Veritas NetBackup™ for NDMP Administrator's Guide

Before you try to troubleshoot a suspected problem, review the following operating notes:

• A tape that was created on an NDMP storage unit is in backup format. It cannot be restored from a non-NDMP storage unit. If you duplicate an NDMP backup image, the new copy is still in backup format. It cannot be used for restores on a non-NDMP storage unit.

• In the backup selections list for an NDMP policy, you can include only directory paths. Individual file names are not allowed. Wildcard characters are allowed in backup selections, though some limitations apply to some filers. More information about wildcards in NDMP backup selections is available:

  See Wildcard characters in backup selections for an NDMP policy.

• In a NetBackup NDMP policy, you cannot include a path in the file list that is more than 1024 characters long. This limitation may be further restricted for certain vendors. See the NetBackup NAS Appliance Information guide for path name length information for specific filers.

• Observe the following restrictions to the use of the ALL_FILESYSTEM directive and the VOLUME_EXCLUDE_LIST directive:

  • A VOLUME_EXCLUDE_LIST statement may include a maximum of 256 characters. Create multiple VOLUME_EXCLUDE_LIST statements if necessary to avoid exceeding the limit of 256 characters. If you specify more than 256 characters, the volume list is truncated. A truncated statement may result in a backup job failure, and the error message Invalid command parameter(20) is displayed. VOLUME_EXCLUDE_LIST applies only to ALL_FILESYSTEMS. It does not apply to explicit backup selections or wildcard-based backup selections.

  • With NetBackup Replication Director, if the backup selection includes read-only volumes or full volumes, an NDMP backup job fails with the status code 20 (Invalid command parameter(20)). If you encounter a similar NDMP backup job error, review the ostfi logs to identify the volumes for which the failure occurred. You can use VOLUME_EXCLUDE_LIST statements with the ALL_FILESYSTEMS statement to exclude the read-only volumes and the volumes with insufficient space.

    Note:

    This restriction applies only to NetBackup Replication Director environments.

  More information about these directives is available:

  See ALL_FILESYSTEMS and VOLUME_EXCLUDE_LIST directives.

• The NDMP protocol uses port 10000 for communication.

• On UNIX systems, the NetBackup avrd process uses Internet Control Message Protocol (ICMP) to ping NDMP hosts to verify network connectivity. This protocol is required for the NetBackup for NDMP product.

• If backup jobs or restore jobs are running slowly, verify that the network interface cards (NIC) are set to full duplex. Half-duplex often causes poor performance. For assistance viewing and resetting duplex mode for a particular NAS host, consult the documentation that the manufacturer provides. You may be able to use the ifconfig (or ipconfig) command, as explained in the NetBackup Troubleshooting Guide.

• Do not perform incremental backups of the same NDMP data from two different policies. Incremental backups performed by one of the policies may be incomplete, because NDMP filers perform level-based incremental backups instead of time-based incremental backups. Consider the following example:

  Policy A performs a full backup of /vol/vol1 (level 0).

  Policy B then performs a full backup of /vol/vol1 (level 0). The filer now considers the policy B backup to be the last full (level 0) backup of /vol/vol1.

  Policy A performs an incremental backup of /vol/vol1 (level 1). The policy A incremental backup captures only the data that changed since the full backup that was done by policy B. The incremental backup misses any changes that occurred between the policy A full backup and the policy B full backup.

• NDMP restore jobs may complete successfully even though no data (0 KB) has been restored. This situation can occur when a target volume does not have enough space for an image you are trying to restore.

  • Workaround: Check the restore job details for entries similar to the following messages:

    mm/dd/yyy hh:mm:ss PM - Info ndmpagent(pid=11071) fas2050c1: RESTORE: We recommend that 19 
      inodes and 907620 kbytes of disk space be available on the target volume order to restore 
      this dump.	You have 466260 inodes and 5316 kbytes of disk space on volume /vol/abc_15gb
    mm/dd/yyy hh:mm:ss PM - Info ndmpagent(pid=11071) fas2050c1: RESTORE: This restore will
      proceed, but may fail when it runs out of inodes and/or disk space on this volume.

    Confirm that the target volume does not have enough space for the restore image. If it does not, either free up enough space on the volume to complete the restore job successfully or specify a different restore volume.