NDMP (EMC filer) backups are failing with EXIT STATUS 229

Article:TECH69091  |  Created: 2009-01-25  |  Updated: 2013-10-28  |  Article URL http://www.symantec.com/docs/TECH69091
Article Type
Technical Solution


Issue



When using the NetBackup for NDMP agent, a backup of an EMC filer can fail with EXIT STATUS 229.  The failure will not occur until all of the backup data have been sent to the storage unit, at which point NetBackup performs a validation check of the data that were sent by the EMC filer.   If the backup image is deemed to be invalid, the backup will report the EXIT STATUS 229 error code. 


Error



events out of sequence - image inconsistency

 

BPTM log:
14:24:53.579 [4434] <16> setup_tir_read: read of TIR file failed: events out of sequence - image inconsistency
14:24:58.590 [4434] <2> bptm: EXITING with status 229 <---------
 
BPBRM log:
14:24:58.841 [4429] <8> bpcr_send_signal: couldn't send signal to pid 4433 on client <client_name>, error = 3
14:24:58.843 [4429] <2> bpbrm kill_child_process: start
14:24:58.843 [4429] <2> bpbrm Exit: client backup EXIT STATUS 229: events out of sequence - image inconsistency
 
BPDBM log:
14:24:57.604 [10143] <32> Function: CatImg_FinishBackup: Corrupt .f file, inode entries missing directory info: File: nbe_cat_c_api.cpp Line: 1518: memory map invalid data,
14:24:57.606 [10143] <2> bpdbm: request complete: exit status 229 events out of sequence - image inconsistency
 

Environment



EMC filer

 


Cause



The reason for the failure relates to the inode map, that is sent to NetBackup by the filer.  If the data sent from the filer contains more than one file with the same inode number, this causes the NetBackup validation check to fail the backup with EXIT STATUS 229.  This problem is usually seen with very large NDMP volumes and/or volumes where files are constantly being created and removed, which can result in newer files/directories using the same inode details which were related to other files/directories that were backed up earlier in the same backup.

 


Solution



There are two possible workarounds for this problem.
Either:
Enable EMC's SnapSure option, by adding "set snapsure=y" above the volume path(s) in the backup selection list of the NDMP backup policy.
Using SnapSure creates a snapshot, which avoids inode re-use in the backup image.  A sufficient amount of free disk space must be available to create a snapshot.  Please refer to the EMC SnapSure documentation for more information.
Or:
If SnapSure cannot be used, add "set type=tar" above the volume path(s) in the backup selection of the NDMP backup policy.
The "tar" backup format is not affected by the problem, because it does not use the inode information.  Using the "tar" backup format will not noticeably affect the backup or restore speeds.
Please note that using the "tar" backup format can present problems for data restore, if there is a requirement to be able to selectively restore data from incremental backups - see the information in TECH202412 (link below).
 
Please refer to the Knowledge Base documents linked below for additional information.



Legacy ID



321567


Article URL http://www.symantec.com/docs/TECH69091


Terms of use for this information are found in Legal Notices