Hi Mario,
there was a new version of Argus rolled out this weekend (1.4), if your
system (nightly yum update) or one of your team updated the Argus-host,
the Argus-service is shut down afterwards. We (Argus PT) recommend to
rerun Yaim after any update.
Please check your messages if the update was performed shortly prior to
the Argus-shutdown. Note that Argus-logs are in Zulu-time, while your
messages might be in any other timezone.
cheers,
Joël
On 26.09.11 09:31, Mario Kadastik wrote:
> Hi,
>
> I'm baffled by ARGUS, that suddenly decided to shut down in the night between Sat and Sun. The logs look as if someone had logged in and shut it down:
>
> pap:
> 2011-09-25 02:22:09.066Z - INFO [log] - Shutdown request received from 127.0.0.1.
> 2011-09-25 02:22:09.072Z - INFO [PAPService] - Shutting down PAP service...
> 2011-09-25 02:22:09.072Z - INFO [PAPService] - Shutting down distribution module...
> 2011-09-25 02:22:09.079Z - INFO [PAPService] - Shutting down authorization module...
> 2011-09-25 02:22:09.079Z - INFO [PAPContextListener] - PAP service shutdown complete!
>
> pdp:
> 2011-09-25 02:22:05.089Z - INFO [ShutdownCommand] - Service shutting down
> 2011-09-25 02:22:10.195Z - INFO [JettyShutdownTask] - Server stopped.
> 2011-09-25 02:22:10.299Z - INFO [JettyShutdownTask] - Server stopped.
>
> pepd:
> 2011-09-25 02:22:01.874Z - INFO [ShutdownCommand] - Service shutting down
> 2011-09-25 02:22:06.885Z - INFO [JettyShutdownTask] - Server stopped.
> 2011-09-25 02:22:06.890Z - INFO [JettyShutdownTask] - Server stopped.
>
> prior to that all logs look normal authorization ones. Last log doesn't show anyone in the node at the time:
>
> root@mercury pepd]# last
> root pts/0 192.168.1.248 Mon Sep 26 10:22 still logged in
> root pts/0 192.168.1.248 Mon Sep 5 16:31 - 10:28 (17:56)
> root pts/0 192.168.1.248 Fri Sep 2 11:09 - 12:14 (01:05)
>
> Also system log doesn't seem to show anything around that time. So I'm baffled somewhat. I've started the services again, but such a random stopping killed all of our CE's as it's a single point of authentication and therefore a single point of failure... That brings up the question on wether there is a way to configure a HA mode of ARGUS on multiple nodes and have all CE's request auth data from the pool of them?
>
> Mario Kadastik, PhD
> Researcher
>
> ---
> "Physics is like sex, sure it may have practical reasons, but that's not why we do it"
> -- Richard P. Feynman
--
SWITCH
Serving Swiss Universities
--------------------------
Joël Casutt, Informatics Engineer, Middleware
Werdstrasse 2, P.O. Box, 8021 Zurich, Switzerland
email: [log in to unmask] phone: +41 44 268 1573
|