Restore Stuck in “In Progress”

The Monitor page shows restore status as “In Progress” even though the restore is either completed or failed.

This article is for Zmanda EBR versions 4.1, 4.2, and 5.1.

Analysis

If the restore job is running for a long time without completing or restoring any data to the destination, there is a possibility that the job might be stuck.

In such cases, the Monitor page will display the job status as “In-Progress” and aborting the job from the Monitor page will fail.

It is possible to verify if the job is actually running from the backend by using the command:

# ps -ef | grep amfetchdump | grep <backupset>

If you don’t see any active process, it means that the job is either completed or failed, and the status is not synced correctly with the ZMC database.

Solution

Clearing the restore job involves updating the corresponding database tables:

1. Connect to the  ZMC database.
   

   # /opt/zmanda/amanda/bin/mysql ZMC

2. Run the following query to find all the restores stuck in In-Progress:
   

   SELECT restore_run_id, datetime, state, receiving FROM restore_restorerunreport WHERE state="In-Progress";

   

3. Update the state and receiving columns in the table for the affected restore_run_id:
   

   UPDATE restore_restorerunreport SET receiving="Completed", state="Success" WHERE restore_run_id = <restore_run_id>;

4.  After exiting the database, check the restore status once again. If it is still showing as “In-Progress”, remove the restore_running.yml file inside the backup set directory.

    /etc/amanda/<backupset>