Release Notes

A list of known issues and limitations with the current release are provided within this section.

2.1.0

Release Date

2021-05-04

What's New

  1. Architecture Change: Earlier TVK used to ship two different scope CustomResourceDefinitions (CRDs) : Cluster and Namespaced . Now, we only ship Namespaced Scope CRDs. However TVK still support both the "Cluster" & "Namespaced" installation mode as before.

  2. Multi Cluster Management: TVK now supports managing multiple TVK clusters/installations through the UI.

  3. Velero Integration: TVK now ships with velero integration. It will now allow you to list and see the details of your existing velero backups, restore & target in the TVK UI directly.

  4. Analytics Controller: TVK now comes with an analytics controller which populates running stats in trilio resources providing extra information about individual resources.

  5. Helm2 & v1alph1 CRD Deprecation: helm2 support is now deprecated from the TVK and you will not be able to backup/restore the helm2 applications. v1alpha1 version of the TVK CRD's is also deprecated.

  6. Krew Plugins for preflight and log-collector: The shell based pre-flight checks and the go based log-collector tool have been migrated to Krew based plugins. Additional information around usage of these tools is available here

  7. Datamover Performance: This release includes performance improvement to the NFS based targets

  8. Lightweight webhook & OLM bundle: TVK has simplified installation process by moving resource creation to it's Helm chart and the OLM bundle, making the webhook init process lightweight

  9. View/Download YAML: Ability to view raw YAML files for Backup, Restore, Backupplan, Target, Hooks,Retention Policy.

  10. DR Plan: Disaster Recovery Plan feature to support a plan via a single workflow to recover your cluster state in case of any disaster.

  11. Tooltips for all UI forms

  12. Performance and resource analysis via StormForge Experiments - purpose built solution to analyze the resource requirements for TVK pods.

Bugs Fixed

  1. Multiple bugs with respect to the management console and backend APIs have been fixed.

Known Issues/Limitations

  1. Multi-namespace backup in the same workflow not supported yet.

  2. Custom certificates for on-premise S3 targets not supported.

  3. MinIO targets are not supported as MinIO does not support empty folders. Will be fixed in the next release.

2.0.5

Release Date

2021-04-22

What's New

  1. No new features as part of this release.

Bugs Fixed

  1. This release includes a patch fix of taking clusterIP service backup and restore without using transformation on k8s 1.20+

Known Issues/Limitations

  1. OLM Based Upgrade/Install

    1. Clustered

      1. The v2.0.4 version will be automatically upgraded to v2.0.5 via OLM if the subscription is set to automatic. If not, manually approve the subscription upgrade.

      2. v1.1.1 or v2.0.0to v2.0.5 - either incrementally upgrade to the next version up to v2.0.3 or uninstall the current operator and reinstall the 2.0.4 operator from OperatorHub.

    2. Namespaced

      1. Install - Use a custom catalog source to install TVK at a namespace scope.

      2. Upgrade - Uninstall the current operator and install the 2.0.5 operator via custom catalog source.

    3. Clustered <-> namespaces

      1. Same as before - full reinstall

  2. In 2.1.0 (Early May) all clustered installations would have to go through full reinstalls as namespace to cluster and vice-versa upgrades will be supported.

  3. Multi-namespace backup in the same workflow not supported yet.

  4. Custom certificates for on-premise S3 targets not supported.

  5. MinIO targets are not supported as MinIO does not support empty folders.

2.0.4

Release Date

2021-04-01

What's New

  1. No new features as part of this release.

Bugs Fixed

  1. Issue related to cross-cluster migration and custom CRDs failing is now fixed.

  2. 504 Gateway timeout Error Toast Message on UI on Launch of Target Browser

  3. There was a Red Hat OperatorHub issue where Trilio images were not loading for 2.0.3 which has been fixed with 2.04. As a result, 2.0.2 customers can upgrade to 2.0.4 via OLM.

Known Issues/Limitations

  1. OLM Based Upgrade/Install

    1. Clustered

      1. The v2.0.2 version will be automatically upgraded to v2.0.4 via OLM if the subscription is set to automatic. If not, manually approve the subscription upgrade.

      2. v1.1.1 or v2.0.0to v2.0.4 - either incrementally upgrade to the next version up to v2.0.3 or uninstall the current operator and reinstall the 2.0.4 operator from OperatorHub.

    2. Namespaced

      1. Install - Use a custom catalog source to install TVK at a namespace scope.

      2. Upgrade - Uninstall the current operator and install the 2.0.4 operator via custom catalog source.

    3. Clustered <-> namespaces

      1. Same as before - full reinstall

  2. In 2.1.0 (Early May) all clustered installations would have to go through full reinstalls as namespace to cluster and vice-versa upgrades will be supported.

  3. Multi-namespace backup in the same workflow not supported yet.

  4. Custom certificates for on-premise S3 targets not supported.

  5. MinIO targets are not supported as MinIO does not support empty folders.

  6. Kubernetes 1.20 adds a new field known as ClusterIPs for the service resource that is auto-populated from the clusterIP field. Trilio wipes out the clusterIP field from a service resource before a restore, but the ClusterIPs field cannot be non-empty with the clusterIP field. As a result, the restore validation fails. Trilio will fix this issue in a subsequent release, but until then restore transformations can be used to workaround this issue. A video explaining the transform can be found here

2.0.3

Release Date

2021-03-08

What's New

  1. No new features as part of this release.

Bugs Fixed

  1. Issue related to cross-cluster migration and custom CRDs failing is now fixed.

  2. 504 Gateway timeout Error Toast Message on UI on Launch of Target Browser

Known Issues/Limitations

  1. OLM Based Upgrade/Install

    1. Clustered

      1. The v2.0.2 version will be automatically upgraded to v2.0.3 via OLM if the subscription is set to automatic. If not, manually approve the subscription upgrade.

      2. v1.1.1 or v2.0.0to v2.0.3 - either incrementally upgrade to the next version up to v2.0.3 or uninstall the current operator and reinstall the 2.0.3 operator from OperatorHub.

    2. Namespaced

      1. Install - Use a custom catalog source to install TVK at a namespace scope.

      2. Upgrade - Uninstall the current operator and install the 2.0.3 operator via custom catalog source.

    3. Clustered <-> namespaces

      1. Same as before - full reinstall

  2. In 2.1.0 (late March) all clustered installations would have to go through full reinstalls as namespace to cluster and vice-versa upgrades will be supported.

  3. Multi-namespace backup in the same workflow not supported yet.

  4. Custom certificates for on-premise S3 targets not supported.

  5. MinIO targets are not supported as MinIO does not support empty folders.

  6. Kubernetes 1.20 adds a new field known as ClusterIPs for the service resource that is auto-populated from the clusterIP field. Trilio wipes out the clusterIP field from a service resource before a restore, but the ClusterIPs field cannot be non-empty with the clusterIP field. As a result, the restore validation fails. Trilio will fix this issue in a subsequent release, but until then restore transformations can be used to workaround this issue. A video explaining the transform can be found here

2.0.2

Release Date

2021-01-30

What's New

  1. No new features as part of this release.

Bugs Fixed

  1. OCP dynamic forms - inability to enter certain mandatory fields. All mandatory fields are supported now.

  2. Restore by location from Target Browser not loading metadata for transformations and exclusions.

Known Issues/Limitations

  1. OLM Based Upgrade/Install

    1. Clustered

      1. The v2.0.1 version will be automatically upgraded to v2.0.2 via OLM if the subscription is set to automatic. If not, manually approve the subscription upgrade.

      2. v1.1.1 or v2.0.0to v2.0.2 - either incrementally upgrade to the next version up till to 2.0.2 or uninstall the current operator and reinstall the 2.0.2 operator from OperatorHub.

    2. Namespaced

      1. Install - Use a custom catalog source to install TVK at a namespace scope.

      2. Upgrade - Uninstall current operator and install the 2.0.2 operator via custom catalog source.

    3. Clustered <-> namespaces

      1. Same as before - full reinstall

  2. In 2.1.0 (late March) all clustered installations would have to go through full reinstalls as namespace to cluster and vice-versa upgrades will be supported.

  3. Multi-namespace backup in the same workflow not supported yet.

  4. Custom certificates for on-premise S3 targets not supported.

  5. MinIO targets are not supported as MinIO does not support empty folders.

2.0.1

Release Date

2021-01-14

What's New

Management Console:

  1. Namespace and application toggle for changing monitoring.

  2. Filters added for backup view on landing page under monitoring panel

  3. Filters available for backup overview page

  4. Status log with information specific to each PV is available.

  5. Delete backup and delete restore in place from backup tab under monitoring panel on landing page

  6. Consolidated views for backup/restore summary and overview pages will be part of upcoming release

  7. Reusable resources view added to manage backupPlans, retention policies, targets and hooks.

  8. Validation and checks for all workflows added.

  9. Help screens for landing page and Backup/Restore Overview pages added.

Backend:

  1. Ability to set/configure resources configurations for Trilio Pods.

Bugs Fixed

  1. UI related multiple bugs resolved related to user experience and cosmetics.

  2. Post-hook being triggered after upload is fixed.

Known Issues/Limitations

  1. OLM Based Upgrade/Install

    1. Clustered

      1. The v2.0.0 version will not be automatically upgraded to v2.0.1 if the auto publish is enabled as the images for 2.0.0 were being served from the namespace channel. The channel will have to be changed back to cluster to receive the update. A video has been provided on the Upgrade page that demonstrates the process

    2. Clustered <-> namespaces

      1. Same as before - full reinstall

  2. In 2.1.0 all clustered installations would have to go through full reinstalls as namespace to cluster and vice-versa upgrades will be supported.

  3. Multi-namespace backup in the same workflow not supported yet.

  4. Custom certificates for on-premise S3 targets not supported.

  5. MinIO targets are not supported as MinIO does not support empty folders.

2.0.0

Release Date

2020-11-16

What's New

  1. TrilioVault for Kubernetes Management Console

  2. Namespace Level Backups

  3. Restore Enhancements

    1. Transformation

    2. Exclusions

    3. PatchCRD

    4. Skip Operator Resources

    5. Omit Metadata

    6. Restore Hooks

  4. Target Browser

  5. Support for Helm Charts with sub-charts for helm based backup.

Bugs Fixed

  1. Metadata for Helm resources now available in Status sub-resource

Known Issues/Limitations

  1. OLM Based Upgrade

    1. Upgrading from 1.x -> 2.0

      1. Namespace install -> Namespace install

        1. User will have to use custom catalog source for upgrade

        2. Uninstall current Operator

        3. Reinstall Operator via Catalog Source

      2. Clustered install -> Clustered install

        1. Instead of using upgrade mechanism via OLM, do the following:

          1. Uninstall current operator

          2. Reinstall 2.0 from OperatorHub

      3. Clustered <-> namespaces

        1. Same as before - full reinstall

  2. In 2.1.0 (End of January) all clustered installations would have to go through full reinstalls as namespace to cluster and vice-versa upgrades will be supported.

  3. Custom certificates for on-premise S3 targets not supported.

  4. MinIO targets are not supported as MinIO does not support empty folders.

1.1.1

2020-08-25

What's New

  1. No new features.

Bugs Fixed

  1. Licensing bug fixed where Master nodes were being counted for licensing.

  2. Unable to update cron schedule in a BackupPlan.

Known Issues/Limitations

Same as 1.1.0

  1. Sub-charts in case of helm backup not supported.

  2. Custom certificates for on-premise S3 targets not supported.

  3. Metadata information for Helm apps not supported.

  4. Dynamic Forms in OCP should not be used for TVK administration as the forms feature is still under development and does not provide the level of nesting needed by TVK CRDs.

  5. CRDs not deleted when uninstalling - OpenShift does not delete CRDs as part of Operator uninstall. Since TVK supports namespace and clustered installation, this may cause unexpected behavior with the application. Customers must ensure that the CRDs are deleted after Operator uninstall.

1.1.0

Release Date

2020-08-17

What's New

  1. Migration capability from 1.0.0 and v1alpha1 versions.

    1. Customers using Kubernetes lower than v1.15 must enable the CustomResourceWebhookConversion feature gate on the cluster

  2. Restore By Location

Bugs Fixed

  1. Snapshot issue where backups were getting stuck at 6%

Known Issues/Limitations

  1. Sub-charts in case of helm backup not supported.

  2. Custom certificates for on-premise S3 targets not supported.

  3. Metadata information for Helm apps not supported.

  4. Dynamic Forms in OCP should not be used for TVK administration as the forms feature is still under development and does not provide the level of nesting needed by TVK CRDs.

  5. CRDs not deleted when uninstalling - OpenShift does not delete CRDs as part of Operator uninstall. Since TVK supports namespace and clustered installation, this may cause unexpected behavior with the application. Customers must ensure that the CRDs are deleted after Operator uninstall.

1.0.0

Release Date

2020-07-14

What’s New:

  1. IBM CloudPak Certified

  2. IBM MultiCloud Management Certified

  3. Migrated from v1alpha1 to v1 API version

  4. Licensing Enabled.

  5. Label Backup - ability to provide multiple label selectors where the corresponding logic is ORR’ed instead of AND’ed

  6. Upstream Operator Backup - Support for Multiple CRDs

  7. Added support for pre-backup validations

  8. Upstream Operator Backup - Ability to select Operator resources by providing helm release name

  9. Backup Metadata information - Ability to view metadata information of a backup in terms of objects/resources captured.

  10. Grafana Dashboards - 10 Grafana dashboards that are pivoted on Backups/Restores/Targets/Applications (BackupPlans) that provide an overview, summary, and details around TVK resources.

Bugs Fixed:

  1. NFS Timeout issue - Issue with NFS target creation where mount validation would fail due to a timing issue.

  2. Backup by Labels - Backup only top-level items as nested items were being double-counted in the restore, causing it to fail.

Known Issues/Limitations:

  1. Restore by location not supported.

  2. Upgrades are not supported directly and will be supported as a patch before the next released version.

  3. Sub-charts in case of helm backup not supported.

  4. Custom certificates for on-premise S3 targets not supported.

  5. Metadata information for Helm apps not supported.

  6. Dynamic Forms in OCP should not be used for TVK administration as the forms feature is still under development and does not provide the level of nesting needed by TVK CRDs.

  7. CRDs not deleted when uninstalling - OpenShift does not delete CRDs as part of Operator uninstall. Since TVK supports namespace and clustered installation, this may cause unexpected behavior with the application. Customers must ensure that the CRDs are deleted after Operator uninstall.

0.2.3 - 0.2.5

  1. Issues post uninstall:

    1. MutatingWebhookConfiguration and ValidatingWebhookConfiguration is not cleaned up

    2. The Cleaner cronjob is not cleaned up.

    3. Cleaner cronjob will fail to create jobs due to RBAC revocation

    4. In case of scheduled backups defined for any BackupPlan, the cronjob for the same will not be cleaned up unless the BackupPlan is deleted. This will lead to failed jobs due to RBAC revocation Helm charts with sub-charts are not yet supported

  2. In case there is a problem with Target after it is created (change in IP, removal of keys in use), it won’t change the state of Target and will still show Available

  3. OCP UI forms should not be used for resource creation. Instead, the YAML/JSON editor should be used

  4. Listing CRD resources don’t show correct status on OCP UI for operator details

  5. OLM does not consider the scope of CRD it is installing.

    1. Problem: In case TrilioVault is installed in cluster scope and then deleted and after then if the user tries to install in namespaced scope, OLM allows it. On operator deletion, CRD’s are not deleted. So even when the user did the namespaced installation, the underlying CRD’s are installed in cluster scope, leading to issues with application functioning