Source:Troubleshooting NDMP

NDMP

NDMP connect error during backup

Problem

• NDMP server cannot connect to SEP sesam (sbc_ndmp) during NDMP backup. Errors like
NDMP error: ERR NDMP4_DATA_CONNECT NDMP4_CONNECT_ERR NDMP error: ERR NDMP4_DATA_ABORT NDMP4_ILLEGAL_STATE_ERR or NDMP error: NDMP_DATA_CONNECT are shown and backup fails.

Possible causes

• In most cases, this is a firewall issue.
• Other incorrect network settings preventing connection to sbc_ndmp.
• In case of NetApp, multiple interfaces may prevent connection to sbc_ndmp.

⇒ Solution

• Exclude the sbc_ndmp.exe from the firewall scanner and run the backup again.
• Check the configuration of your network. Note that your network must be properly configured to allow the connection to sbc_ndmp.
• In case of NetApp, if multiple interfaces are suspected, set the preferred interface option on your NAS device for specifying the interface which should be used for NDMP:
options ndmpd options ndmpd.preferred_interface <interface_name> For more details, see NetApp-specific NDMP configuration.

NDMP on NetApp

NDMP browsing of non-UTF-8 fails and restore to non-UTF-8 volume may not be possible

Problem

• When performing a restore, it is not possible to browse the content of non-UTF-8 NetApp volumes. You might get an error similar to following:

NDMP error:  ERR NDMP4_SNAP_DIR_LIST  ?0x20500106?

This can happen because:

• Non-UTF-8 volumes are not searchable when objects containing umlauts or special characters.
• Consequently, restore to non-UTF-8 volume is not possible if objects with special characters are included in NDMP backup.

The following CLI command displays the encoding language of all NetApp volumes:

volume show -vserver * -fields language

⇒ Solution

• For NDMP backup tasks from NetApp volumes without UTF-8 encoding language, the backup path can only be specified manually.
• If a backup of a non-UTF-8 volume contains objects with umlauts or special characters then the restore can only be done on an UTF-8 volume.

NDMP DAR restore of single files is too slow

Problem

• When using a direct access recovery (DAR) which enables NDMP selective restore, the restore is too slow, even though the DAR should greatly reduce the time it takes to restore individual files.

Cause

• The general prediction is that DAR greatly increases the restore speed, however, this depends on the amount and size of the backed up data. As stated by NetApp, depending on the number of files in the directory, position of the directory and file on the tape, and the number of tapes needing to be read, the restore can take quite some time. For details, see NetApp documentation.

⇒ Solution

• Separate the backup into smaller pieces to reduce the amount of information that DAR has to manage during the restore.

Restore fails with: "Storing of nlist entries failed."

Problem

• Restore finishes with error: "Storing of nlist entries failed."

⇒ Workaround

• Instead of commonly used version 4 of the NDMP protocol, start the NDMP daemon in version 3 by entering the following commands:
ndmpd off ndmpd version 3 ndmpd on
• Then retry the NDMP restore.

See also

NDMP – NetApp-specific NDMP configuration