This content has been marked as final. Show 10 replies
Is parameter sqlnet.expire_time set (in sqlnet.ora)? 'Something' closes the connection, it could be due to this parameter. If there's a firewall between the client and the auxiliary instance, this firewall could have a defined timeout. Turn on tracing on the client side (from where RMAN is invoked):
TRACE_LEVEL_CLIENT = 16
TRACE_FILE_CLIENT = CLIENT
TRACE_TIMESTAMP_ CLIENT = ON
TRACE_UNIQUE_CLIENT = ON
You may need assistance from Oracle support to analyze this problem.
DBMSDirect- The duplicate was failing 11+ hrs in. And I had already tried restarting.
I did eventually find that note.
I added SQLNET.EXPIRE_TIME=10 to my sqlnet.ora, which seemed to help with the Connection Lost Contact errors.
Even with the note, I wasn't able to work out how to get past:
switch clone datafile all.
Eventually I restarted the duplicate, and it completed successfully.
Thanks all for the replies.
Message was edited by:
It depends on what stage the failure occurred. The most important phase is the restoration of the datafiles. If all the datafiles are restored like in this case, then we can a seperate RUN block to restore the controlfile or manually create the controlfile.
Then again create a different RUN block to restore archivelog files and then recover the auxiliary database using the manual 'recover database until cancel using backup controlfile' by feeding the archive logs restored from this RUN block.
Yes this seems to be a "different style" of creating a controlfile by the DUPLICATE command. It first creates the "basic" controlfile by only adding the sys datafile. Later it adds the remaining datafiles by the "switch clone datafile all".
What is the function of switch clone datafile all
I faced the same problem. It just created the basic controlfile with the sys datafile and when it was running the "switch clone datafile all", suddently "it lost contact".
I am gonna try again with the sqlnet.ora settings. It failed for me after 16 hrs of operation !!