Harness Self-Managed Enterprise Edition Release Notes

Updated 3 weeks ago by Michael Cretzman

This document contains release notes for the Harness Harness Self-Managed Enterprise Editions.

For Harness SaaS release notes, see Harness SaaS Release Notes.

Release notes are displayed with the most recent release first.

Minor Release 75324

Release Date: June 10, 2022

We're pleased to present Harness On-Premise Minor Release 75324.

General What's New and Early Access

New Features and Enhancements

The following new features were added to the Harness platform components:

Delegate
  • Fix LdapSettingsDTO Kryo backwards compatibility (DEL-4099)
    • Fix LdapSettings' DTO class to ensure Kryo backwards compatibility of new manager with old delegate.

Issues Fixed in the Release

  • LDAP searchGroupByName not using CapabilityCheck (PL-25483, ZD-30998)
  • User Cannot Delete Group. After this change, userGroups referencing in a pipeline will not be allowed to delete via graphQl unless those pipelines are deleted (PL-24821, ZD-30056, ZD-30935)
  • Harness internal users are able to delete resources using graphql on customer's accounts (PL-24764)
  • [S3 or GS Based Helm Repo] Multiple Customers reporting - Chartmuseum on the Delegate seems to be on stalled state, so port is always in use. Fixed retry logic to start chart museum server in case the port is already in use. (CDS-38135, ZD-31129)
  • Artifact collection with FF 'DISABLE_ARTIFACT_COLLECTION' on Nexus3 artifact collection is failing with Nexus 3.x does not support getVersion for parameterized artifact stream (CDS-38030)
  • Artifact History option is not shown for already auto collection artifact sources when we enable the FF (CDS-37958)
  • For Docker artifact streams, changing collection value from yaml doesn't work (CDS-37956)
  • Deployment Permissions for a Specific Pipeline No notes req since it was for one cuxtomer (CDS-37918, ZD-30951)
  • AWS Secret Manager not using Delegate Selector for Pulling Secret. Added the execution capability of SM to verify the eligibility of delegate to access the SM (CDS-37797, ZD-30884)
  • Missing Git File error for CloudFormation from GIT does not surface to the Deployment Logs (CDS-37670, ZD-30781)
  • Manage gets No delegate task found made timeout as required field (CDS-37524, ZD-30412)
  • PCF Deployment Fails with SweepingOutputServiceImpl.save Error (CDS-37500, ZD-30761)
  • [UI - Known Issue (API call)] Limitation to render more than 1k elements on `Deployments - > Start new Deployment` Blocks customer use case new change will make call in group of 15 items every time you scroll, and also if you type it will search for those 15 items with the written text. The changes are behind a Feature Flag, so that testing scope is not increased and we can quickly move the changes (CDS-37471, ZD-30667)
  • Azure - Artifactory incorrectly regenerates/reads the artifact path (CDS-36596)
  • Improve error message for past window activation for freeze window (CDS-36119)

Delegate Version

  • 75323

Minor Release 75281

Important information regarding certificates and Delegate upgrade to OpenJDK 11:

As part of our migration to OpenJDK 11, Delegates will download OpenJDK 11 and get started with it instead of the previous OpenJDK 8 that was used. This means that Delegates will have Java installed on a different path. Any user actions which modify or use older Java versions would be affected. Some of the most common scenarios include adding certificates to the default Java keystore and running additional Java applications using Delegates installed Java (e.g. from a profile or init script).

See Information regarding certificates and Delegate upgrade to OpenJDK 11.

Release Date: May 31, 2022

We're pleased to present Harness On-Premise Minor Release 75281.

General What's New and Early Access

New Features and Enhancements

The following new features were added to the Harness platform components:

Platform
  • Support skipValidation for secret created using Custom Secret Manager (PL-23731)
CD
  • For CloudFormation steps, we are going away from AWS CloudFormation update() API and moving to AWS CloudFormation Change sets to update stacks. (CDS-37148)
    • This will help customers properly validate CF Templates if there are no changes done. This is behind FF and impacts only stack updates. The creation of the stack will work as it is working right now.
  • Apply deployment tags (Pipelines and Workflows) during the execution. (CDS-29002)
    • With this change, Pipeline tags are resolved after every stage is finished. Previously they were being resolved only twice: before Pipeline execution started and after Pipeline execution finished.
Delegate
  • Remove iterator logging and build first class support with OpenCensus.
    • As part of streamlining our monitoring and reducing the cost of logging, all the background jobs metrics are moved from log based metrics to OpenCensus metrics providing us better visibility into our critical paths. (DEL-3835)

Issues Fixed in the Release

  • cURL query for createFreeze window returns error message on using API key (PL-24899, ZD-29519, ZD-30296)
  • Delayed Sync with Harness UI and Github Config as Code Repository. We had some slowness in case a file was renamed in Harness -→ git. We have improved it going forward. (PL-24662, ZD-28316)
  • UI is not preserving the error message format which BE is sending. Although backend was sending newline characters, frontend wasn't honouring them. It is fixed now using a CSS rule. (PL-23726)
  • Delegate watcher throwing null pointer exception on DELEGATE_SHUTDOWN_TIMEOUT. Handle null value scenario while reading delegate data. Watcher was into a loop of reading Delegate data. Couldn't make progress because of this NPE. Fixed the code path by making correct assumptions. (DEL-4039)
  • Manifest collection task assignment is failing (DEL-3999)
  • delegateScopeList API call not working. Fixed the bug of listing Delegate scopes when the environment is not provided (DEL-3991, ZD-30462)
  • Added support for special characters in proxy password for delegate proxying. (DEL-3348, ZD-26493)
  • Triggers fail when optional input of type allowed values is not passed in the payload. Issue scenario: If you had an optional workflow variable with some allowed values, and the value was not provided for that workflow variable in trigger payload, and trigger executions starting failing with error: Trigger rejected because a passed workflow variable was not present in allowed values. Fix: Changed the validation to only occur when variable value is provided in trigger. (CDS-37510, ZD-30760, ZD-30769, ZD-30773, ZD-30777, ZD-30782, ZD-30797, ZD-30798, ZD-30810, ZD-30812)
  • Bamboo step is failing and Delegate is throwing Class is not registered: software.wings.delegatetasks.BambooTask$BambooExecutionResponse execption (CDS-37264, ZD-30567)
  • SSH Step -> Download Artifact: NPE with DownloadArtifactCommandUnit.java. Gives UNKNOWN_ERROR on the UI (CDS-37193, ZD-30299, ZD-30498, ZD-30499)
  • ArtifactMetadataEvaluator Class is not registered in QA (CDS-37055)
  • Artifactory Docker autosuggestion shows wrong value for Artifactory on-prem/server for "Docker Repository Server" (CDS-36996, ZD-30023, ZD-30402, ZD-30568, ZD-30623, ZD-30628)
  • Clicking Task Definition in newly created ECS service: "Something went wrong...". Added null check (CDS-36995, ZD-30422)
  • Canary failed: [Canary failed - Invalid request: Infra Definition is not associated with Environment error. When the environment change, we clear the infrastructure definitions to force the user to review every phase and manually fix it. That behavior is a feature and not a bug. For a reason, after the clone operation, the workflow is not set as incomplete and that is what we fixed here. (CDS-36932, ZD-30269)
  • [Rancher] NPE on Rancher Rollout Deployment - 400-rest/src/main/java/software/wings/delegatetasks/rancher/RancherTaskHelper.java:105. Bug resolved by adding null checks (CDS-36740, ZD-30133)
  • Kyro exception seen in the QA manager logs for Jenkins server while fetching builds information (CDS-36729)
  • Cannot Select all User Groups in Freeze Window (CDS-36681)
  • When using REST or GraphQL the workflow execution status return a value different of the UI. To get the current status value is required to transpose the execution tree and this take sometime. This refresh status operation is behind a FF named WORKFLOW_EXECUTION_REFRESH_STATUS. (CDS-36623, ZD-29921)
  • Error when editing a Cron-based Trigger where the existing cron value is now in the past. Trigger editing problem with cronjob, where frontend was sending an unused field with a value that was exceeding the Long max value from java. The field was marked to be ignored by the backend. (CDS-36570, ZD-29822, ZD-30469)
  • Workflow won't abort: showing up as paused. Change TTL of WaitInstance to 90 days, and add a log when a WaitInstance is not found to help us in future troubleshooting. (CDS-36568, ZD-29994)
  • Cannot Select all User Groups in Freeze Window. We were not supporting selecting all the user group and user had to click one by one to select all user group if he wants to notify all. Going forward we can select "ALL" option to notify all at once. (CDS-36556, ZD-29926)
  • Custom Approval Step - Shell Script - Secret exposed (CDS-36154, ZD-29588)
  • Artifactory incorrectly regenerates/reads the artifact path - AWS Lambda (CDS-36045)
  • YAML Validation is missing for service and environment selection while creating the Deployment Freeze Windows (CDS-35705)
  • JENKINS_GET_JOB task not listing all jenkins job. Fixed an issue where manager could not connect to delegates with scopes related to application. (CDS-35628, ZD-29044, ZD-29114)
  • Unclear Error message on Fetch Instance MARK_EXPIRED. Added a proper message when aborting fetch instance step. (CDS-35519, ZD-29062)
  • Concurrent Execution of Create AMI Stack fails upon Select Nodes. We fixed certain workflows which use AWS Infrastructure with Tags that were getting in a race condition if tags are provided as an expression. This also resolved the incorrect instance counts for such services. (CDS-35152, ZD-29973)
  • Usergroup is getting deleted even if it is used in notification strategy in workflow. Restricted the deletion of any usergroup when it is used in any notification strategy in any workflow (CDS-34159)
  • Powershell script is not terminated when the shell script step timeouts. When the workflow is aborted or expired, the running shell script step and all its child processes now terminate immediately. (CDS-24447, ZD-20817, ZD-29183)
  • Services Dashboard - Build# does not match reality (CDS-2992, ZD-18477, ZD-18483, ZD-29101)
  • CE Weekly Report was being sent out every 30 minutes. (CCM-7595, ZD-30545)
  • User group creation now automatically adds `VIEW_CE` account permission which breaks everyone using Terraform. Fixed the issue with CE_VIEWER permission not getting disabled for user groups. (CCM-7324)

Delegate Version

  • 75275

Minor Release 74969

Release Date: May 11, 2022

We're pleased to present Harness Self-Managed Enterprise Edition Minor Release 74969.

New Features and Enhancements

The following new features and enhancements were added to Harness Self-Managed Enterprise Edition:

  • Custom selector API for Delegate NG Token management is available now. (DEL-3558, ZD-21717, ZD-27515)
  • Harness can now send key Workflow and Pipeline deployment events to a URL endpoint as a JSON payload. This helps in analyzing how Workflows and Pipelines are performing, using other tools that consume and build dashboards for the events. This feature is now public and not behind the Feature FlagFF APP_TELEMETRY. (CDS-35193)
  • You can now filter out the desired application on the Deployments page using the search option in the Application dropdown. (CDS-34070)
  • We are reducing the heap size of Current Gen Delegate processes from 4gb to 1.5gb. We are also reducing the k8 memory requirement from 8gb to 4gb. This is behind a feature flag and will be rolled out with various accounts. (DEL-3365)
  • We now have the new application filter component that supports API-driven BE search as well as infinite scrolling. So we won't fetch all the applications at once, but instead fetch them in a batch size of 15 and relevant to the search text typed. (CDS-36365)

Issues Fixed​

  • We have removed the Walkme integration from Harness. (PLG-988, ZD-30008)
  • When pasting the YAML definition in a Workflow, the Update failed. Reason: NullPointerException error message was thrown without any reason for the failure. (PL-24456, ZD-29097)
  • API Explorer wasn't using the API_URLenvironment variable. A special code was removing the “gateway” prefix from the API calls. (PL-24415, ZD-27919, ZD-29434)
  • Fixed issue: If cluster lacked IPv6 support, delegate-proxy and harness-ui crashloop due to being unable to bind to any IPv6 interfaces on port 8080 [::}:8080 (PL-24002, ZD-28675)
  • Fixed issue: GraphQL Deployment Stats missing (PL-22091, ZD-21917)
  • Fixed issue: Config-As-Code stopped syncing from On-Premise to Git. (PL-18485, ZD-17190, ZD-17744, ZD-18077, ZD-18312, ZD-20225)
  • The setup as code API for performing CRUD with YAMLs had a bug in which the YAML processing status wasn't correctly returned. This issue is being addressed in order to return the correct YAML processing status. (PL-17678, ZD-16868)
  • Config-As-Code YAML upsert API used to return 200 even if the zip was incorrect. In such cases, it will now return HTTP response code 400. (PL-17676, ZD-16868, ZD-16873)
  • AddDelegateScope output payload wasn't displayed correctly. (DEL-3757)
  • Unused column from datagrid has been removed and tooltip user experience has been improved​. (DEL-3742)
  • Fixed issue: Older Delegates were missing Name and selector after automatic restart. (DEL-3714, ZD-29353)
  • When querying the GraphQL API for Delegates, the lastHeartBeat field was a string instead of an integer. (DEL-3712)
  • Fixed issue: Delayed notifications for Delegate Down alert. (DEL-3696, ZE-28359)
  • Error message wasn't displaying the reason why a Delegate task failed. A Delegate task can fail to be assigned to a Delegate for various reasons, like incorrect Selectors, or Scope setup. (DEL-3686, ZD-28494, ZD-29304)
  • The Watcher program communicates with the Delegate using a local file-based channel. A NullPointerException was thrown while reading messages from this channel. (DEL-3623)
  • Ratio API for Delegate wasn't working as expected. (DEL-3603)
  • Fixed issue: Delegate Page wasn't loading​. (DEL-3571, ZD-28886, 28935)
  • The ECS Delegate alerts are not unless all Delegates associated with an ECS group become unavailable. (DEL-3565)
  • For on prem and oss installations, delegate and watcher jars are downloaded from the backend and hence CDN URLs don't need to be present in the YAML. (DEL-3126)​
  • GraphQL API had no method to list Delegate Scopes for an account. (DEL-2963)
  • Docker Delegate installation was failing whenever there was a space in the hostname. (DEL-2597)
  • Displayed the service artifact version and enhanced the design on hovering over the details. (CDS-36735)
  • Fixed issue: Unable fetch children Jenkins jobs in the Jenkins step of the workflow. (CDS-36701)
  • We have now added infinite scroll and search capability for the Triggers page in CD, similar to cloud providers (CDS-36625, ZD-30004)
  • The CloudWatch verification stage was failing because the instance IDs of the hosts were not properly rendered. (CDS-36609, ZD-30035)
  • The expressions $ami.newAsgName and $ami.oldAsgName for AMI Blue Green incrementally shift traffic workflows weren't set correctly. (CDS-36563, ZD-29905)
  • For Deployment Freeze, an empty description was getting saved in the database. (CDS-36562, ZD-29969)
  • Incorrect Delegate selection logs were getting displayed when a Step failed or was skipped. (CDS-36480, ZD-29628)
  • Approval step with slack had an issue in very corner case which was fixed​. (CDS-36429)
  • The users were not getting redirected to the Pipeline execution / Workflow execution page based on the Pipeline / Workflow. (CDS-36352, ZD-29800)
  • Error Message "Duplicate State" was not displaying "Duplicate Step Name" and was displaying an incomplete notification. (CDS-36298, ZD-29769)
  • Jira's issue metadata was not being fetched correctly. (CDS-36158, ZD-29622, ZD-29628, ZD-29775)
  • Pipeline Approval Step was throwing an error UNKNOWN_ERROR. (CDS-36152, ZD-29557)
  • Fixed UI issue in Deployment screen. (CDS-35976)
  • Audit Trail was not showing deleted Applications. (CDS-35910, ZD-29197)
  • An invalid parentheses character within the step name prior to the expression evaluation/resolution wasn't normalized. (CDS-35795, ZD-29253)
  • Jira Approval was failing with an unclear error message. (CDS-35786, ZD-29259))
  • Helm version 3.5 and above was not working with Native Helm Services as the URL name was changing and force-updateflag wasn't working. (CDS-35667, ZD-29165, ZD-29194)
  • An incorrect error message Invalid argument(s): Task has expired was being displayed when adding a Jenkins Provider. (CDS-35524, ZD-29044)
  • Some Workflows were not rendering dynamic Infrastructure Definition expressions for Terraform. This failure happens for Pipelines with multiple Terraform Workflows that use the same Service and Infrastructure Definition expressions. (CDS-35486, ZD-28977, ZD-28997, ZD-29166)
  • Concurrency issues were occurring when multiple processes execute helm repo add, helm repo update, and helm repo pull in parallel. (CDS-35483, ZD-27796)
  • Fixed issue: UI logs and trace blank for failed workflow (CDS-35343, ZD-28913)
  • Fixed issue: Can't find workflow when creating new filter (CDS-35342, ZD-28902)
  • The default failure and notification strategies are not added when a Workflow is created via Git Sync. Git Sync to Harness was adding notification rules into Workflows. (CDS-35112, ZD-28644)
  • Fixed issue: Failure Strategy in Rollback section wasn't working in Post Deployment Rollback (CDS-35099, ZD-28331)
  • Error message for “Delegate task timed out” in Workflow executions is enhanced. (CDS-34845, ZD-28496)
  • When a User Group is used in a Workflow Approval Step or a templatized User Group variable in a Pipeline stage, deletion of User Groups is blocked. (CDS-34698, ZD-28460)
  • When configuring Kubernetes Cloud Provider, skipValidation now works correctly with Custom Secret Managers. (CDS-34599, ZD-28345)
  • Fixed issue: User fails to adopt Shell Parameter Expansion (not JEXL) on Harness Shell Scrip Step. (CDS-34548, ZD-28271)
  • Fixed an issue with WinRM script execution ignoring timeout. (CDS-34168)
  • PCF perpetual tasks that has been failing for more than 7 days for any reason will now be deleted and we will not be executed any further. (CDS-33630)
  • Added pagination to the Cloud Providers page:
    • We will now load only 50 CPs on opening the page
    • There will be search box that the user can use to look for the CP they want (search by name)
    • Added index on the BE to further improve the API search request. (CDS-25818, ZD-16352, ZD-18775, ZD-27387)
  • Fixed the inconsistency in service name in services page. (CDS-25445, ZD-18756)
  • FIxed a bug when Kubernetes rollback wasn't skipped if execution failed before Kubernetes deployment step. (CDS-4025, ZD-15002)
  • Now the error HelmClientException is caught at an earlier stage. This avoids unnecessary execution and saves time for user. (CDS-2022)

Minor Release 74410

Release Date: March 30, 2022

We're pleased to present Harness Self-Managed Enterprise Edition Minor Release 74410.

Important

Harness accounts that use Single Sign-On (SSO) with SAML need to update the redirect URL in the SAML provider to include /gateway because gateway is added in this release.

New Features and Enhancements

The following new features were added to the Harness platform components:

Platform
  • New Authorization UI is now available in On-Prem platforms. (PL-24016)
Delegate
  • We are reducing the heap size of delegate process of CG delegates from 4gb to 1.5gb. We are also reducing the Kubernetes memory requirement from 8gb to 4gb. This is behind a Feature Flag and we will roll this out slowly with various accounts. (DEL-3365)
CDS
  • When you edit a Helm Connector (except the credentials and name), previously collected Helm charts will get deleted and updated. We now show a Confirm | Cancel dialog box so you have the option to cancel the operation. (CDS-29032)
  • The Deployments page has new options for configuring rollbacks after provisioning from a manual intervention, action after retries, and normal rollback vs. rollback after provisioning. (CDS-28963)

Issues Fixed in the Release

  • Anonymous users sometimes report Refiner surveys in CG. (PLG-782)
  • Infinite scroll has been added to the User Groups list page. You can also search for your user groups in this page. A search fetches all user groups across the application. (PL-20963. ZD-18775, ZD-20187, ZD-21829, ZD-26436, ZD-28694)
  • With Nexus3, the artifact.url variable was picking up a random URL from a list of files present in the version. It didn't honor the classifier and extension configured for the artifact source. Now, if both classifier and extension are configured, Harness will pick the matching URL. If they aren't configured, Harness picks a random URL, which is the current behavior. (CDS-25641, ZD-20723)
  • When doing a collection for Nuget, Harness first fetches a list of all versions and then the corresponding data for every version. The latter isn't needed for versions that are already collected. If a version is already collected, Harness doesn't call the API to fetch its corresponding metadata. (CDS-25580, ZD-20739)
  • There was a race condition issue between starting the step execution and the execution going to final status.
    For example: if the step expired before the step execution is started, Harness marks the execution expired. Then, when the step execution kicks in, Harness considers the step failed and starts checking for failure strategies.
    Honoring failure strategies here isn't correct because the execution is already in final status. This could cause a few issues. For example, if the failure strategy set was manual intervention, the workflow is already expired, but it'll show we are waiting on manual intervention.
    As a fix for this, in such race conditions, once the step has gone into final status, and then step execution kicks in, Harness considers the step in final status and won't check for any failure strategies or go into the rollback phase. (CDS-24530, ZD-20175)
  • In Code-as-Config for Azure VMSS BG workflow, the field downsizeOldVMSSS in YAML had a typo ( an extra S). We fixed it by updating the field name, which is directly reflected in YAML files.
    If you have Git Sync enabled, please replace downsizeOldVMSSS with downsizeOldVMSS in yaml files related to VMSS workflows. (CDS-2173)

Delegate Version

74410

Minor Release 73807

Release Date: Feb 23, 2022

We're pleased to present Harness On-Premise Minor Release 73807.

General What's New and Early Access

New Features and Enhancements

The following new features were added to the Harness platform components:

Platform
  • N/A
CD
  • You can now customize your Kubernetes deployments using Kustomize patches. Kustomizations let you create specific Kubernetes deployments while leaving the original manifests untouched. Currently, this feature is behind the Feature Flag KUSTOMIZE_PATCHES_CG. Contact Harness Support to enable the feature. When you enable this Feature Flag, you will be able to use Kustomize version 4.0.0. (CDP-19430)
    • You can add patch file(s) to Kustomize Patches. In Harness Service, click Kustomize Patches in Configuration and then click Add Patches. You can add multiple files by using Add Patches multiple times.
    • Kustomize doesn't natively support variable substitution but Harness supports variable substitution using Harness variable expressions in Kustomize patches.
    • You can also use Harness secrets in patches.
    • You can override the Service settings for Kustomize Patches in a Harness Environment using Service Configuration Overrides.
    • See Kustomize for Kubernetes Deployments.
You cannot use Harness variables in the base manifest or kustomization.yaml. You can only use Harness variables in kustomize patches you add in Kustomize Patches.
CCM
  • N/A
CV
  • N/A

Issues Fixed in the Release

  • N/A

Component Versions

The following table lists the component versions in this release.

Name 

Version

Manager 

73807

Delegate

73807

Watcher

73807

Verification Service

73807

UI

73100

Learning Engine

66100

Minor Release 73610

Release Date: Feb 1, 2022

We're pleased to present Harness On-Premise Minor Release 73610.

General What's New and Early Access

New Features and Enhancements

The following new features were added to the Harness platform components:

Platform
  • Add Azure Active Directory users that belong to more than the 150 group limit (PL-21670) via SAML.
    • When Azure AD users have large numbers of group memberships, the number of groups listed in the token can grow the token size. Azure Active Directory limits the number of groups it will emit in a token to 150 for SAML assertions.
    • To add users that belong to more than 150 groups, you simply add the Client ID and secret for your registered app to Harness.
    • This feature is behind the feature flag AZURE_SAML_150_GROUPS_SUPPORT.
    • See SAML SSO with Azure Active Directory.
CD
  • You can now add non-containerized artifacts for Harness Azure Web App deployment. For example, WAR, NuGet, and Zip. (CDP-19058, CDP-16747)
  • You can now pull Helm Charts of a specific version when needed from the manifest history by selecting Manually pull artifact in the Harness Service. This version can then be used in future deployments and is useful when Helm Chart collection is disabled. (CDC-16422)

Issues Fixed in the Release

  • N/A

Component Versions

The following table lists the component versions in this release.

Name 

Version

Manager 

73610

Delegate

73610

Watcher

73610

Verification Service

73610

UI

72901

Learning Engine

66100

Minor Release 73406

Release Date: Jan 18, 2022

We're pleased to present Harness On-Premise Minor Release 73406.

General What's New and Early Access

New Features and Enhancements

  • N/A

Issues Fixed in the Release

  • VMSS: downsizeOldVMSS appears as downsizeOldVMSSS (with extra s) in YAML (CDP-19407)
    • In Config-as-Code for Azure VMSS BG Workflow, the downsizeOldVMSSS field in YAML had a typo: extra s. Fixed by updating the State field name which is directly reflected in YAML files.
    • Action: If you have Git Sync enabled, please replace downsizeOldVMSSS with downsizeOldVMSS in YAML files related to VMSS workflows.
  • When using NuGet repository type on Nexus with a lot of artifacts, a task might take too long to complete and expire (CDC-16359, ZD-20739)
    • When doing a collection for NuGet we first fetch a list of all versions and then for every version we fetch corresponding metadata. The later is not needed for versions that are already collected. If a version is already collected we do not make API calls to fetch its corresponding metadata.
  • Nexus artifact collection picking up a random URL (CDC-16156, ZD-20723)
    • With Nexus 3 the artifact.url variable was picking up a random URL from a list of files present in the version. It was not honoring the classifier and extension configured for the artifact source.
    • Going forward if both classifier and extension are configured then we will pick the matching URL. If they are not configured we will be going with the current behavior of picking a random URL.
  • Customer is unable to abort running Pipeline in paused state (CDC-15969, ZD-20175)
    • There was a race condition issue between starting the step execution and the execution going to final status.
    • For example, say, the step expired before step execution is started, we will mark the execution expired and then when the step execution kicks in, we will consider step failed and start checking for failure strategies.
    • Honoring Failure Strategy here is not correct since the execution is already in final status. This could cause few issues. For example, if Failure Strategy set was Manual Intervention, the Workflow is already expired but it will show we are waiting on manual intervention.
    • As a fix for this, in such race conditions, once the step has gone into final status, and say then step execution kicks in, we will consider the step in final status and won't check for any failure strategies or go in rollback phase.

Component Versions

The following table lists the component versions in this release.

Name 

Version

Manager 

73406

Delegate

73406

Watcher

73406

Verification Service

73406

UI

72700

Learning Engine

66100

Minor Release 73225

Release Date: Jan 6, 2022

We're pleased to present Harness On-Premise Minor Release 73225.

General What's New and Early Access

New Features and Enhancements

The following new features were added to the Harness platform components:

Platform
  • N/A
CD
  • Use IAM AssumeRole with Terraform Provisioner (CDP-19707)
    • This feature is behind the feature flag TERRAFORM_AWS_CP_AUTHENTICATION.
    • If you want to use a specific AWS role for Terraform Provision, Apply, and Destroy steps' provisioning, you can select the AWS Cloud Provider, Region, and Role ARN.
    • See Option: AWS Cloud Provider, Region, Role ARN.
  • Users Groups with just the Execute Pipeline permission are now allowed to perform actions (abort/pause/manual intervention) on Workflows in Pipelines. (CDC-14155)
CCM
  • N/A
CV
  • N/A

Issues Fixed in the Release

  • Pipeline failed with Workflow grayed out (CDC-15677, CDC-16042, ZD-20466)
    • When pipeline had many stages in parallel, pipeline execution used to intermittently fail with workflow being grayed out. We suspected the root cause of this to be a race condition between threads. We did fix it and added log lines to confirm the fix.
  • Workflow incomplete on adding Custom Shell Script Approval with Delegate Selector step (CDC-15205, ZD-20011, ZD-20527, ZD-21104)
    • Intermittent issues with corrupt Workflows and Pipelines when using Delegate Selectors with Custom Shell Script Approval. This change fixes the issue.

Component Versions

The following table lists the component versions in this release.

Name 

Version

Manager 

73225

Delegate

73225

Watcher

73225

Verification Service

73225

UI

72400

Learning Engine

66100


How did we do?