gizmo with multitude of loadable
modules and innumerable configuration files just to do basic
authentication? How safe do you really feel when your system security
depends on a beast with so many points of failure? (here is a small
example: In all linux distros I've tried, the sshd_config file contains
the directives: 'PasswordAuthentication no' and 'UsePAM yes'. So guess
what happens when you try to ssh in with just a password? want to be
the first to write a HowTo for configuring PAM to require host-keys but
only for ssh? [I'm not being facetious -- I admin some systems that are
linux and cannot be converted to OpenBSD.])
-Jacob
> /juan
| Greg KH | [GIT PATCH] driver core patches against 2.6.24 |
| Tarkan Erimer | Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3 |
| Amit K. Arora | [RFC] Heads up on sys_fallocate() |
| Chuck Ebbert | Why do so many machines need "noapic"? |
git: | |
| Jarek Poplawski | [PATCH] pkt_sched: Destroy gen estimators under rtnl_lock(). |
| David Miller | [GIT]: Networking |
| Gerrit Renker | [PATCH 27/37] dccp: Integration of dynamic feature activation - part 2 (server side) |
| Natalie Protasevich | [BUG] New Kernel Bugs |
