Summary
When trying to run Oracle backup jobs, you may encounter errors claiming that the database initialization file format is invalid.
Symptoms
When running Oracle backup jobs, you may find the following error in the sbtio.log file:
SBT-1167402 2008/05/19 16:21:48: (debug): cm_recv_rec(1): got 19 bytes
SBT-1167402 2008/05/19 16:21:48: (debug): job reply parsed: 0 (s/n: 252000, status: -1, jobid: 18446744073709551615)
SBT-1167402 2008/05/19 16:21:48: (error): job manjit_NCPW creation failed; please check logs
SBT-1167402 2008/05/19 16:21:48: (error): could not establish a job to work within
SBT-1167402 2008/05/19 16:21:48: (debug): exit sbtopen: -1 (7009, 0)
SBT-1187916 2008/05/19 16:21:48: (debug): cm_recv_rec(1): got 11 bytes
SBT-1187916 2008/05/19 16:21:48: (debug): got a base serial number: 294000
SBT-1187916 2008/05/19 16:21:48: (debug): generated sn: 294000
SBT-1187916 2008/05/19 16:21:48: (debug): sent 1515 bytes on socket 1
SBT-1187916 2008/05/19 16:21:53: (debug): cm_ready was interrupted; continuing
Similarly, the *.jrl file on your Backup Express (BEX) master server may display the below error message:
Mon May 19 16:21:48 2008 SNBGEN1110I LogFile(/BEX-mastr/BEX/logs/0009705c.jrl) opened by Module(mmjrun)
Mon May 19 16:21:48 2008 SNBGEN1111I Module Build ID:[arch(aix43) ver(2.3) build_time(23:15:04) build_date(May 27 2006)]
Mon May 19 16:21:48 2008 SNBGEN1113I Process Info:[pid(618588) user(maestro) cmdline(mmjrun -r -n manjit_NCPW -m backup -d 00000000 -c /tmp/daaaJiiid.cfg -j /tmp/daaaJiiid.jid -f NCPWdfh_10_655142803_1 -p)]
Mon May 19 16:21:48 2008 SNBGEN1112I Env:SSPRODIR(/BEX-mastr/BEX)
Mon May 19 16:21:48 2008 SNBGEN1112I Env:SSICMAPI(-a 6123 -sbs 65535 -lcl yes)
Mon May 19 16:21:48 2008 SNBmmj2006I Processing database initialization file /tmp/daaaJiiid.cfg
Mon May 19 16:21:48 2008 SNBmmj2008E Database initialization file format is invalid
Mon May 19 16:21:48 2008 SNBmmj1279I ckup
Mon May 19 16:21:48 2008 SNBmmj1031E Error calling fn (ob_read_inifile) rc (-1)
Mon May 19 16:21:48 2008 SNBmmj1031E Error calling fn (ob_process_jobinifile) rc (-1)
Resolution
From the *.jrl file you can see that the Database initialization file is not being successfully parsed. Ensure the sbt11cfg.bex or <nodename>.bex files are using the correct file formats. The above parsing error was caused by an accidental removal of comment tags from the .bex file. Should a similar error occur within your enterprise make sure your .bex file strongly resembles the following example file:
# Backup Express SBT v 2.1 configuration file
#
#
# BEX_DATABASE INST_DATABASE # Important - do not remove this line
# BEX_VERSION v2.1.x
#
BEX_JOB_NAME RMAN_NodeBK
#
# SSICMAPI and SSPRODIR are optional. no spaces around `='!
ENVIRONMENT SSICMAPI=-a 6123
ENVIRONMENT SSPRODIR=/backup/BEX/
# Master Service Database Login Variables
#
# TUNE THIS: Backup Express Server hostname
BEX_MASTER_HOST BEXMaster
#
# Backup Source and Destination Variables
#
# TUNE THIS: Client Hostname as BEX knows it
#
BEX_HOST_NAME_TO_BACKUP RMAN_Node
BEX_HOST_OS_TYPE_TO_BACKUP UNIX
#
# TUNE THIS: Where was the backup the we will restore made in first place
#
BEX_HOST_NAME_TO_RESTORE RMAN_Node
BEX_HOST_OS_TYPE_TO_RESTORE UNIX
#
# TUNE THIS: timeout for automatically created jobs in seconds
BEX_OBEX_TIMEOUT 60
#
BEX_TAPE_POOL BEXS_UX49day_1
BEX_DEVICE_POOL 5510_bkpsvr2
#BEX_DEVICE
#
# Backup Options, Leave blank to override with
# Master Server Job Defaults.
#
# Retention Period must be set to 32767
#
# Options That do not apply to obsi jobs have not been listed.
#
BEX_NETCOMPRESS
BEX_DATACOMPRESS
BEX_ENCRYPTION
BEX_MAXTASKS
BEX_MAXDRIVES 1
BEX_TWIN
BEX_TWINTAPEOFFSITE
BEX_ORIGTAPEOFFSITE
BEX_TAPEWAITINTERVAL
BEX_TAPEMOUNTRETRY
BEX_LABELNEWTAPE
BEX_EOJ_ACTION U
BEX_CHECKSUM
BEX_VERIFYLEVEL
BEX_APPENDING
BEX_MAILADDR
BEX_CM_TIMEOUT 0
#
# The following split options should not be changed
#
BEX_RETENTION 49
BEX_MULTITHREAD N
BEX_MULTFS N
BEX_SPLITTHRESHOLD 0