Error In Execution Backup RMAN with OSB - TAPE_WRITE_INT: UNEXPECTED ERROR - CANNOT ALLOCATE MEMORY (Doc ID 2866332.1)

Last updated on JULY 20, 2024

Applies to:

Symptoms

From time to time below error is received while taking a backup:

tape write failed due to media error
Warning: can't find rogue piece <piece_name>.bkp
object not found (OB database mgr)

and at backup statistics we can see:

piece <piece_name>.bkp completed, status 93

Analyzing through logs what we found in obndmpdms.log below info, happening each time backup fails:

2022/02/26.10:34:09 [22594] tape_write_int: unexpected error - Cannot allocate memory                   >>>>>>>>>>>>>>>>>>>>>
2022/02/26.10:34:09 [22594] svc_movertapewrite: tapeWrite failure causing mover halt (internal error) >>>>>>>>>>>>>>>>>>>>>
2022/02/26.10:34:09 [22594] svc_movererror: sent notify_mover_halted request
2022/02/26.10:34:09 [22594] svc_dataerror: sent notify_data_halted to DMA
2022/02/26.10:34:14 [22594] get_cksum_values: checksum = 888718679, total_bytes = 90373619712
2022/02/26.10:34:14 [22594] svc_moverstop: bytes moved 0x150AB00000
2022/02/26.10:34:14 [22594] svc_datastop: bytes processed 0x150AB00000
2022/02/26.10:34:16 [22594] NDMP_TAPE_CLOSE wrote a filemark
2022/02/26.10:36:12 [25264] ndmpdshmcreate: shared memory area filename is '/usr/tmp/obndmpd_25264_25264_164588257225264a'
2022/02/26.10:36:17 [25264] blocks written 0 current tape pos 0038F3F6 2022/02/26.10:36:18 [25264] svc_movererror: sent notify_mover_halted request
2022/02/26.10:36:23 [25264] get_cksum_values: checksum = 3366890609, total_bytes = 1048576
2022/02/26.10:36:23 [25264] svc_moverstop: bytes moved 0x100000
2022/02/26.10:36:25 [25264] NDMP_TAPE_CLOSE wrote a filemark

But customer had enough memory allocated on the server. We also looked at historical use of memory and there is no such high memory usage.

Cause

My Oracle Support provides customers with access to over a million knowledge articles and a vibrant support community of peers and Oracle experts.