ORA-04068 - Existing State Of Packages Has Been Discarded
Feb 4, 2011
We are experiencing an issue with ORA-04068: existing state of packages has been discarded. We're running Oracle 10.2.0.4 on Solaris 10.
I'll provide the timeline of what occurred:
1.) Last night during a maintenance window I recompiled a package and package body. I ran "create or replace" statements from a script via sqlplus to recompile both package and package body. Because I am required to do all DDL under my own login (and not the schema owner login) the scripts require that the schema name be included in the "create or replace" statements. Otherwise an object will be incorrectly created in my schema. This is what happened with the package body. I discovered this when the package body would not compile.
2.) I ran a "drop package body" statement to drop the package body from MY schema.
3.) I added the proper schema name to the script and recompiled both the spec and body under the proper schema.
4.) I ran a script to check for any invalid objects under the application user schema. There were none.
5.) During the night a batch process ran which executed the package. There were several failures with the error ORA-04068: existing state of packages has been discarded.
6.) During the early AM today the error stopped. Not knowing this at the time and believing that the issue was still occurring I began searching the internet for information regarding this issue. One site I found said this could be due to DDL being done on the object (which was the case for us) and there could be an entry in the shared pool marked as invalid which our application's session may be accessing. It recommended running "alter package recompile" for the spec and body. I did this but it seems to have triggered the event again.
I also ran utlrp.sql.
I found some information suggesting the use of DBMS_SESSION.MODIFY_PACKAGE_STATE and/or RESET_PACKAGE but I'm not seeing mention of the risks/implications.
I can verify that the object is no longer in my schema and that the proper object owner has no invalid objects. I am not sure what to do to resolve the issue or keep it from occurring again.
I am trying to use dbms_jobs in Oracle 8i database.
This is the syntax I am using:
begin dbms_job.submit(:jobno, 'SYSTEM.SHRINK_ROLLBACK_SEGS', '07-FEB-2007', NULL); commit; end;
i have defined the variable jobno as number!
And i am getting the error :
* ERROR at line 1: ORA-04068: existing state of packages has been discarded ORA-04063: package body "SYS.DBMS_JOB" has errors ORA-06508: PL/SQL: could not find program unit being called ORA-06512: at line 2
I then go and look at the Sys.dbms_job package and see its not compiling
I have this error:
(1): PLS-00201: identifier 'ANY_INSTANCE' must be declared (2): PL/SQL: Item ignored (3): PLS-00801: internal error [21076] (4): PL/SQL: Item ignored
I am getting the following error only for the ""first execution"", even the package is in "valid" state. If the execute the same package again "second execution", it runs ok.
usually we complie all the "invalid objects" in the schema when we move the code to TEST or PRE-PROD env.
but i am getting the following error, even if the package is in "valid state".
ERROR at line 1: ORA-04068: existing state of packages has been discarded ORA-04061: existing state of package "TEST_OWNER.TEST_PKG" has been invalidated ORA-04065: not executed, altered or dropped package "TEST_OWNER.TEST_PKG" ORA-06508: PL/SQL: could not find program unit being called: "TEST_OWNER.TEST_PKG" ORA-06512: at line 2I have gone through the following thread, blu explained the cause of this error, but do we get this error even when the package is in "valid" state too?
I'm am getting an error that says that a role doesn't exist when I can cleary see it in DBA_ROLES. I can successfully grant other roles in this session. I am using Oracle 11.1 on RAC.
GRANT "CONNECT" TO BAKERD Error at line 1 ORA-01919: role 'CONNECT' does not exist select * from dba_roles where role = 'CONNECT'
I'm not sure if this belong to this place. move it to the proper place.
I'm having an annoying problem: (I'm not sure if it is a problem) - In a package body I'm trying to delete/update/insert/select the contents of a table in other schema. - The respective synonym exists. - I had created a role with the respective privileges over the synonym. - I granted the role to the package's owner. - I try to compile the package, but it keeps returning compilation errors. (Not table found) - In standard SQL, I can delete/insert/update/select over the table. - The only way to compile the package, without errors, is to grant the privileges directly to the package's owner.
Anand basic Anand pascal Juliana cobol Kamala dbase Mary oracle Mary c Partick c++ Qadir assembly Qadir c Ramesh dbase Remitha c Remitha assembly Revathi pascal Revathi basic
Wanted to know whether it is possible to know a trigger is getting fired due to which all packages/procedures. If it has got fired for a particular transaction say multiple times, then wanted to get details of the packages/procedures which caused an event so that trigger got fired?
i dont have access to any Oracle XML related pacakges, however I have to pull the data from the following xml.Eg: The alternativeIdentifier.Name should be a column and the value should be the data for the column. Same applies to other tags under characteristics tag.
we have a database application which is done frequently.in these we load data throught Sql loader, we create an DB instances, we do several DML operation on the database.
now for such task in an application we need to keep an logging track of each task performed in PL/SQl procedure packages.
I am really wondering about data loading to the cloud, I've just read the documentation, but regarding the packages, I have more than 100 package and I have to "add to cart" one by one the problem is that the package body is not added by default, so you have to expand the packages one by one then select the package body right click and add to cart. This task is very tedious, is there any other options.”
Tried to use the yum command "yum install gcc" but get the following error. Error: cannot find a valid baseurl for repo:update
I think this is due to the "CentOS-Base.repo" file, which I've tried to update by changing the mirror URL's but no success.Then tried to find the rpms online ([URL]...
i want to do a schema export from Database A. There are hundreds of users under this schema.I have to import this schema into other database say B. My question's are:
1) Do i need to pre-create only schema user or all the users under it.
2) Will the schema export all the roles,procedures,packages,synonyms,funsctions and triggers?
2437312:ORA-12801: error signaled in parallel query server P009 2437313:ORA-01502: index 'POS.XIETBK_POS_FACT_TRAN_DATE' or partition of such index is in unusable state
and tried to rebuild the index and i got following error.
ORA-14086: a partitioned index may not be rebuilt as a whole.
The table size for the index is large. we need to rebuild the hole index.
1. when the oracle session changed from active to inactive? 2. what is the time for active session? 3. session is changed to inactive from active. but it is still showing in v$session. 4. in v$session, can i see the ipaddress of client machine ?
When i am writing a function (Stand alone or inside a Package) i know what i am writing, i know weather i am changing a Package state or weather i am changing a Database State. So what is the use for giving a PRAGMA RESTRICT_REFERENCE?
as for the other PRAGMAs EXECPTION_INIT is needed to Specify an error name to a specifiec error number so that i can use the error name to handle the exception. AUTONOMOUS_TRANSACTION is used to execute the SQL Operation inside a Block as a child Transaction.SERIALLY_REUSABLE states that a package variable doesnot persists throughout the session.
RESTRICT_REFERENCE states that the code should not do the following RNPS : Read no package state, WNPS : Write no package state RNDS : Read no database state WNDS : Write no database state
what is there to state as i know what my code is doing.
without specifing EXECPTION_INIT, AUTONOMOUS_TRANSACTION or SERIALLY_REUSABLE i cannot get a handler for an exception which doesnot have a handler, i cannot execute SQL Operations from inside the CODE autonomously or i cannot reuse the package variable, but without the RESTRICT_ REFERENCE is can ensure that my code in not doing (RNPS,WNPS,RNDS,WNDS).
I understand the architecture of solid state drives favor different types of I/O. Oracle 11g introduced the database flash cache which allows flash storage to be used as a secondary cache. With database flash cache serving as a secondary cache to my Oracle buffer cache.
My understanding is I need to config two database parameters DB_FLASH_CACHE_FILE and DB_FLASH_CACHE_SIZE. If i want to have certain tables/indexes to take advantage of this extra cache I would set the value KEEP within the parameter.
Is there anything that needs to be done to the tablespaces / datafiles.
so here is the query : For every state, list the most popular product. Here is the query I have so far :
SELECT C.State, P.Product_Name, Count(*) cnt FROM Product P, Customer C, OrderTable O, LineItem L WHERE C.CID = O.CID AND O.OID = L.OID AND L.PID = P.PID GROUP BY C.State, P.Product_Name;
Result : STATE PRODUCT_NAME COUNT(*) ---------- -------------------- ---------- New Jersey Computer 3 Texas Computer 1 New Jersey Speaker 2
I would need the result to only say New Jersey Computer and Texas Computer because I only want a list of the states with the product name that is sold the most in each state.All I need to do is have the query only select the product name with the max count for each state...
I started database in mount state;startup mount; alter tablespace users offline;
I tried to make tablespace offline In another case online but showing error database not open same for command alter tablespace begin backup; do think we can make tablespace offline in mount...
after merging two partitions into single partition (partition is by list) of a table ,when i analyzed the table it is giving this error : ORA-01502 INDEX TEST.PK_ID or partition of such index is in unusable state.
The function is supposed to tell whether numbers stored as a string value evaluates to a date.I copied into an SQL windown in PL/SQL Developer and typed "commit;" after and then tried to view the results as I posted below to make sure the function worked.
However, I get an ora-06575 Package or Function GETDATE is in an invalid state.So would like to know first, if the code is correct? And if it is, do I need to do something else to make Developer take the code? Here is the code that gave me the error I used to try to test the function:
select GetDate ('010312', 'dd-mm-yyyy') from dual
Here is the code I used to create the function:
create or replace function GetDate (vDate varchar2, vMask varchar2) return varchar2 is
vReturnDate date; begin vReturnDate = to_date(vDate, vMask); return vReturnDate; exception when value_error then return 0; end GetDate; commit;
At a test database instance of Oracle 11g (11.2.0.3.0) at CentOS 6 i've imported a tablespace from another instance to test my backup strategy.
First i've done a fullbackup with RMAN: ---------- # rman target / RMAN> BACKUP DATABASE PLUS ARCHIVELOG DELETE INPUT; ---------- Then i thought to test the database i could delete something and restore and recover the database from the backup. So i deleted a table in the imported tablespace: ---------- # sqlplus / as sysdba SQL> drop table TESTTABLE; ---------- After that i would start the restore and recover process by restoring and recovering the tablespace in which I deleted the table with RMAN. At first i've set the tablespace offline: ---------- SQL> alter tablespace TESTTABLESPACE offline immediate; ---------- Then I go to RMAN and start the restoring: ---------- RMAN> restore tablespace TESTTABLESPACE; ---------- Now there is a problem when i execute the recover command. The recover command includes the recovering from the redolog files, right? As I dropped the table from the tablespace this operation was written in the redolog files. So when i make a : ---------- RMAN> recover tablespace TESTTABLESPACE; ---------- The dropped tablespace would be dropped also, because the last redolog file told RMAN to drop the table. So i think I have to restore and recover my tablespace from an older timestamp, so RMAN would ignore the dropping of the table in the redolog file. Is that right so?
And when it is right how can i restore a tablespace from an older date so RMAN would ignore the dropping?
