Hello everyone,

It seems we pushed the 3.2.0 release out the door a little too fast.  I've had 
a disturbing bug report today that indicates total breakage of NAV's 
eventEngine.  

NAV 3.2.0 introduces the new messages and maintenance subsystems, discarding 
the old combined messages/maintenance interface.  In doing so, the old 
message/maintenance database tables are dropped.  Apparently, though, the 
eventEngine erroneously uses the maintenance schedule tables to retrieve the 
maintenance status of devices, instead of the alert history tables, which are 
authoritative for these kinds of device states.

None of the active NAV developers were aware of this, and the result is that 
the eventEngine tries to use database tables that don't exist, and therefore 
seems to fail to retrieve any meaningful data about devices.


/I would therefore advise everyone to delay upgrading to or installing NAV 3.2, 
while we are working on a bugfix for this problem./


---

For those really persistent souls among you, you can still have the eventEngine 
working decently if you modify the 3.2.0.sql script and comment out the 
following SQL statements before running it:

 DROP VIEW maintenance_view; 
 DROP TABLE maintenance;
 DROP TABLE emotd_related;
 DROP TABLE emotd;

If the tables have already been dropped, or you have made a clean installation 
of NAV 3.2.0, you can install the missing tables in the manage database by 
locating their definitions in 
http://svn.itea.ntnu.no/repos/nav/navme/tags/3.1.1/doc/sql/manage.sql .

Please be aware that even with this "patch" in place, eventEngine will be 
unable to hold back alerts for devices on maintenance, since it is looking in 
the wrong place for maintenance status.

-- 
mvh
Morten Brekkevold
UNINETT
_______________________________________________
nav-users mailing list
[email protected]
http://mailman.itea.ntnu.no/mailman/listinfo/nav-users

Reply via email to