Troubleshooting HDFS to Ozone migration
Interpret common Cloudera Replication Manager failures, scan gaps, and authentication errors when migrating HDFS data into Apache Ozone with Cloudera Migration Assistant.
Symptom
No directories appear after HDFS scans.
Cause
Overly narrow path filter or aborted scan commands.
Remedy
- Review scan logs.
- Rerun the scan with broader paths (often /) until inventories refresh.
Symptom
Ozone scan exposes zero volumes.
Cause
Ozone service offline or degraded.
Remedy
Verify Ozone availability in Cloudera Manager before resubmitting scans.
Symptom
Structure mapping rejects configuration.
Cause
Destination Ozone nodes missing from deployed layout.
Remedy
Create volumes, buckets, and folders described in Managing Ozone storage, then redeploy the layout and remap.
Symptom
Execution fails at Kerberos checkpoints.
Cause
Expired principals or mismatched CMA Agent keytabs.
Remedy
Refresh Kerberos configurations for each CMA Agent, then restart the agents.
Symptom
Cloudera Replication Manager errors mid-run.
Cause
Unreachable Cloudera Replication Manager endpoint or stale credentials stored in CMA Master.
Remedy
Validate Cloudera Replication Manager URLs and secrets on CMA Master, then inspect Master logs for stack traces preceding the failure.
Symptom
Partial datasets under Ozone.
Cause
Skipped files flagged by DistCp owing to POSIX permissions.
Remedy
Review Cloudera Replication Manager job logs for skip reasons, then widen source read permission for the migration role.
