SQL & PL/SQL :: DBA Pending Transaction In Database
Jun 16, 2011
application was not starting due to some pending transaction in database.Pending transactions were rolled back by DBA team.To avoid such situations what I thought was, having a job that will call a procedure that will monitor table status everyday and will send the mail. Now that job is working fine for no pending transactions in DBA_PENDING_TRANSACTIONS.
But now I am in doubt if someday there are PENDING TRANSACTIONs in the table DBA_PENDING_TRANSACTIONS, will SELECT * FROM DBA_ PENDING_ TRANSACTIONs query will work as normal or this whole process of monitoring table and sending mail will work fine?
I am tring to create a query about finding our pending purchuase requestions without PO, we have 4 tables, 2 for PRs & 2 for POs, you can not creat a PO without having a PR,the problem that I tried to find out the pending PRs for all projects and PRs but the query is taking so much time and computer is hanging , is their any way to generate it grouped by project number and PR number and include material in each PR, below is description of all tables related:
PRs table:
MASTER: -------------------- PRNUMBER TYPE MADEBY PROJECTNUM [code]....
I hv a situation where a webservice interacts with the database.
Here the webservice will first make a request to database for some operation but i dont want the database to commit changes in first request itself. A response will be sent to webservice further a second request will be sent to database for committing the changes. So can that be done?
Submitting transaction / master data through API without exposing DB credentials...I am developing a scenario where I will submit transaction / master data like sales order / master item to Oracle apps (EBS) through MS Excel via API / interface table but like to provide only apps (EBS) user credentials rather then DB credentials so DB password could not expose to normal end user.
I have a Data entry form which is a multirecord block;
Question: for example that form has 10 to 25 fields or columns more than that all the data has been entered, but before committing or saving that form i need to cross check the the data with a select query, whether the data entered is correct or not but before committing, that data it should be posted into that table if i find that one data is entered wrongly then i will modify that and again cross check and save the transaction permanently into the database table?
Find the date difference. I need to find that how many days the task is pending, if ACT_NAME field switching from 'SET PENDING%' to 'RESUME PENDING%' by using ACTIONTAKENDATETEXT field in the History table.
Example as needed: NoPendingDays = 23 (8+15)
I have attached Create table and Insert table values sample as SQL file.
Having PL/SQL procedure and calling sys dba procedure to alter rollback segment to online and after that to set the transaction to this rollback. The rollback altered to online but transaction still running on available rollback and not the one I was meant to. The user have executing grant on the sys dba procedure.
Oracle Database 10g Enterprise Edition Release 10.2.0.4.0 - 64bi PL/SQL Release 10.2.0.4.0 - Production CORE10.2.0.4.0Production TNS for Solaris: Version 10.2.0.4.0 - Production NLSRTL Version 10.2.0.4.0 - Production
We are running a code through informatica and are frequently getting the ORA-08007. The error and the code causing it are mentioned below. (The '?' implies a variable which is supplied by informatica). The source is a flat file.
ORA-08007: Further changes to this block by this transaction not allowed
1. Tried initially by reducing the commit interval to commit at 100 rows. But still getting the same error. 2. DBA has installed the patch # 6790768. But it still fails.
A B 1000 B C 2000 A D 3000 C A 3000 B A 4000 A B 5000 A B 1000 B C 2000 C A 3000 B A 4000 A B 5000 A C 10000
Dynamically I am giving user name A. I want to retrieve the data how much money user B taken from user A and how much money user B given to A lly for C and D.
CONDITION: I want to retrieve the records of 2 way transaction that means user A sent money to user B and user B must sent money to user A. I don't want one way transaction records i.e user A sent money to D but user A didn't get money from user D like this records.
OUTPUT LIKE:
USER MONEY_TAKEN_FROM_A MONEY_GIVEN_TO_A B 12000 8000 C 10000 6000
Query in format Query@DB_Link is sometimes dropping transactions. Move it to a package on target dababase "B" called from DATABASE "A" via package.procedure(varaibles);
Is that all I need to do to get the transaction to always go? I tried a commit in the package body, with repeated attempts if it fails. Btu when I use commit, the package doesn't work. (Compiles fine) Why would the commit fail?
My goal is to make that the transaction succeeds.Code for what I tried below. the COMMIT is commented out - if I uncomment it, the transaction fails. I have done this before calling an oracle SP from SQLServer and it worked
Comments:
The loop is to avoid an endless cycle..The concept is to perform the update, commit it, if the commit fails, rollback and try again.I'm not sure if I even need a loop, but I don't know if the unresolved transaction would get fixed by the remote DB, or dropped.
I have a job which runs the procedure below concurrently in 4 different sessions; It updates two tables.
PROCEDURE UPDATE_TAB_1_AND_TAB_2( traceIdTab IN pl_sql table, individualIdTab IN pl_sql table, fileSeqIdTab IN pl_sql table,
[Code].....
Each session is passed a sequence generated file ID; so the records processed in each session are completely different.
However, everytime this job runs, without failure, ONLY TWO sessions process concurrently, while the other two sessions are blocked. Once the first two have finished, then the remaining two sessions start work.
I have being monitoring the sessions on each run daily, and realized that the first two sessions processing are each holding a transaction exclusive lock on objects; hence blocking the other two sessions.
When I tried to find the objects that the two running sessions have a lock on, I cant seem to find anything by searching DBA_OBJECTS.object_id. However, when I look for the objects that are being waited on, its either a partitioned table or a partitioned index on either table_1 or table_2.
At first, I thought it was an ITL lock and after much analysis, I did not find anything which definitely pointed to an ITL lock. However, just to make sure, I increased the INITTRANS value from 2 to 5, but the problem did not go away.
At this point, I have applied all I know to understand the origin of the TX lock and how I can work around this situation, and decided to ask higher powers in Oracle for advise.
I am working with an oracle table that is populated by a trigger on another table.. So Table A is an audit of table B. The trigger also uses sysdate to populate a modification Date column on the Audit table.
I was using this modification Date column in a query interface to get changes that happened on the main table after a certain date/time.
The problem is that there is an application that uses transactions to write to table B and sometime this transaction may not be committed for over a minute so the modification Date is not a reliable way to query the table for changes after a certain time.
Is there a way to update the trigger/create a new one where the sysdate that gets written to the audit table is from when the transaction is committed, not when the transaction starts?
A Java program would call a sequence of 5 PLSQL procedures, and they all need to participates the Tranasction defined in Java.
What do I do for the transactions inside PLSQL to ensure all 5 pariticpate in the Java transaction? I suppose I just do not specify any COMMIT in the PLSQL, but that probably would not work as PLSQL would COMMIT by default after exiting the proc?
I have approximately 1200 transaction to be updated to a master table. There are other columns in the master table but only one column is being updated. I would like to use sqlloader if possible or any other efficient means. Those 1200 record is stored in an excel spreadsheet. The col1 of the excel spreadsheet have to match col1 of the master table inorder for update col2 from the excel spreadsheet. Here is an example of the data. My operation system is HPUX and database is Oracle 10g.
Master table col1 col2 col3 col 4 4238 susan 56e 5879 h698c rich 12g 7091 joyce 34b 0876 mike 25n 7501 k956b robert 87c 9498 angela 67r 3645 doris 92y
Consider a situation: There are two users 'A' and 'B' account number of A: 001234 account number of B: 002678
Suppose, if A wants to transfer $1500 to B's account
update cust_acc set balance = balance - 1500 where account_no = '001234'; [code]....
Assume that, oracle crashes somewhere between first commit and second commit.In this case, A's total money is deducted by $1500..But $1500 is not yet transferred to B's account.How will oracle keep track of these information?
sql statement to query a transaction table that stores transactions of items bought from my organisation.The report i would like to generate is one that lists the items bought and this should be grouped month by month.
The Procedure A extracts and filter some data from the DW, this data is stored on the Global Temporary Table. Another Procedure, named B, use the data from the Global Temporary Table and store it on a normal table using another procedure Named X that Merge the data from Global Temporary against the Normal Table (inserting if not exist and updating some fields if exist).
(X isn´t important on the new flow)
Now, i need to add some steps on the normal flow:
The Procedure A extracts and filter some data from the DW, this data is stored on the Global Temporary Table. Another Procedure, named B, use the data from the Global Temporary Table and store it on a normal table. Using the Data from Global Temporary Teble i must to Store some fields on another normal table, for this i use another Procedure named C that merge the data from Global Temporary Table against the data from normal table, and i must to commit at this point. X Merge the data from Global Temporary Table and the data from the Normal table con the procedure "C" against another Normal Table (inserting if not exist and updating if exist).
The issue that i´m expecting is that i can´t use "C" for merge and commit, because this truncate the data on the global temporary table. I can´t change the on commit delete rows option, because another procedures are using this Global Temporary Table on production.
Before you ask, i try using AUTONOMOUS_TRANSACTION on "C" and didn´t works because "C" can´t found data on the Global Temporary table ( i use a select count on "C" from Global Temporary), this is because The Autonomous_Transaction (i think). So, what i can do? I tried to trace the session Number on C and A and with the AT is the same ( so isn´t session change problem).
I need:Commit on the Procedure "C" without Truncating Global Temporary Table because i need this data for X.
