Netvault has locally attached drives on NetApp filer that are set up as shared with another NetApp filer
NDMP backup fails with “Failed to start NDMP transfer” in binary log as seen below and at the same time we see “tape is currently in use by other operations 40:1156.” In the filer /etc/messages file.
From Binary log:
Background 2012/02/27 14:24:17 105109 Media LDSBAK01 Mover has halted as all data has been transferred
Error 2012/02/27 14:27:14 105109 Media LDSBAK01 LDSBAK01 Tape pmfiler03:nrst1a (HP Ultrium 4-SCSI): had transfer aborted
Error 2012/02/27 14:27:14 105109 Data Plugin LDSBAK01 Failed to start NDMP transfer
Error 2012/02/27 14:27:14 105109 Jobs LDSBAK01 Fatal error: Backup failed
Error 2012/02/27 14:27:14 105109 Jobs LDSBAK01 Job Status: Backup Failed
Job Message 2012/02/27 14:27:14 105109 Jobs LDSBAK01 Finished job 105109, phase 1 (instance 1)
From messages file:
Mon Feb 27 14:22:17 GMT [pmfiler03: smtape.backup.err:error]: Backup of '/vol/exch01_mailbox_corptax01' to 'NDMP_LOCAL_000000' failed: tape is currently in use by other operations 40:1156.
Mon Feb 27 14:22:28 GMT [pmfiler03: smtape.backup.err:error]: Backup of '/vol/exch01_mailbox_corptax02' to 'NDMP_LOCAL_000000' failed: tape is currently in use by other operations 40:1156.
This is an OnTap problem below is NetApp explanation of the problem.
Netvault re-use the same ndmp session on the filer for another backupas well
and called newData(), and newData zero the old data structure, and modifying sessionid to 0 and re-start the session.
That why all multiple session is showing sessionid '0', and could not proceed parallel backup.
See NetApp KB for full details using URL below:
http://support.netapp.com/NOW/cgi-bin/bol?Type=Detail&Display=526717
Workaround:
Using ndmpdData.setSessionId in newData() for re-using the same old sessionid for further smtape transfer.
Problem fixed in following versions of OnTap:
Data ONTAP 7.3.7RC1 (First Fixed) - Fixed
Data ONTAP 7.3.7 (GA) - Fixed
Data ONTAP 8.0.3 (GA) - Fixed
Data ONTAP 8.1.1RC1 (RC) - Fixed
