ORA-21779: duration not active problem solution

ORA-21779: duration not active problem solution

The user tells a RAC environment database to generate a large number of ORA-21799 error messages in the log, the alert Log part is truncated as follows:

Mon Jan 06 10:27:24 2014
Errors in file d: \ Oracle \ product \ 10.2.0 \ admin \ orcl \ bdump \ orcl1_smon_22688.trc:
ORA-21779: duration not active.
ORA-06512: at line 1


Mon Jan 06 10:27:26 2014
Errors in file d: \ oracle \ product \ 10.2.0 \ admin \ orcl \ bdump \ orcl1_smon_22688.trc:
ORA-21779: duration not active.
ORA-06512: at line 1

Mon Jan 06 10:27:27 2014
Errors in file d: \ oracle \ product \ 10.2.0 \ admin \ orcl \ bdump \ orcl1_smon_22688.trc:
ORA-21779: duration not active.
ORA-06512: at line 1

Mon Jan 06 10:27:29 2014
Errors in file d: \ oracle \ product \ 10.2.0 \ admin \ orcl \ bdump \ orcl1_smon_22688.trc:
ORA-21779: duration not active.
ORA-06512: at line 1

Mon Jan 06 10:27:31 2014
Errors in file d: \ oracle \ product \ 10.2.0 \ admin \ orcl \ bdump \ orcl1_smon_22688.trc:
ORA-21779: duration not active.
ORA-06512: at line 1

Mon Jan 06 10:27:40 2014
Errors in file d: \ oracle \ product \ 10.2.0 \ admin \ orcl \ bdump \ orcl1_smon_22688.trc:
ORA-21779: duration not active.
ORA-06512: at line 1

Mon Jan 06 10:27:41 2014
Errors in file d: \ oracle \ product \ 10.2.0 \ admin \ orcl \ bdump \ orcl1_smon_22688.trc:
ORA-21779: duration not active.
ORA-06512: at line 1

Further analyze the d: \ oracle \ product \ 10.2.0 \ admin \ orcl \ bdump \ orcl1_smon_22688.trc log mentioned in the log. The content is as follows:

* ** Service name :( SYS $ BACKGROUND) 23:24:57. 404
* ** Session id: (981.1) 23:24:57. 404
* ** 23:24:57. 404
Start recovery for domain 0, valid = 0, flags = 0x0
Validate domain 0
Validated domain 0, flags = 0x0
* ** 11:23:32. 519
Drop transient type: paipte10f0jkthw7t5an5zerha =
* ** 11:23:32. 519
SMON: following errors trapped and ignored:
ORA-21779: duration not active.
ORA-06512: at line 1
Drop transient type: paipte10f0jkthw7t5an5zerha =
* ** 11:23:33. 299
SMON: following errors trapped and ignored:
ORA-21779: duration not active.
ORA-06512: at line 1
Drop transient type: paipte10f0jkthw7t5an5zerha =
* ** 11:23:34. 188
SMON: following errors trapped and ignored:
ORA-21779: duration not active.
ORA-06512: at line 1
Drop transient type: paipte10f0jkthw7t5an5zerha =
* ** 11:23:35. 202
SMON: following errors trapped and ignored:
ORA-21779: duration not active.
ORA-06512: at line 1
Drop transient type: paipte10f0jkthw7t5an5zerha =
* ** 11:23:36. 216
SMON: following errors trapped and ignored:
ORA-21779: duration not active.
ORA-06512: at line 1

Installing Oracle 12C in Linux-6-64

Install Oracle 11gR2 (x64) in CentOS 6.4)

Steps for installing Oracle 11gR2 in vmwarevm

Install Oracle 11g XE R2 In Debian

We can clearly see that the error message is generated when you perform the Drop transient type: delepte10f0jkthw7t5an5zerha = operation. What is this operation? You can query metalink as follows:

 

 

SYMPTOMS

 

SMON generated the following errors in the alert log file:

SMON: following errors trapped and ignored:
ORA-21779: duration not active
ORA-06512: at line 1
Drop transient type:
SYSTPfQMEpjI7QJ7gQ6wVAuBAng==ÿ��CAUSE

This error is caused by SMON not able to clean up some transient types and this problem has been reported a few times in Oracle10gR2.

SOLUTION

Other then producing large trace files, this error has no impact. SMON is erroring whilst evaluating a SYSTEM trigger when executing cleanup of TYPEs that are no longer needed, so currently we have the following options to address this:

1. Bounce the DB and see if the error is persistent
   
   OR
2. As suggested by Oracle Development in similar bugs, set the following event at system level to delay the cleanup of these types:
   
   SQL> alter system set events '22834 trace name context forever, level 1'

 

When you see the errors again (ORA-21779), please tail the alert logs on all instances and find out which one spits the error continuously. Then run the oradebug below against those instances:

$ Sqlplus/as sysdba
Oradebug setospid <smon pid>
Oradebug event 22834 trace name context forever, level 1

> Tail th SMON trace, it shoshould generate tons of logs for event 22834, once it stops, also monitor alert log,

When ORA-21779 is not generating anymore, disable the event as follows:

Oradebug setospid <smon pid>
Oradebug event 22834 trace name context off

This issue does not affect database functionality and the only problem is that the alert. log is filling up with internal error messages (also after restarting the instance the error shocould not occur again ).

The command "alter system flush shared_pool" cocould also resolve the problem.

It is a small BUG. It indicates that it does not affect usage, but a large number of logs will be generated, resulting in a rapid growth of alert logs. The solution is to set 22834 internal events, however, this will lead to the growth of transient types in the database, which obviously does not work in the production database. Finally, we mentioned that we can restart the database instance, and this error may not occur again, you can also run the alter system flush shard_pool command.

Clearing the shared pool can also solve this problem.