ReinitializePrimarySNMPInterface event causes SNMP interfaces manually marked as "Collect" to revert to "No collect"

This should have been resolved by the fix tot

Based on the similarities in the overall problem, this might very well be related to , but I've opened a separate bug since it's not an exact duplicate.

To confirm that I wasn't going crazy, I set up a cron job to capture the state of the snmpinterface table for a given node every 5 minutes. The query was:

select snmpifindex,snmpifname,snmpcollect from snmpinterface where nodeid=30 order by snmpifname;

At 0700, all of the interfaces I marked for collection were properly set in the database (snmpcollect='C').

At 0705, all of the interfaces I marked had been reset to snmpcollect='N'.

Combing the events database for this node, there were only two events between 0600 and 0800 for this node: nodeScanCompleted and reinitializePrimarySnmpInterface. See below:

opennms=# select * from events where nodeid=30 and eventcreatetime > '2011-09-30 06:00:00' and eventcreatetime < '2011-09-30 08:00:00';
eventid | eventuei | nodeid | eventtime | eventhost | eventsource | ipaddr | eventdpname | eventsnmphost | serviceid | eventsnmp |
eventparms | eventcreatetime |
eventdescr | eventloggroup |
eventlogmsg | eventseverity | eventpathoutage | eventcorrelation | eventsuppressedcount | eventoperinstruct | eventautoaction | eventoperaction | eventoperacti
onmenutext | eventnotification | eventtticket | eventtticketstate | eventforward | eventmouseovertext | eventlog | eventdisplay | eventackuser | eventacktime | alarmid | ifindex
--------+--------------------------------------------------------------------------------------------------------------+-------------------------------------------------------------+--------------------- --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- ---------------------------------------------------------------------------------------------------+----------------+--------------------+---------------------------------------------------+------------- -------------------------------------------------------------------------------------------------------------------------------------------------------------+-------- 624906 | uei.opennms.org/internal/provisiond/nodeScanCompleted | 30 | 2011-09-30 07:02:07-05 | | Provisiond | | undefined | | | | foreignSource=Networ
k - MASKED(string,text);foreignId=1316746428188(string,text) | 2011-09-30 07:02:09.756-05 | A message from the Provisiond NodeScan lifecycle that a NodeScan has completed: <p>The Node with Id: 30; ForeignSource
: Network - MASKED; ForeignId:1316746428188 has completed.</p> Typically the result of a request of an import request or a scheduled/user forced rescan. | | <p>The Node with Id: 30; ForeignSource
: Network - MASKED; ForeignId:1316746428188 has completed.</p> | 3 | | | | | | |

624905 | uei.opennms.org/nodes/reinitializePrimarySnmpInterface | 30 | 2011-09-30 07:02:07-05 | | Provisiond | MASKED.7 | undefined | | | |

(2 rows)

The node scan and reinitializePrimarySNMPInterface are triggered every 24 hours by the default provisiond node rescan. The device in question here is a Catalyst 6509 with a couple hundred interfaces.