Release notes
Release 115; Build 1036
March 28, 2023
Latest Portworx version compatibility
PDS is now compatible with the Portworx 2.13.x versions.
You can find a list of supported target clusters and the corresponding Portworx versions in the Prerequisites topic.
Known issues
| Issue Number | Issue Description |
|---|---|
| DS-5105 | Some PDS functionality is dependent on database access. Examples include monitoring and backups. If you change the password for the pds user, you need to also update the corresponding Kubernetes secret for the deployment. The secret can be found in the namespace with the name <DEPLOYMENT-NAME>-creds.For more information, see Update Kubernetes secret after changing the pds password. |
| DS-5105 | Cassandra backups require database access. If you change the password for the pds user, you need to also update the cqlshrc file located on each pod. The file is located under ~/.cassandra/cqlshrc.For more information, see Update the pds password in the cqlshrc file for Cassandra pods. |
Fixed issues
This release fixes the following issues:
| Issue Number | Reported in Release | Fix Description |
|---|---|---|
| DS-4612 | 114 | It is now possible to select menu items by scrolling through the dropdown menus in the Deploy <data-service> window. |
| DS-4700 | 114 | After rebooting any of the Couchbase pods, error messages will not appear in the pod logs. |
Release 114; Build 1028
February 24, 2023
New Features
Static IP support
PDS now supports static IPs, which you can whitelist in your firewall rules. For more information, see Static IPs support.
Improvement
The Portworx CSI provisioner is used as a default provisioner for Prometheus volumes in new PDS installations. The Portworx in-tree provisioner is deprecated and its support may be discontinued in future PDS releases.
Known issues
| Issue Number | Issue Description |
|---|---|
| DS-4579 | When you deploy Couchbase on RKE2 and upgrade from previous image, the Failed to pull and unpack image error message appears. This issue occurs, because the containerd versions older than 1.6.10 had issues deflating images with extended attributes.Workaround: Couchbase on RKE2 requires the containerd version 1.6.10 or newer. |
| DS-4612 | When you click a dropdown menu (Backup, Target, and so on) in the Deploy <data-service> and Edit Deployment windows, you cannot pull the scroll bar down to select the menu items.Workaround: Use the down arrow on the keyboard, the wheel button on mouse, or any peripheral devices to scroll down and select the dropdown menu items. |
| DS-4700 | When you reboot any of the Couchbase pods, the nodes are restored. However, you will see unhealthy error messages in the pod logs. These messages are harmless and would not affect the actual data traffic. |
| DS-4727 | In a multi-node target cluster, if the node that is hosting the pds-operator-target-controller-manager pod goes down, then there will be an expected downtime of 5 minutes between PDS control plane and target cluster. During this time, the target cluster would become disconnected from the PDS control plane and all the data services on this target cluster will be inaccessible from PDS UI.Workaround: Wait for five minutes until the target cluster is reconnected. Once it is connected, you can access all the relevant data services from PDS UI. |
Release 113; Build - 1025
January 26, 2023
New Features
Availability of new data services
This release includes the following new data services:
- Consul
- Couchbase
- Elasticsearch
- MongoDB Enterprise
For more information about the versions supported for these data services, see Supported versions.
Ability to add any users
The PDS adminsitrator can now add users without a valid Portworx account to an existing PDS account. For more infromation about adding users, see Add users.
Deprecation announcement
From this release, support for the following legacy Portworx installations is deprecated:
- Portworx with CSI driver disabled: By default, the CSI driver will be enabled on new Portworx installations. On the older Portworx installations, you can enable the CSI driver.
- DaemonSet installation: PDS is compatible with the Portworx Operator installation. You can migrate the exisitng Portworx installed from a DaemonSet to the Operator.
Known issues
| Issue Number | Issue Description |
|---|---|
| DS-2371 | If you deploy a single-node Redis cluster, then you cannot modify the node count in the Edit Redis Deployment -> Number of nodes dropdown. |
| DS-3265 | When you backup a Redis cluster, the backup job runs just on one container. So, PDS backs up only a part of the data. To perform a full backup, you can update the backup operator or use a new version to run backup jobs on each Redis container. However, updating the backup operator is not backward compatible since the operator runs backup script on each Redis container and it causes race condition on old containers that makes backup to fail.Workaround: To backup Redis data service without issues, use the new Redis image. |
| DS-4014 | In most of the MySQL cluster deployments, the DNS host is taking more than 30 minutes to become available.Workaround: Sometimes, this issue occurs due to the slow propagation of DNS records from AWS Route 53 to clients. You must wait for the public DNS caches to clear. |
| DS-4236 | PDS UI does not properly report the long running MySQL backups.Workaround: To get up-to-date report, check the backup.backups.pds.io and backupjobs.backups.pds.io resources in your Kubernetes target cluster. |
| DS-4356 | Existing backup targets go out of sync after some time, after being successfully synchronized to some EKS target clusters.Workaround: Create a new backup target and check if the sync is successful. |
Release 112; Build - 1022
December 15, 2022
This release provides a seamless experience with many UI improvements reported in the earlier versions.
Known issues
| Issue Number | Issue Description |
|---|---|
| DS-3935 | Sometimes, even for a successfully deployed data service, intermittent error messages may appear in the data service Deployment dashboard after idling in the same screen for a long time.Workaround: Refresh the dashboard and re-login to get rid of the error messages. |
Release 111; Build - 1020
December 08, 2022
New Features
New data service versions
This release includes the following new data service versions:
MySQL: 8.0.31
Redis: 7.0.5
Known issues
| Issue Number | Issue Description |
|---|---|
| DS-3792 | When all nodes in your MySQL cluster crash with OutOfMemory error due to high load, the MySQL cluster could fail to restart due to diverged GTIDs.Workaround: To learn more about this issue and for troubleshooting instructions, see the Advanced data recovery topic. |
| DS-3803 | Data service endpoints are not reachable in the OCP 4.11 cluster. This issue occurs, if the external-dns change request is longer than 32000 characters.Workaround: In the external-dns deployment spec, reduce the batch size for external-dns (for example, 100) by updating the aws-batch-change-size flag. |
| DS-3818 | You cannot backup a data service, while workload is in progress. This issue occurs when there are no available resources for backup process. For example, if a database is serving heavy workload consuming all available resources, the backup attempt may fail.Workaround: You can restrict the workload inside a database to run with limited resources to provide the background processes (for example, backup) enough resources to run. Alternatively, you can retry backup when the database is slightly less busy in consuming resorces. |
| DS-3825 | Sometimes, even though the data service backup is completed successfully, the backup job indicator displays red.Workaround: Delete the data service backup, and perform the backup job again. |
| DS-3838 | PDS allows new database deployments to namespaces that are in the terminating state, even though such deployments always fail.Workaround: Run the following command to check if the deployment’s database is in the terminating state:If the database is in the terminating state, then:
Or Do not deploy into a namespace that is in the terminating state. |
| DS-3874 | If you create a Kubernetes cluster without installing Portworx, and register the created cluster to control plane, then the cluster gets added as Unhealthy. When you remove the Unhealthy cluster from the UI, it is removed successfully. However, a new entry for the same cluster is populated, which gets stuck in the Initializing phase.Workaround: Before removing the Unhealthy cluster from the UI, you must uninstall the PDS Helm chart: |
Release 110; Build - 1016
November 15, 2022
New Features
Portworx CSI support
PDS now allows you to choose Portworx CSI with the PDS chart version 1.9.0 or newer.
For more information about the new storage options, see Add or manage a storage option template.
Ability to add user API keys
You can add user API keys, in the PDS Profile page, to authenticate and authorize API calls to PDS and automate your work with PDS.
For more information, see Add user API keys.
Filter data services and deployment targets
The new search and filter options in the data service deployments and the deployment targets pages help you find one or more specific entries from the entire list.
For more information, see Filter data services.
Anthos support
PDS now supports the Anthos platform version 1.21.
New data service versions
This release includes the following new data service versions:
Cassandra: 4.0.6
Kafka: 3.2.3
RabbitMQ: 3.10.9
IMPORTANT: From this release, you must install the PDS chart version 1.9.0 or newer to ensure new PostgreSQL versions are compatible with older versions (until the chart version 1.8.1).
You can view all supported versions for each data service in the Supported versions topic.
New Kubernetes versions
PDS supports the following new Kubernetes versions in this release:
Vanilla Kubernetes: version 1.24.0
OCP: versions 4.11 and 4.10
Latest Portworx version compatibility
PDS is now compatible with the latest Portworx version 2.12.0.
You can view all supported target clusters and compatible Portworx versions in the Prerequisites topic.
Known issues
| Issue Number | Issue Description |
|---|---|
| DS-3461 | Unable to connect to the PostgreSQL server when the storage reaches its maximum limit.Workaround: Manually resize the PVC to connect to the PostgreSQL server: |
| DS-3494 | Data service deployment fails in the on-premises cluster, if you choose XFS as the filesystem type when adding a storage options template. Note: This issue occurs only with the Portworx version 1.12.0. Workaround: Choose the Ext4 filesystem type, instead, to deploy a data service. |
| DS-3653 | The filter option(s) that you selected in the All Deployment Targets dropdown are valid to only the selected data service. When you select another data service with the same filter options, then the search results will not be accurate.Workaround: When you select another data service, reset the filter and then select new deployment targets from the dropdown to get accurate search results. |
Release - 109; Build - 1011
October 18, 2022
New Features
Sample templates to configure data services
PDS provides the sample Application Configuration and Resource Settings templates with predefined parameters. You can copy these sample templates, modify the parameters (if necessary), and easily configure your data services for deployment.
For more information about using these sample templates, refer to the Configure data services topic.
Update a data service
PDS automatically detects if an updated version is available for a data service. You can then update the data service to the latest version with a single-click. For more information, refer to the Update a data service topic.
Availability of MySQL
PDS now includes the MySQL version 8.0.30 as a data service.
Ensure that you install the pds-target Helm chart 1.8.1 or a newer version before deploying the MySQL data service.
New target cluster
PDS now supports the Google Kubernetes Engine (GKE) version 1.23.8-gke.1900 as a target cluster.
For more information about all supported target clusters, refer to the Prerequisites topic.
Ensure that you install the pds-target Helm chart 1.8.1 or a newer version before deploying the target cluster.
New PostgreSQL version
PDS now supports the PostgreSQL version 14.5.
To know the supported versions for each data service, refer to the Supported versions topic.
Improvements
PDS has upgraded or enhanced functionality in the following areas:
Persistent Volumes created for data service deployments in a target cluster contain PDS labels to make them easier to identify.
The annotations or labels to mark loadbalancers in target clusters as internal were extended to cover more cloud providers.
Known issues
| Issue Number | Issue Description |
|---|---|
| DS-3097 | A downtime of few seconds is expected when you upgrade the PostgreSQL and MySQL data service. |
| DS-3176 | In each data service deployment details page → Additional Details window → Connection tab → >Nodes list displays the 0 (zero) node twice. |
| DS-3211 | The URLs in the Settings -> Data Services -> |
| DS-3222 | You cannot backup a data service when it is getting updated.Workaround: Do not select the Backup button, though it is enabled during the data service update process. |
| DS-3246 | During the data service deployment process, the deployment dashboard displays failure and warning messages.Workaround: Ignore these messages, since the deployment will be eventually successful. |
| DS-3252 | In the MySQL metrics page -> Application Details section, limits get varied for the following metrics:
|
| DS-3282 | The backup targets created in PDS are not reachable from the target clusters for TLS disabled S3 endpoints or TLS S3 endpoints not signed by a trusted certificate authority.Workaround: Enable TLS with a trusted TLS certificate to reach backup targets from target clusters. |
Release - 108; Build - 1007
September 15, 2022
New Features
New target clusters
PDS now supports vSphere with Tanzu, Rancher Kubernetes Engine (RKE1), and RKE2 as target clusters.
Ensure that you install the pds-target Helm chart 1.7.1 or a newer version before deploying these target clusters.
Google Cloud backup credential and target
You can now add Google Cloud compatible backup credentials and backup targets in PDS.
For more information, refer to the Configure backup credentials and Configure Google Cloud targets topics.
Known issue
| Issue Number | Issue Description |
|---|---|
| PWX-26445 | On PDS clusters, volume creation fails with the unauthorized error message.Workaround: Restart Portworx pods.Note: If you are running Portworx using the Portworx Operator, then you can label your nodes with px/service=restart and the Operator will restart the specific Portworx node. For more information, refer to the Troubleshoot Portworx on Kubernetes topic. |
| DS-2936 | On Tanzu Kubernetes clusters only, when you create or edit a ZooKeeper data service deployment, the pods will sometimes crash with the following error message:After a few minutes, this issue will eventually resolve by itself and the pods will run normally. |
Release - 107; Build - 1001
August 18, 2022
New Features
Upgrade Kubernetes cluster
In the Settings → Deployment Targets page → each deployed cluster, an Upgrade button is available now.
For more information, refer to the upgrade PDS system on a target Kubernetes cluster topic.
Release and build number
PDS now includes the About page in the Settings menu, which displays the current PDS release, build details, and the end user terms for reference.
Removal of advanced deployment settings
The dns_zone and load_balancer_source_ranges parameters in the data service deployment dialog box → Advanced section are removed from this release.
Known issues
| Issue Number | Issue Description |
|---|---|
| DS-2628 | In the RabbitMQ deployments, the Application Details → Incoming messages chart in the Metrics page displays only the messages that were sent with confirmation. |
| DS-2659 | The Cassandra data service fails to come up after rebooting the Kubernetes nodes.Workaround: Delete the pod, which is in CrashLoopBackOff state, so that it re-spawns normally:kubectl delete pod <crashing_pod_name> -n <name_space> |
July 19, 2022
New Features
Availability of Redis
PDS now includes Redis version 7.0.2 as a data service.
Data service version and image selection
You can now deploy more than one data service version in PDS, and the corresponding image for that version.
For more information about:
Supported version(s) for each data service, refer to the Supported versions topic.
Selecting the version and image, refer to the Procedure to deploy a data service section.
Restrictions for selecting nodes per target cluster
PDS now allows you to select only a specific number of nodes to deploy each data service on a target cluster. You can select the nodes from the Deploy a data service dialog box → No. of Nodes dropdown.
For more information about the available number of nodes for each data service, refer to the Node restrictions section.
Add an empty application configuration template
You can now add an empty Application Configuration Template, and specify the application configuration parameters later when you deploy a data service.
For more information about adding an empty template, refer to the Add an empty template topic.
Known issue
| Issue Number | Issue Description |
|---|---|
| DS-2399 | The Redis data service does not include the default, empty application configuration and resource settings templates, in this release.Workaround: Before deploying the Redis data service in PDS, the Admin must add an empty Application Configuration template and a Resource Settings template. |
June 30, 2022
This release provides a seamless experience with a reduced number of vulnerabilities reported in the earlier versions.
New Feature
You can now view the status of your deployments. To view real-time events:
In the Deployments page, select a deployed data service.
In the deployment details page, select the vertical ellipsis button next to Edit. From the options, select View Deployment Health.
In the Deployment Health window, select the Events tab.
TIP: You can also click the node hexagons, in the deployment details page, to view the Deployment Health window.
The status descriptions help you know the following information about your deployment:
whether it is failed or created successfully
reasons for the failure
timestamp
Known issue
| Issue Number | Issue Description |
|---|---|
| DS-2334 | If your backup does not complete within the timeout value of 24 hours, then the backup eventually fails on the target cluster.Workaround: Since the backup process will keep running inside the data service pod(s), you can terminate the backup process. |
Jun 20, 2022
Known issue
| Issue Number | Issue Description |
|---|---|
| DS-2181 | In the Metrics of a PostgreSQL deployment, under Application Details, the Server Session Count chart shows wrong Total count. |
| DS-2219 | As an Admin, after adding a backup target, you can only edit the Name of the backup target but cannot change the Cloud account, Bucket name, Container name and Region of the backup target.To change the fields, Pure Storage recommends to delete the backup target and a new backup target with the desired options in the fields. |
May 25, 2022
Known issue
| Issue Number | Issue Description |
|---|---|
| DS-2078 | When multiple PDS data services are deployed on the Portworx cluster, deployments may fail due to a lack of free space on Portworx volumes. The filesystem enters read-only mode when the storage reaches 90% of its capacity, and it may be unavailable if some Portworx node pools go offline.As PDS deployments depend on Portworx volumes, ensure that Portworx clusters are running without errors to avoid a data service deployment failure. |
Apr 28, 2022
Announcing general availability for Portworx Data Services!
Known issues
| Issue Number | Issue Description |
|---|---|
| DS-1113 | If you delete backup credentials manually using the pxctl credentials delete command, PDS fails to detect and recreate the credentials on the target cluster. |
| DS-1607 | In some edge cases when all RabbitMQ pods are terminated at once, the first pod fails to start and the pod’s log contains the following message:Waiting for Mnesia tables for 30000 ms, X retries leftWorkaround: Run kubectl exec -n <namespace> <pod_name> /srv/pds/pds-utils/rabbitmq_force_boot.sh to fix the issue. |
| DS-1927 | Due to a bug in PostgreSQL exporter, an extra stale database connection does not allow deletion of a PostgreSQL database and shows the following error:DROP DATABASE "database-name" ; ERROR: database "database-name" is being accessed by other users DETAIL: There is 1 other session using the database.Workaround:Force drop: Use DROP DATABASE <database-name> WITH (FORCE); instead of DROP DATABASE <database-name>;. It may fail if the database user does not have permissions to terminate existing database connections.Exclude database from metrics collection: You can temporarily or permanently exclude a database from metrics collection which will not allow the postgresql exporter to connect to the database:
|