Skip to content

Latest commit

 

History

History
91 lines (64 loc) · 4.14 KB

ChangeLog.md

File metadata and controls

91 lines (64 loc) · 4.14 KB

Splunk Operator for Kubernetes Change Log

0.1.0 Alpha (2020-03-20)

  • This release depends upon changes made concurrently in the Splunk Enterprise container images. You must use the latest splunk/splunk:edge nightly image with it, or alternatively any release version 8.0.3 or later.

  • The API has been updated to v1alpha2, and involves the replacement of the SplunkEnterprise custom resource with 5 new custom resources: Spark, LicenseMaster, Standalone, SearchHeadCluster and IndexerCluster. Please read the revised Custom Resources and Examples documentation for details on all the changes. This is a major update and is not backwards-compatible. You will have to completely remove any older versions, and any resources managed by the operator, before upgrading to this release.

  • Scaling, upgrades and other updates are now more actively managed for the SearchHeadCluster and IndexerCluster resources. This helps protect against data loss and maximizes availability while changes are being made. You can now also use the "kubectl scale" command, and Horizontal Pod Autoscalers with all resources (except LicenseMaster, which always uses a single Pod).

  • A new serviceTemplate spec parameter has been added for all Splunk Enterprise custom resources. This may be used to define a template the operator uses for the creation of (non headless) services.

  • Splunk Enterprise clusters may now be created without having to provide a license file via the licenseURL parameter. When no license is provided, a default trial license will now be used.

  • Annotations and labels from the managed custom resources are now appended to any corresponding Pod and Service objects that the operator creates.

  • A unique pass4SymmKey secret will now be randomly generated, to resolve cluster master warnings about using the default value.

  • Integrated with CircleCI and Coverall for CICD and code coverage, and added a bunch of unit tests to bring coverage up to over 90%.

0.0.6 Alpha (2019-12-12)

  • The operator now injects a podAntiAffinity rule to try to ensure that no two pods of the same type are scheduled on the same host

  • Ingest updates: HEC is now always enabled by default; added S2S port 9997 to indexer pod specs, added splunk-indexer-service, doc updates

  • Fixed bugs and updated docs, YAML deployment files, roles and bindings to better accomodate using a single instance of the Splunk operator to manage SplunkEnterprise stacks across an entire cluster (cluster scope).

  • Fixed deadlock condition between deployer and search heads when apps_location is specified (note this this also requires a patch in splunk-ansible splunk/splunk-ansible#312

  • Fixed issues with randomly generated secrets sometimes containing characters that broken ansible's YAML parsing, causing failures to provision new stacks. All randomly-generated secrets now use only alpha-numeric characters. Exception is hec_token which now uses a UUID like format.

  • DFS spark worker pool now uses a Deployment instead of StatefulSet

  • Added annotations to get istio to not intercept certain ports

  • Removed unused port 60000 from splunk-operator deployment spec

  • Pod template change now log the name of the thing that was updated

0.0.5 Alpha (2019-10-31)

  • Added port 8088 to expose on indexers, and only exposting DFC ports on search heads
  • Bug fix: The spark-master deployment was always updated during reconciliation

0.0.4 Alpha (2019-10-22)

  • Updates to SplunkEnterprise objects are now handled, enabling deployments to be upgraded and modified after creation
  • Added liveness and readiness probes to all Splunk Enterprise and Spark pods
  • All Splunk Enterprise containers now run using the unprivileged splunk user and group
  • The minimum required version of Splunk Enterprise is now 8.0
  • The splunk-operator container now uses Red Hat's Universal Base Image version 8

0.0.3 Alpha (2019-08-14)

  • Switched single instances, deployer, cluster master and license master from using Deployment to StatefulSet

0.0.2 & 0.0.1

  • Internal only releases