10 Troubleshooting Zero Downtime Migration
This section describes how to handle migration job failures.
For more information about troubleshooting Zero Downtime Migration and known issues in the current release, see the Zero Downtime Migration Release Notes.
10.1 Handling Migration Job Failures
If your migration job fails, the following logs can help you discover the issue.
Migration Job Output Logs
If your migration job encounters an error, refer to the migration job output logs, Zero Downtime Migration service logs, and server-specific operational phase logs present at the respective source or target database servers.
If the migration job encounters an exception (that is, fails) then the logs can provide some indication of the nature of the fault. The logs for the migration procedures executed in the source and target environments are stored on the servers in the respective source and target environments. The Zero Downtime Migration command output location is provided to you when the migration job is run with the ZDMCLI
command migrate database
. You can also find the log file location (Result file path
) in the output of the ZDMCLI
command query job -jobid job-id
.
Zero Downtime Migration Service Host Log
Determine which operational phase the migration job was in at the time of failure, and whether the phase belongs to the source (phase name contains SRC
) or target (phase name contains TGT
) . Check the Zero Downtime Migration service host log at $ZDM_BASE/crsdata/zdm_service_host/rhp/zdmserver.log.0
, and access the respective source or target server to check the log associated with the operational phase in $ORACLE_BASE/zdm/zdm_db_unique_name_job-id/zdm/log
.
If the Zero Downtime Migration service does not start, then check the Zero Downtime Migration service logs for process start-up errors to determine the cause of the error reported. The Zero Downtime Migration service log can be found at $ZDM_BASE/crsdata/zdm_service_host/rhp/zdmserver.log.0
.
Data Pump Log
For logical migration jobs on co-managed target databases, refer to the Data Pump log in the specified DATA_PUMP_DIR
. For Autonomous Database targets, the Data Pump logs are uploaded to a data bucket specified with DATAPUMPSETTINGS_DATABUCKET_*
in the response file. If the data bucket is not specified, then you will need to upload the dump from Autonomous Database to Object Storage to access it. See How To View Import Log Generated For ADW/ATP (Doc ID 2448060.1).
If a migration job fails, you can fix the cause of the failure and then re-run the job while monitoring the logs for progress.
The best practice for troubleshooting a migration job is to query the job
and then, select the logs shown, and tail them. You can tail -f absolute_path_to_log_file
tail
the job and server logs to get a real time information about the migration jobs.
Tailing the job output file, displayed in the output of the query job with label Result file path
, is the preferred way of monitoring the progress of a specific job. This avoids excessive tracing in ZDM server logs caused by the query job
command.
However, it is not recommended to run the query command in a loop, as this command only provides a high level overview on a stage by stage basis, for more detailed analysis, proceed as described above.