This content has been marked as final. Show 12 replies
have you looked into this1 person found this helpful
Is it possible that your Primary does not have forced logging on?
SELECT force_logging FROM v$database;
My friend CPKT has a good note on a roll forward here :
Sunny kichloo, as I understand in your testcase DB was in NO FORCE LOGGING through all steps you did. In my case I turned DB in FORCE LOGGING mode before taking backup.
Hm, in the post you mentioned I saw rerestoring controlfile on standby... I didn't do that. May be it's the issue. Ok. I'll try it and post result.
Just try restoring control file as mentioned by you and Share the output.
Sunny kichloo ,
No, just backup and recovery of standby controlfile didn't help. FIRST_NONLOGGED_SCN stayed unchanged.
in general, it looks like FORCE NOLOGGING mode doesn't turned on (but it turned). Because after steps I posted, FIRST_NONLOGGED_SCN changed to greater value, and difference between it and current_scn on primary DB much shorter.
So, I'd say that while I doing incremental backup and restoring it on standby db, NOLOGGING operations continue playing on primary db, so by the time I mount standby I have new FIRST_NONLOGGED_SCN.
I read somewhere, can't remeber exactly, but something like 'alter database force logging waits nologging operation to complete'. Is it possible that I need to restart primary db in turn to FORCE LOGGING mode take effect?
Hello again;1 person found this helpful
No, I don't believe you have to restart the database for forced logging.
still can't resyncronize standby.
I noticed on primary db:
select min(unrecoverable_change#) from v$datafile where unrecoverable_change# > 0;MIN(UNRECOVERABLE_CHANGE#)
So i still have unrecoverable data, but db in force logging mode:
select force_logging from v$database;FOR
Edited by: bas_ua on 15.02.2013 15:41
How big is the database?
What is the current Error?
rman report unrecoverable command on primary db showed almost all users datafiles.
So I made full backup of primary db, and then just run duplicate target for stadby.
I know that's not prity well practice for large databases, but it solved the problem.
thanks for help