Unexpected NDMP error communicating with tape server Error Code: [39:218]

Hi ​@drPhil 

Thanks for reaching out, from the above description it indicates the backup runs before it hits this condition, although it runs slow.

One of the first things that sticks out is could be contention on the server hosting the volumes we’re backing up.

Depending on the number of backups/streams we’re running against the filer, we can reduce the number of jobs or streams per job running against the NDMP filer and retry the backup that you’re seeing the job hitting the “Unexpected NDMP error communicating with tape server” condition.

The NASbackup.log should give more information as well to assist with troubleshooting.

Hi ​@Emils - thank you so much for you contribution.

It seems really strange that only one volume is being backed up at a time, resulting in minimal traffic, which shouldn't cause any unusual issues. However, this particular volume is taking an exceptionally long time to back up, while other volumes on the filer are completing much faster. This suggests there might be a network issue, but it's challenging to verify this assumption. 

The logs indicate a significant wait time - [99.9% (99.7% Wait)][899 Secs (897 Wait Secs)] - but it's unclear what the backup is waiting for. Unfortunately, NDMP is one of the most complex and confusing protocols :(

This amount of wait time indicates the backup is waiting on the filer to send the data. This could be that the data is traversing via slower interface or the volume being backed up has a lot of io. Depending on the filer vendor, you can try to back up from snapshot using steps in https://documentation.commvault.com/v11/expert/performing_backup_from_snapshot.html.

Also, if you know the ip if the traffic should come from you can use data interface pairs using the steps in https://documentation.commvault.com/v11/expert/setting_up_data_interface_pair_between_two_clients.html

To force the correct interface.