Details
Assignee
fookerfookerReporter
JianYetJianYetHB Grooming Date
Aug 15, 2023HB Backlog Status
Refined BacklogComponents
Sprint
NoneFix versions
Affects versions
Priority
Medium
Details
Details
Assignee
fooker
fookerReporter
JianYet
JianYetHB Grooming Date
Aug 15, 2023
HB Backlog Status
Refined Backlog
Components
Sprint
None
Fix versions
Affects versions
Priority
PagerDuty
PagerDuty
PagerDuty
Created August 10, 2023 at 7:26 PM
Updated October 4, 2023 at 2:08 PM
I have tried adding environment variables like below into the
docker-compose.yaml
file.environment: - OPENNMS_TRAPD_ADDRESS=127.0.0.1 - OPENNMS_TRAPD_PORT=10162 - OPENNMS_TRAPD_NEWSUSPECTONTRAP=true - OPENNMS_TRAPD_INCLUDERAWMESSAGE=true - OPENNMS_TRAPD_THREADS=1 - OPENNMS_TRAPD_QUEUESIZE=12000 - OPENNMS_TRAPD_BATCHSIZE=1200 - OPENNMS_TRAPD_BATCHINTERVAL=300
The result in
/usr/share/opennms/etc/trapd-configuration.xml
is only the address and the port are changed. The others are untouched.I also tried variable mapping by way of confd. For example, in
horizon-config.yaml
--- postgres: host: database port: 5432 user: postgres password: postgres opennms: dbname: opennms dbuser: opennms dbpass: opennms timeseries: strategy: rrd rrd: storebyforeignsource: true strategyclass: org.opennms.netmgt.rrd.rrdtool.MultithreadedJniRrdStrategy library: jrrd2: /usr/lib64/libjrrd2.so trapd: address: 127.0.0.1 port: 10162 newsuspectontrap: true includerawmessage: true threads: 1 queuesize: 12000 batchsize: 1200 batchinterval: 300
The result is only the address and the port are changed. The others are still the default values.
Container logs attached.
What’s the point of overriding trapd-configuration.xml with confd? This has caused issue in persisting the changes made in the container directly as with every boot the trapd configuration is overridden by confd. Also, it’s introducing extra complications as each snmp v3 device that’s sending traps to opennms needs to be added there. How do we control this from confd? In helm chart, the only way to do this by creating a overlay file load it as configmap. If the file gets over the 1MB file size limit for helm, then it would break. We need to incorporate this into the change proposed by https://opennms.atlassian.net/browse/NMS-15822