Disk-level snapshot restore fails if the original disk is detached from the instance
This issue occurs if you are performing a disk-level snapshot restore to the same location.
When you trigger a disk-level snapshot restore to the same location, NetBackup first detaches the existing original disk from the instance, creates a new volume from the disk snapshot, and then attaches the new volume to the instance. The original disk is automatically deleted after the restore operation is successful.
However, if the original disk whose snapshot is being restored is manually detached from the instance before the restore is triggered, the restore operation fails.
You may see the following message on the NetBackup UI:
Request failed unexpectedly: [Errno 17] File exists: '/<app.diskmount>'
The NetBackup coordinator logs contain messages similar to the following:
flexsnap.coordinator: INFO - configid : <app.snapshotID> status changed to
{u'status': u'failed', u'discovered_time': <time>, u'errmsg': u'
Could not connect to <application> server localhost:27017:
[Errno 111]Connection refused'}Workaround:
If the restore has already failed in the environment, you may have to manually perform a disk cleanup first and then trigger the restore job again.
Perform the following steps:
- Log on to the instance for which the restore operation has failed.
Ensure that the user account that you use to connect has administrative privileges on the instance.
- Run the following command to unmount the application disk cleanly:
# sudo umount /<application_diskmount>
Here, <application_diskmount> is the original application disk mount path on the instance.
If you see a
"device is busy"message, wait for some time and then try the umount command again. - From the NetBackup UI, trigger the disk-level restore operation again.
In general, if you want to detach the original application disks from the instance, use the following process for restore:
First take a disk-level snapshot of the instance.
After the snapshot is created successfully, manually detach the disk from the instance.
For example, if the instance is in the AWS cloud, use the AWS Management Console and edit the instance to detach the data disk. Ensure that you save the changes to the instance.
Log on to the instance using an administrative user account and then run the following command:
# sudo umount /<application_diskmount>
If you see a
"device is busy"message, wait for some time and then try the umount command again.Now trigger a disk-level restore operation from the NetBackup UI.