In the last week we released new Meridian and Horizon versions, and worked on a bunch of bugs, Nephron flow processing, moving BMP support in-core, node geohashes, the Prometheus collector, Minion configuration, ToS/QoS aggregation, and much more.
Github Project Updates
Internals, APIs, and Documentation
- Dustin and Christian continued to work on improvements to Nephron flow processing and clock skew handling
- Bonnie and Ronny did more work on the migration of the admin guide to Antora
- Chandra worked on a fix for the elasticsearch forwarder when using custom prefixes
- Marcel made confd templates for the Mattermost notification strategy
- Dino fixed an issue with non-integer counters in the Prometheus collector
- Chandra continued to work on moving our OpenBMP integration to the OpenNMS core
- Jesse added support for geohashes to the node metadata
- I fixed a serialization bug I introduced when fixing timezones in Grafana reports
- Dustin worked on some flink test infrastructure
- Pierre fixed an issue with custom server certificates in the Minion
- Dustin made a fix to initializing telemetryd when multiple metrics have the same name
- Stefan added a CLI option to
opennms to start paused waiting for a debugger
- Chandra worked on an SNMP poller exception bug
Web, ReST, UI, and Helm
- Stefan worked on updates to the flow ReST API and OpenNMS.js to handle raw ToS queries
- Patrick continued his work on handling topology map issues with applications
- Bonnie and Chandra worked on clarifying the Provisioning UI scan prompt
Thanks to the following contributors for committing changes since last OOH:
- Benjamin Reed
- Bonnie Robinson
- Chandra Gorantla
- Christian Pape
- Dino Yancey
- Dustin Frisch
- Jesse White
- Marcel Fuhrmann
- Matthew Brooks
- Patrick Schweizer
- Pierre Bouffard
- Ronny Trommer
- Stefan Wachter
In November we released updates to all supported Meridians as well as Horizon 27.
You might have noticed that Horizon 27 and Meridian 2020 jump by 2 releases in this announcement compared to last month. This brings us to the report serialization regression.
Report Serialization Regression
A fix for timezone handling in Grafana reports that went into Horizon 27.0.1 and Meridian 2020.1.2 turned out to trigger a bug in loading existing reports, because of changes to the underlying Java objects that get stored in the database.
The regression was released, so we did an off-schedule release of updated Horizon 27 and Meridian 2020 packages to minimize the impact.
If you created new scheduled reports with Horizon 27.0.1 or Meridian 2020.1.2, you will need to recreate them.
On Tuesday, we released Horizon 27.0.1, followed by Horizon 27.0.2 yesterday.
- Horizon 27.0.1 _(Pan-Galactic Gargle Blaster)_ contains a number of bugfixes including a critical CVE fix for Jetty, as well as a few other smaller changes and improvements.
- Horizon 27.0.2 _(Deep Thought)_ adds to that a fix for the report serialization bug, a fix for JCIFS support, and an update to node metadata to support geohashes.
Like Horizon 27, Meridian 2020 had two releases: 2020.1.2 on Tuesday, and 2020.1.3, yesterday.
Meridians 2019.1.14 and 2018.1.24
As is common in maintenance releases, the older Meridians got just a couple of bug fixes -- the Jetty CVE fix, and a fix to RRD file creation in certain cases when provisioning new nodes.
For more details, see the release notes:
The next OpenNMS release day is January 5th, 2020.
The current expectation is that there will be a Horizon 27.0.3 release with bug fixes.
Next Horizon: 28 (Q1 2021)
The next major Horizon release will be Horizon 28.
We're still early in the development cycle for it, but at a high level it is expected to contained our finished work moving OpenBMP's functionality into OpenNMS, as well as enhancements to handle ToS/QoS in flows.
Next Meridian: 2021 (Q2 2021)
With the recent release of Meridian 2020, plans are still tentative.
However, the current plan is that Meridian 2021 will be based on Horizon 28.
We'll know more once development plans start to firm up.
Note that this is just based on current plans; dates, features, and releases can change or slip depending on how development goes.
The statements contained herein may contain certain forward-looking statements relating to The OpenNMS Group that are based on the beliefs of the Group’s management as well as assumptions made by and information currently available to the Group’s management. These forward-looking statements are, by their nature, subject to significant risks and uncertainties.
...We apologize for the excessive disclaimers. Those responsible have been sacked.
Mynd you, møøse bites Kan be pretti nasti...
We apologise again for the fault in the disclaimers. Those responsible for sacking the people who have just been sacked have been sacked.
Until Next Week…
If there’s anything you’d like me to talk about in a future OOH, or you just have a comment or criticism you’d like to share, don’t hesitate to say hi.
Resolved Issues Since Last OOH
- NMS-12700: Add support for custom.system.properties on Sentinel
- NMS-12949: Persist basic BMP messages in OpenNMS
- NMS-12952: Handle stats for BMP
- NMS-12982: Create confd templates to handle Slack properties
- NMS-12989: Change installation guide to run as non-root user
- NMS-13004: Create confd templates to handle Mattermost properties
- NMS-13007: Prometheus Collector attempting to persist non-integer values to counters
- NMS-13017: When using a custom prefix, the Elasticsearch Forwarder for events and situation-feedback creates a wrong template.
- NMS-13034: OpenNMS fails to start, when more than one active listener is referencing the same parser
- NMS-13035: Bouncycastle JAR version 1.67 breaks CIFS Monitor
- NMS-13036: Add "geohash" support to the meta-data DSL
- NMS-13037: report timezone changes break reading pre-existing reports from Quartz