On Tue, 2013-01-15 at 15:23 +0100, Jakub Hrozek wrote:
On Tue, Jan 15, 2013 at 08:27:32AM -0500, Simo Sorce wrote:
> On Tue, 2013-01-15 at 09:02 +0100, Jakub Hrozek wrote:
> >
https://fedorahosted.org/sssd/ticket/1757
> >
> > When the user entry was missing completely after initgroups, we would
> > never invalidate the user entry from cache. This led to dangling cache
> > entried in memory cache if the user was removed from the server while
> > still being in memory cache.
>
> 0001: Why are we unconditionally deleting the user from the memory cache
> even if just groups changed ?
>
I was trying to keep the flow as simple as possible because I think that
user deletes or even group modifications updates on the server are so rare
that if we loose the memcache, it's not a big deal.
But I moved the user update separately in this version.
> 0002: Please add a comment in be_initgroups_prereq() that the first id
> it he user's primary gid and 'might' be duplicated later ?
Done
> Also why do we skip verifying the primary gid ?
>
Ah, not sure, I think I was just copying the previous flow that skipped
the primary gid too. The attached version checks them all.
Thank you for the review.
Looks good to me, ACK!
Simo.
--
Simo Sorce * Red Hat, Inc * New York