Chapter 1. Migration toolkit for applications 8.3

You can convert immutable PersistentVolumeClaim (PVC) storage classes within a single cluster context by using the mta-ops transfer-pvc command. By specifying matching --source-context and --destination-context parameters alongside --dest-storage-class, mta-ops clones volume data securely to a newly provisioned PVC. When combined with the --pvc-rename-map parameter during the mta-ops transform stage, the pipeline automatically updates all referencing workload manifests (Deployments, StatefulSets, Pods, and Jobs) to bind to the new volume.

For more information, see Converting volume storage classes within the same cluster.

MTA-7421

You can replicate persistent volume claim (PVC) data across disconnected, air-gapped, or cross-cloud environments without direct inter-cluster network routes. Passing the --cloud-storage flag to mta-ops transfer-pvc activates a two-phase sequential transfer by using an embedded rclone engine using its s3 backend. The source cluster uploads volume data to an S3-compatible bucket, and the target cluster downloads it to the destination PVC.

Amazon Web Services (AWS) S3 is the currently validated storage provider; other storage services exposing an S3-compatible API are expected to work when configured through rclone’s s3 backend, but are not currently tested.

The feature supports rclone client-side encryption (--encrypt), credential management via Kubernetes Secrets (--rclone-config-secret) or local configuration files (--rclone-config-file), and automatic cloud storage cleanup.

For more information, see Migrating persistent volume data indirectly through cloud storage.

MTA-7449

Every mta-ops command-line interface (CLI) command execution (export, transform, apply, validate, transfer-pvc) automatically records a persistent JSON Lines audit log to audit/.crane-audit.log in your active working directory. The audit log captures full debug-level diagnostic output, operational metadata, and timestamps regardless of terminal verbosity settings. Sequential command executions append to the existing log file, preserving a complete execution history. You can override the default log file location by using the --audit-log <path> option.

For more information, see Analyzing the mta-ops audit log.

MTA-7454

The mta-ops transfer-pvc command displays a structured [i/N] phase progress tracker, an opening context banner detailing cluster endpoints, and an end-of-run summary showing transfer duration and completion status. If a transfer fails, the error message is reported inline within the failing phase, while automatic cleanup of temporary transfer resources is still performed.

MTA-7453

You can use agents with migration-specific skills or directly with large language models (LLMs) to run application analysis and code resolutions in sandboxed containers. You can enforce role-based control to maintain cluster configurations, resources for agent runs, and to run agent-based migrations with real-time streaming output. The LLM provider configuration provides you with control over maximum context window to prevent context override when running an agent.

For more information, see Configuring and using MTA Agents.

MTA-7422

Use a multi-stage agent workflow run for complex projects that require context continuity. Multi-stage agents run a pre-determined sequence of stages or a pipeline. Each stage has an agent that completes a task. To maintain context continuity between two tasks, the agents read the handoff file from a shared Git branch.

For more information, see Configuring and using MTA Agents.

MTA-7423

When you use a wildcard to import packages used by an annotation in your Java source code, an MTA analysis does not trigger a violation for such annotations. As a consequence, you cannot fix all occurrences of an annotation defined in a rule because MTA triggered no incidents for some annotations.

No known workaround exists.

MTA-6530

The analysis report downloaded from the MTA web console has duplicate labels in the Technology tab, leading to confusion when viewing technology categories.

No known workaround exists.

MTA-6142

After connecting MTA to a JIRA instance, the web console does not display JIRA tickets in the application inventory or for migration waves. As a consequence, you cannot track application migrations in the MTA web console.

No known workaround exits.

MTA-5930

The MTA Visual Studio Code extension 8.1 renders inconsistently in the Red Hat OpenShift Dev Spaces dashboard when you use the Firefox version 149.0. To work around this problem, use the MTA Visual Studio Code extension in Red Hat OpenShift Dev Spaces through the Google Chrome browser.

MTA-6735

If the audit log file (audit/.crane-audit.log or a custom path specified by using --audit-log) cannot be created or opened for writing due to restricted directory permissions or an invalid destination path, the mta-ops command-line interface (CLI) prints the following warning and continues executing commands without writing audit logs:

level=warning msg="Failed to open audit log file /root/bad/path.log: mkdir /root: read-only file system"

Operations performed during this run are not recorded in the audit trail.

MTA-7870