Server Administration :: Objects Invalid Due To Auto-compilation
Jun 27, 2012
One of my friends is facing a peculiar problem where objects are getting "Invalid" during execution I suspect it is happening as they are changing system date during their testing (time travel) which can create conflicted last_ddl_time on objects having dependencies
Consider a scenario
[1] system date is 10-06-2012 there are total 10 objects which has status as 'valid'
[2] the system date is changed to 10-07-2012 Now out of 10 Only 5 objects are compiled During execution ORA-04065,ORA-06508, ORA-06512 are observed
[3] the system date is brought back to 10-06-2012 Again during execution ORA-04065,ORA-06508, ORA-06512 are observed
suppose in step 2 objects are compiled whereas there synonyms are compiled in step 1, only thus last_ddl_time for objects will be later to that of its' synonym...
Does database validate last_ddl_time for objects having dependency during execution and then auto-compiles or invalidates the objects?
I am getting below error while connecting to sqlplus.
SQL*Plus: Release 10.2.0.5.0 - Production on Tue Mar 22 12:47:48 2011 Copyright (c) 1982, 2010, Oracle. All Rights Reserved. ERROR: ORA-06550: line 1, column 7: PLS-00201: identifier 'DBMS_OUTPUT.DISABLE' must be declared ORA-06550: line 1, column 7: PL/SQL: Statement ignored
[code]....
Executed the below scripts but it didnt resolve the issue, whereas some of the SYS objects and catproc got invalid...
dbmsotpt.sql dbmsapin.sql
Now even after reexecuting the catproc.sql and utlrp...Sys objects and the catproc status is still INvalid.
I tried to manually compile the sys objects, but it didnt work.
OWNER SUBSTR(OBJECT_NAME,1,40) OBJECT_TYPE -------------------- ---------------------------------------- -------------------- SYS DBMS_XPLAN PACKAGE BODY SYS AQ$AQ_SRVNTFN_TABLE VIEW SYS DBMS_LOGREP_DEF_PROC PACKAGE SYS DBMS_LOGREP_DEF_PROC PACKAGE BODY
[code]....
how to go about making the SYS objects and catproc VALID and resolve the error which i mentioned above.
I listed invalid object, but when compile object, say that not exist.
SQL> SELECT OBJECT_NAME, OBJECT_TYPE FROM DBA_OBJECTS WHERE STATUS='INVALID' AND OBJECT_NAME ='PCK_PIR_NFEL_MILLENIUM' AND OWNER='PIRAMIDE' OBJECT_NAME OBJECT_TYPE --------------------------------- -------------------- PCK_PIR_NFEL_MILLENIUM PACKAGE
SQL> ALTER PACKAGE PCK_PIR_NFEL_MILLENIUM COMPILE; ALTER PACKAGE PCK_PIR_NFEL_MILLENIUM COMPILE * ERROR at line 1: ORA-04043: object PCK_PIR_NFEL_MILLENIUM does not exist
Is there a way to find out when a datafile for an undo tablespace with autoextend enabled actually extended? I've done a few tests, and nothing is written to the alert log or any trace file that I've found. I can't find any V$ or DBA view that will give me the history of a file's size.
I want to OFF tablspace AUTOEXTEND on a prodution system, we have many RAC databses and that will be done on all stations. i have got a document from net which was written on 29-Jun-2007 and it says that if need to OFF the AUTOEXTEND of a TABLESPACE so you need to ist make it off on the underlying datafiles of that tablespace so this doc is for Oracle 8.1.7.2.0
I am getting the following errors when I try drop a tablespace.
I already did the following.
a) The tablespace & its datafiles offline.
b) I have purged dba_recyclebin.
SQL> drop tablespace db_maintenance including contents and datafiles; drop tablespace db_maintenance including contents and datafiles * ERROR at line 1: ORA-00604: error occurred at recursive SQL level 1 ORA-38301: can not perform DDL/DML over objects in Recycle Bin
I was importing one schema from Oracle 10g to 11g using traditional import. I imported as a SYS user, so all the objects created in SYS schema. how can I remove these objects and retain only default SYS objects
i want create view to select all invalid objects in database.So i create this one:
CREATE OR REPLACE FORCE VIEW INVALID_OBJECTS_DETAILS ( DETAILS ) AS SELECT DISTINCT a.owner || ', ' || a.object_name FROM dba_objects a, dba_source b WHERE a.owner = b.owner AND a.object_name = b.name AND a.object_type = b.TYPE AND a.status != 'VALID' AND b.text NOT LIKE '%@%';
But I want only select invalid objects without a database link .
I am Modifying a table structure, so dependent objects(triggers,packages etc) are getting invalid. So i thought of compiling those invalid objects which are related with the modified table. I used below query to get the invalid objects,
select obj.object_name,obj.object_type from user_objects obj,all_dependencies dep where referenced_name='DEPT' and obj.object_name=dep.name and dep.owner='SCOTT' and obj.status='INVALID'
Q1)What is wrong with this query, sometimes it works sometime it doesn't.? Q2)All_dependencies : does this view show all dependent object on a table even the Invalid one's? Q3) Is there any alternative to find the dependent invalid objects or even dependent object on a table ?
Quote:by default the job runs once every 24 hours.
When you see LAST_ANALYZED being days, weeks, or months in the past do not be alarmed.
If/when the data in a table does not change, then the statistics do not need to change.
Oracle collects new statistics when enough of the data (about 10%) has changed.
Like above statement for DBMS_STATS job, is there a automatic job that runs every day to compile invalid objects for schema or we need to compile it manually ?
we are doing database upgradation 10g2 to 11gr2, while doing pre check before upgrdation ..we have found few duplicate objectes on sys and system schema..
SQL> column object_name format a30 select object_name, object_type from dba_objects where object_name||object_type in (select object_name||object_type from dba_objects where owner = 'SYS')
[code]...
As per metalink note "How to Clean Up Duplicate Objects Owned by SYS and SYSTEM Schema [ID 1030426.6]"..
im going to do drop only below objects ..
DROP TABLE SYS.HELP; DROP INDEX SYS.HELP_TOPIC_SEQ; DROP TABLE SYSTEM.PLAN_TABLE;
and ignore below objects ...
DROP TABLE SYSTEM.AQ$_SCHEDULES; DROP INDEX SYSTEM.AQ$_SCHEDULES_PRIMARY; DROP PACKAGE SYSTEM.DBMS_REPCAT_AUTH; DROP PACKAGE BODY SYSTEM.DBMS_REPCAT_AUTH;
I want to compare two users on different databases, actaully there are two users one in user a on database a and another user b on database b they have same tables, and everytime when a table or object is created on user a (database a) i will take the table name ,procedure or any other object from user_objects based on ddl_created date and then i need to recreate the same on user b on database b, is there a way to find out tables which are not only created but also i need to check whether if there is any column added or any change in procedure or any other objects.Is there a way to generate the scripts based on list of objects selected from user_objects. all i want is.
a)Find out the list of objects added along with creation scripts b)find out the list of objects modified along with creation scripts.
Recently I migrated our Oracle to new machine using exp/imp on schema basis. After import finished I had tons of invalid objects in database. I ran utlrp.sql script and lots of them got validated. Than I recompiled manually in EM those are left but two invalid objects (MGMT_JOB_UI description and body) in SYSMAN schema gave error while recompiling.
Now when I click on any scheduled jobs to edit it or view its schedule, EM throw following error:
X Error jobType - jobType page property expected
I think its related to that invalid package. The errors while compiling the specification are as follow:
Line # = 50 Column # = 1 Error Text = PL/SQL: Declaration ignored Line # = 65 Column # = 9 Error Text = PLS-00201: identifier 'JOBRUNTABLETYPE' must be declared Line # = 88 Column # = 1 Error Text = PL/SQL: Declaration ignored Line # = 107 Column # = 9 Error Text = PLS-00201: identifier 'JOBEXECTABLETYPE' must be declared
The Package specification is as below,
AS---------------------------------------------------------------------------------- type definitionTYPE CURSOR_TYPE IS REF CURSOR;-- Get the targets for this step, if any-- if p_return_display_names is false, return the target's internal nameFUNCTION get_step_targets ( p_step_id NUMBER, p_return_display_names BOOLEAN DEFAULT true ) RETURN SMP_EMD_STRING_ARRAY;-- Get the targets for this step, if any, as a comma separated string-- if p_return_display_names is false, return the target's internal nameFUNCTION get_step_targets_str ( p_step_id NUMBER, p_return_display_names BOOLEAN DEFAULT true ) RETURN VARCHAR2;-- Get the parameters for this job, filter as specified for this jobtypePROCEDURE get_visible_params ( p_job_id RAW, p_exec_id RAW, p_params_out OUT CURSOR_TYPE);-- Get the URI for this uri_use-- see emSDK/job/dtd/UriSource.java for uri_use constantsFUNCTION get_display_uri ( p_job_type IN VARCHAR2, p_uri_use IN
find out the list of referenced objects which are dependent on INVALID objects.
I wrote the below query to get the referenced objects which depends on invalid objects.
select owner,name,type, referenced_owner,referenced_name, referenced_type, referenced_link_name , dependency_type from dba_dependencies where type not in ('JAVA CLASS') AND referenced_type NOT IN('JAVA CLASS') AND (NAME, TYPE) IN ( select object_name, OBJECT_TYPE from dba_objects WHERE STATUS='INVALID')
And the output is something like below -
Sr. No OWNERNAMETYPEREFERENCED_OWNERREFERENCED_NAMEREFERENCED_TYPEREFERENCED_LINK_NAMEDEPENDENCY_TYPE 1PUBLICVEHICLE_INSPECTION_REQUESTSYNONYMINSBIN$Uu99fysmRj6Ppn2QppCTWg==$0TABLEHARD 2PUBLICPRODUCT_HIERARCHY_MAP_TEMPSYNONYMCONFSYSBIN$ndGddLcKSDWRwsn5g91Rcg==$0TABLEHARD 3PUBLICACPKG_SUB_RECEIPTINGSYNONYMINSACPKG_SUB_RECEIPTINGPACKAGEHARD
[code].....
Our requirement is to drop all the invalid objects. But we need to know whether it effecting to any other valid objects?Do any valid objects gets INVALID after dropping of the INVALID objects. We need to know the hierarchy of it.
I have a problem some user modified the objects and now objects became invalid .How to trace the ip address and operating system username and service ip who modified the objects?
I am looking at a performance issue at the moment and trying to replicate on a test system. I am initially looking at the impact of upto-date statistics on the main schema's objects.
For this I wanted to:
first run the batch with whatever stats were present in the database Flashback the db to before the batch . Gather stats Re-run the batch with updated stats and compare results.
However, I inadvertently ran the stats job before running the load the first time! I have the SCN from when the environment was set up like production (ie before the stats were run) so am I correct in saying that if I flashback to this point then the stats will be "old" and I can just run the batch then? I know I can verify this when I Flashback the database by looking at LAST_ANALYZED on tables etc but it would be good to know this before hand as it's a 12 hour batch.
