On Wed, Dec 09, 2015 at 03:19:54PM +0100, Lukas Slebodnik wrote:
On (09/12/15 14:29), Jakub Hrozek wrote:
>On Wed, Dec 09, 2015 at 02:15:32PM +0100, Lukas Slebodnik wrote:
>> On (09/12/15 14:11), Lukas Slebodnik wrote:
>> >On (09/12/15 13:21), Sumit Bose wrote:
>> >>On Wed, Dec 09, 2015 at 12:24:38PM +0100, Jakub Hrozek wrote:
>> >>> Hi,
>> >>>
>> >>> see the attached trivial patch. It was verified by a user on IRC
that
>> >>> decreasing the timeout helped quite a bit when a laptop with AD
user
>> >>> went offline and the AD user tried to unlock the screen.
>> >>
>> >>ACK (waiting for CI to finish ...)
>> >>
>> >http://sssd-ci.duckdns.org/logs/job/34/28/summary.html
>> >
>> >rawhide fail is unrelated.
>> >
>> >new libtool is already in koji
>>
>http://sssd-ci.duckdns.org/logs/job/34/28/fedora_rawhide/ci-build-debug/ci-mock-build.log
>> >
>> master:
>> * a687f4473bf305bc2ccb075cd93154c9d661b638
>>
>> BTW should it be fixed also in stable branch?
>> At least this patch was trivial. if we do not push other patches
>> for ticket #2866
>
>I would only push this patch into the stable branch if the patch for
>https://fedorahosted.org/sssd/ticket/2866 is pushed.
I think we should use correct search timeouts and patch is safe.
sssd-1-13:
* e4231dff16f7b6ff9ef68718a67ea01cec62bab6
sssd-1-12:
* fe022b2c87c82693d24f6abc90e36ad8c11c34ca
What I tried to say is that unless you also push the patch for #2866,
this use-case (disconnected AD trust user on a laptop) would fail even
sooner. Sure, this patch is not harmful, but the way I found out about
it was that I gave the reporter test builds with patch for #2866 and he
reported back that while he no longer gets a System Error on login, the
login takes more than a minute...