Stork 0.14.0, December 9th, 2020, Release Notes
Welcome to the Stork 0.14.0 release. This is a development release of the Stork dashboard.
The changes introduced in this version are:
-
Large configuration handling: Stork 0.14 is now much better suited for handling large configurations, e.g. with several thousand subnets. The process of adding new subnets has been greatly streamlined and now takes much less time. Subnets are now indexed (#421 (closed)). Another problem solved was that the Stork server triggered the periodic configuration pull too early. If the initial update took a long time, the first periodic renewal kicked off before the initial update was complete. This race condition is fixed (#446 (closed)). Furthermore, Stork's periodic configuration pull is now smarter. It calculates a hash and is able to detect a configuration change, bypassing unnecessary processing in the very likely case of the configuration not changing (#460 (closed)).
-
Reconfigure events: Stork now records the reconfiguration events detected in Kea. The information about the user who authorized the change is not logged. This information may be added in a future Stork release (#353 (closed), #460 (closed)).
-
Documentation updates: Grafana and Prometheus integration is now documented properly, with examples (#451 (closed)). A note has been added about missing lease statistics and how to fix the problem, by loading stat_cmds hook for Kea (#433 (closed)).
-
Events filtering: Earlier Stork versions showed all events on various pages. This was confusing, as events related to machine A were also presented when viewing machine B. Right now the widgets show only relevant events (#429 (closed)).
-
Tools update: The Go compiler has been updated to 1.15. This is not a user visible change. Nevertheless, it's worth mentioning that the Stork team keeps up the pace with the latest developments in the industry. We're also working on the update to Angular 10 (#457 (closed)).
-
Test improvements: The Stork team takes testing and other QA practices very seriously. We've spent a significant amount of time improving system tests (#434 (closed)), stabilizing Selenium tests for our UI interface (#453 (closed)) and made other UI test improvements (#435 (closed)).
Please see this link for known issues: https://gitlab.isc.org/isc-projects/stork/-/wikis/Known-issues.
Release Model
Stork has monthly development releases, with some exceptions.
We encourage users to test the development releases and report back their findings on the stork-users mailing list, available at https://lists.isc.org/mailman/listinfo/stork-users.
This text references issue numbers. For more details, visit the Stork GitLab page at https://gitlab.isc.org/isc-projects/stork/issues.
License
Stork is released under the Mozilla Public License, version 2.0.
https://www.mozilla.org/en-US/MPL/2.0
Download
The easiest way to install the software is to use native deb or RPM packages. They can be downloaded from:
https://cloudsmith.io/~isc/repos/stork/
The Stork source and PGP signature for this release may be downloaded from:
https://downloads.isc.org/isc/stork
The signature was generated with the ISC code-signing key which is available at:
ISC provides documentation in the Stork Administrator Reference Manual. It is available on ReadTheDocs.io at https://stork.readthedocs.io/en/latest/, and in source form in the docs/ directory.
We ask users of this software to please let us know how it worked for you and what operating system you tested on. Feel free to share your feedback on the stork-users mailing list (https://lists.isc.org/mailman/listinfo/stork-users). We would also like to hear whether the documentation is adequate and accurate. Please open tickets in the Stork GitLab project for bugs, documentation omissions and errors, and enhancement requests. We want to hear from you even if everything worked.
Support
Professional support for Stork will become available once it reaches the 1.0 milestone. Existing ISC customers that consider themselves very early adopters may get involved in the development process, including roadmap, features planning, and early testing, but the software maturity level does not constitute a typical professional service before the 1.0 milestone.
Free best-effort support is provided by our user community via a mailing list. Information on all public email lists is available at https://www.isc.org/mailinglists/. If you have any comments or questions about working with Stork, please share them to the stork-users list (https://lists.isc.org/mailman/listinfo/stork-users). Bugs and feature requests may be submitted via GitLab at https://gitlab.isc.org/isc-projects/stork/issues.
Changes
The following summarizes changes and important upgrades since the Stork 0.13.0 release.
* 121 [func] marcin
Events received over SSE and presented on various Stork pages are
now filtered and only the events appropriate for the current view
are shown. Prior to this change all events were always shown.
(Gitlab #429)
* 120 [func] marcin
When Stork server pulls updated Kea configurations it detects those
configurations that did not change since last update using a fast
hashing algorithm. In case when there was no configuration change
for a daemon, Stork skips processing subnets and/or hosts within
this configuration. This improves efficiency of the configuration
pull and update. In addition, when configuration change is detected,
an event is displayed informing about such change in the web UI.
(Gitlab #460)
* 119 [doc] tomek
Prometheus and Grafana integration is now documented. Also, updated
requirements section pointing out that stat_cmds hook is needed for
Stork to show Kea statistics correctly.
(Gitlab #433, #451)
* 118 [bug] marcin
Prevent an issue whereby Stork server would attempt to fetch updated
machine state while the request to add this machine is still being
processed. This used to cause data conflict errors in the logs and
network congestion.
(Gitlab #446)
* 117 [build] marcin
Upgraded Go from 1.13.5 to 1.15.5 and golangcilint from 1.21.0 to
1.33.0.
(Gitlab #457)
* 116 [perf] marcin
Improved performance of connecting large Kea installation with many
subnets to Stork. Adding subnets to the database is now much more
efficient as it avoids extensive subnet lookups. Instead it uses
indexing techniques.
(Gitlab #421)
Thank you again to everyone who assisted us in making this release possible.
We look forward to receiving your feedback.