000029055 - Capture state shows "WAITING FOR DICTIONNARY REDO: FIRST SCN"

Document created by RSA Customer Support Employee on Jun 14, 2016Last modified by RSA Customer Support Employee on Apr 21, 2017
Version 2Show Document
  • View in full screen mode

Article Content

Article Number000029055
Applies ToRSA Product Set: RSA Data Protection Manager
RSA Product/Service Type: RSA Data Protection Manager Appliance 3.2.x and 3.5.x
 
IssueCapture state shows "WAITING FOR DICTIONNARY REDO: FIRST SCN"
CauseKnown defect
Workaround

Install the CS status pages, version 2.0.13 minimum is required.


First, identify the minimum archivelog filename and sequence number required. Run the following script, provide the SCN from the error message, and get the archivelog filename


/opt/rsa/setup/sh/cs_get_archivelog_for_scn.sh


When prompted with "Enter value for scn", use the value from the error message, the one from the "red box".


Example:


[root@dpm-1 sh]# /opt/rsa/setup/sh/cs_get_archivelog_for_scn.sh
Enter value for scn: 62934654
old 1: SELECT SEQUENCE#, FIRST_SCN, NEXT_SCN, NAME FROM DBA_REGISTERED_ARCHIVED_LOG WHERE FIRST_SCN <= &&scn AND NEXT_SCN >= &scn
new 1: SELECT SEQUENCE#, FIRST_SCN, NEXT_SCN, NAME FROM DBA_REGISTERED_ARCHIVED_LOG WHERE FIRST_SCN <= 62934654 AND NEXT_SCN >= 62934654
SEQUENCE# FIRST_SCN NEXT_SCN Log File Name
---------- ---------- ---------- ----------------------------------------------------------------------------------------------------
3977 62934642 62934654 /opt/oracle/oradata/ONERKM/archdest/1_3977_774283138.dbf
3978 62934654 62935433 /opt/oracle/oradata/ONERKM/archdest/1_3978_774283138.dbf

Now check if the second file (and all subsequent following files) exists:


Example:


[root@dpm-1 sh]# ll /opt/oracle/oradata/ONERKM/archdest/1_3978_774283138.dbf
-rw-r----- 1 oracle oinstall 15314432 May 9 04:42 /opt/oracle/oradata/ONERKM/archdest/1_3978_774283138.dbf


If it does exists in the Oracle archdest folder. Just restart the capture process and the Oracle log miner will fix things automatically:


Example:


[root@dpm-1 sh]# /opt/rsa/setup/sh/cs_stop_capture.sh [capture name to stop]
[root@dpm-1 sh]# /opt/rsa/setup/sh/cs_start_capture.sh [capture name to start]


 


If the file does not exists,


Re-register the archive logs, run :


su - oracle
rman target /
catalog start with '/opt/oracle/oradata/ONERKM/archdest';


Repeat the above command till all the files are registered. it registers 15 files per iteration.


In both cases above (file exists or not), if you fixed the "WAITING FOR DICTIONARY REDO: FIRST SCN" error, run the following command to generate an Oracle archivelog containing the dictionary redo. This will bump up the value of FIRST SCN, which is good:


./cs_first_scn.sh -b


If after catalog and re-register of archive log files dictionary redo has been found.  Restart capture process as shown above.


The archivelog file may have been zipped by the appliance's internal jobs. Get the filename returned by the query above then run the following command (where 1_454_ 774283138.dbf is an example filename used):


for x in /opt/log_archive/*.tgz; do echo "checking $x"; tar ztvf $x | grep "1_454_774283138.dbf" && echo "Found in $x"; done


If the above query returns a filename, then you are not out of luck. Extract this backup file (and all subsequent newer tgz files) to get ALL archivelogs missing. Extract this way: 


mkdir /opt/tmp
tar zxvf /opt/log_archive/<filename.tgz> -C /opt/tmp


Move the extracted .dbf files in  /opt/oracle/oradata/ONERKM/archdest/ . You can use this command to move the extracted files back in the Oracle archdest folder if the files were extracted in /opt/tmp :


find /opt/tmp/opt/log_archive/ -type d -name archdest -exec mv {}/*.dbf /opt/oracle/oradata/ONERKM/archdest \;


If the archivelog containing the missing SCN cannot be found on the host where the error is, it may be on a different host. You can try searching for it.

Attachments

    Outcomes