DMS Release Notes

1.3.0 (2025-06-10)

🎁 Features

• Added support for limiting the maximum number of concurrent workflows. This limit can be configured via the maxConcurrentWorkflowTasks variable. Its default value is set to [10].

• Added support for optionally disabling the creation of new workflows via Temporal UI. This mode can be configured via the temporal_start_workflow_disabled variable. Its default value is set to [true] (disabling the creation of new workflows from the UI).

• Added support for configuring a maximum timeout for cloneVolume and verifyClone activities. This timeout can be configured via the startToCloseTimeout variable. Its default value is set to [15m].

• DMS service health can be monitored via the following alerts:

  • InstanceDown- DMS service is down/not responding to scrapes. This alert is based on the Up metric.
  • HealthFail- DMS service (or one of its sub-services) is unhealthy. This alert is based on the dms_service_health_state metric. Note that these alerts and metrics were available previously. This release fixes and improves some logic related to them and also introduces a minor name change (UpDown->InstanceDown) for improved consistency and clarity.

• Added support for optionally specifying compression mode for the thick-clone operation. When explicitly specified, the given value would be used for the cloned volume/snapshot. When no value is specified, the compression mode used by the source snapshot will be used.

• Added support for optionally specifying sector-size mode for the thick-clone operation. When explicitly specified, the given value would be used for the cloned volume/snapshot. When no value is specified, the sector-size used by the source snapshot will be used.

• Upgraded temporal stack images, as well as observability images.

• Upgraded the version of discovery-client to v1.19.0 included with the DMS release.

🐞 Bug Fixes

• Updated the HealthFail recording rule to point to the correct metric.
• Fixed an issue in the Health probe that erroneously always returned the SERVING value.
• Improved error handling in both the refresh and list clusters.
• Fixed issues with inconsistent use of wid/rid key+values in the log, for improved debug/automated parsing of logs.
• ansible-dms-role: backup dms.yml if updated on upgrade.

🚧 Chores

• Bumped to version 1.3.0.
• Bumped protoc-gen-go to v1.36.5.

1.2.0 (2025-02-13)

🎁 Features

• observability: Added the new dms-dashboard that exposes some useful metrics.
• Added the dms_service_health_state metric.
• Added the cluster_connection_state metric.
• Made optional verifyClone defaulted to false for thick-clone APIs.
• ansible-docker-role: allowed adding linux-users to Docker group.
• Bumped temporal-ui version to temporalio/ui:2.35.0.

🐞 Bug Fixes

• Ran discovery-client with network_mode=host (LBM1-36509)
• ansible-dms-role: unexposed ports 7233 and 5432 for enhanced security.
• Updated metrics emitted during clone/verify activities.
• Updated and exposed port for dms-worker sdk metrics.

📄 Documentation

• Updated API docs with summary and descriptions.
• Added metrics of interest to monitor error issues with the DMS stack.
• Added day-two-operations.md to elaborate on maintenance and troubleshooting of the DMS server.
• Updated dmscli.md with response types.

🚧 Chores

• Bumped version to v1.2.0.
• Enriched metrics and logs of clone/verify with the src/dst endpoint.
• Excluded multipath devices for node-exporter diskstats.

1.1.0 (2025-01-22)

🎁 Features

• Added dms.alert.rules.yml.
• Added node-exporter.alert.rules.yml.
• Added support for optionally specifying the sector size (512/4096) on the destination resource (issue: LBM1-36164).
• Added a new grpc_go dashboard for dms service grpc visualization.
• dms.env: upgraded discovery-client to version 1.18.0.

🐞 Bug Fixes

• ansible: Fixed an issue of attaching multiple DMS services to the same cluster, due to the same pubKeyID being used across different DMS hosts (issue: LBM1-36163).
• thick-clone: Fixed an issue of thick cloning from a snapshot that originally had a sector size of 512 bytes (issue: LBM1-36164).
• metrics: Exposed metrics using http handler over /metrics url.
• prometheus.yml: Updated the reference Prometheus config to scrape the DMS service correctly (the target destination was incorrectly specified).
• docker-compose.yml.j2: Fixed an issue with the discovery-client cfg file name.
• Fixed a permission access issue with DMS dynamicconfig files.

📄 Documentation

• Documented metrics and alerts.
• Packaged proto files in the dms-docs tarball.

1.0.0 (2024-12-30)

🎁 Features

• Support of a single persistent volume for all DMS service files.
• DMSCLI simplified the attach cluster/get credentials flow, by exporting the base64 encoded public key directly to the file.
• Added support for pagination in list workflows.
• Upgraded grpc-gateway to v2.25.1.

🐞 Bug Fixes

• Adjusted lbclient timeouts to 10s for every op and add retries.
• Added verification and protections for snapshots and volumes that already existed on clusters with the same name.
• Fixed incomplete cleanup after some failed thick clone operations.

📈 Performance Improvements

• Enabled configuration of the number of concurrent workflows, changing the default from 2 to 10.
• Enabled configuration of the number of go-routines per workflow, and changed the default from max cores to the min (32, max cores).

API Changes/Product Changes

• Updated lb-ansible version from v11.0.0 to v9.13.0.

• Renamed the following APIs (removing the create prefix):

  • CreateThickCloneVolumeRequest -> ThickCloneVolumeRequest.
  • CreateThickCloneVolumeResponse-> ThickCloneVolumeResponse.
  • CreateThickCloneSnapshotRequest-> ThickCloneSnapshotRequest.
  • CreateThickCloneSnapshotResponse-> ThickCloneSnapshotResponse.

• Added proto buff and generated go code to the docs bundle.