NDMP Backups fail with the following error in the job log: "Storage device reported an error on a request to write a media location identifier (file mark)."
Problem
When NDMP data is backed up to a tape drive, some amount of data gets backed up and the jobs fails with the following error in the job log:
"Backup- //san-01.xxxxxxxxx.com/vol/vol0
Storage device "XXX" reported an error on a request to write a media location identifier (file mark).
Error Message
Error reported:
A device attached to the system is not functioning.
V-79-57344-34036 - An unknown error has occurred.
NDMP log shows the following errors:
NDMP Log Message: DUMP: Wed Dec 17 23:31:43 2008 : We have written 226664 KB.
NDMP Log Message: DUMP: dumping (Pass IV) [regular files]
NDMP Log Message: DUMP: Tape write failed.
NDMP Log Message: DUMP: DUMP IS ABORTED
NDMP Log Message: DUMP: Deleting "/vol/vol0/../snapshot_for_backup.10" snapshot.
NDMP Data Halted: Connect Error
NDMP Log Message: Connection or IO Error.
NDMP Mover Halted: Internal Error
NDMP Log Message: MoveletOutput: Internal Error.
Cause
On a NetApp/IBM filer running the ONTAP operating system, you can optimize performance for either minimal transmission delay or throughput. By default, the performance is optimized for overall throughput, in which the packets are grouped in time blocks of 200 ms. If the communication performance is optimized for minimal transmission delay, the queued packets are sent immediately.
Solution
Depending on the version of ONTAP you need to run either
options ndmp.tcpnodelay.enable {ON.EN_US}
OR
ndmpd.tcpnodelay.enable {TRUE.EN_US}
on/true optimizes for minimal transmission delay.
For more information on these commands and its implications on your environment contact vendor or refer to Data ONTAP Data Protection Tape Backup and Recovery Guide
http://www-01.ibm.com/support/docview.wss?rs=1147&uid=ssg1S7002464
http://www-01.ibm.com/support/docview.wss?uid=ssg1S7002615&aid=1