Go Premium for a chance to win a PS4. Enter to Win

x
  • Status: Solved
  • Priority: Medium
  • Security: Public
  • Views: 715
  • Last Modified:

How to gracefully kill an update command that stalls

Hi,
I was trying to do an update to a table that gets data from one table and inserts it into another table.

There are about 40,000 rows and although it's an old UNIX server an hour seems like a long time to wait for the update.

How do I stop the update without damaging the data in the tables?
Jerry
0
geraldw
Asked:
geraldw
  • 3
  • 2
  • 2
  • +1
1 Solution
 
morphmanCommented:
if its running under an sqlplus session, just ctrl+c and it will rollback, might take just as long to undo though.

Alternativeley, just kill the oracle process (not the sqlplus session, the actual Oracle process which should be a parent child process of the sqlplus session.

if you are unix savvy, you will already know how to do this, otherwise perform the following.

ps -ef | grep sqlplus

find the child parent process for sqlplus, (it will be the left of the 2 numbers from the above command).

THen do a ps -ef | grep nnnnn

where nnnnn is the process number you just got from sqlplus.

then just do

kill nnnnn

and it will gracefully kill the processes related to the oracle process including sqlplus.
0
 
vc01778Commented:
Just kill the session that performs the update and the changes will be rolled back.  That is if you update with a single statement.

It will take approximately as much time for Oracle to rollback the transaction as it did to get to the point where you are now,  though.

VC
0
 
morphmanCommented:
hint: note the time the oracle process started and make sure it coincides with the correct time you started the updates.
0
Independent Software Vendors: We Want Your Opinion

We value your feedback.

Take our survey and automatically be enter to win anyone of the following:
Yeti Cooler, Amazon eGift Card, and Movie eGift Card!

 
geraldwAuthor Commented:
Yes, I can kill the process but before I kill it is there any reason I should just leave it running?  I am trying to achieve some changes here.

I mean, is it possible that it's just taking that long?
Jerry
0
 
vc01778Commented:
It's possible that the process can complete any time.  To see if the transaction is actually doing something,  you can use this:

select used_ublk, status, start_time, log_IO, phy_io from v$transaction

USED_UBLK      STATUS      START_TIME      LOG_IO      PHY_IO
272      ACTIVE      04/02/04 16:14:51      28107      141

The 1st, 4th and 5th columns should be growing.

VC
0
 
n4nazimCommented:
Hi,

If the process is taking time.. check if u have indexes on the columns in the WHERE clause..

Say if u have

UPDATE tablea
SET columna = 'aa'
WHERE columnb = 'ddd' and columnc = 'ttt'

I checked this for more than 30000 records, if i dont have indexes on columnb and columnc, it takes hell lot of time.

But if i do have indexes, it takes less than a minute.

Try for yrself and let me know abt it...
Post yr sql here if needed..


Rgds
Nazim M
0
 
morphmanCommented:
Hi

sounds like your database needs a bit of tuning, as 40,000 updates on a simple query should not take an hour. if there is a complex query behind the updates, then I guess an hour is relative to the complexity. However, could it be that your updates are waiting on locks from other users updating the system?
0
 
n4nazimCommented:
Hi,

Any replies on this ..

Rgds,
Nazim M
0

Featured Post

Veeam and MySQL: How to Perform Backup & Recovery

MySQL and the MariaDB variant are among the most used databases in Linux environments, and many critical applications support their data on them. Watch this recorded webinar to find out how Veeam Backup & Replication allows you to get consistent backups of MySQL databases.

  • 3
  • 2
  • 2
  • +1
Tackle projects and never again get stuck behind a technical roadblock.
Join Now