Inability to "Kill" a Process

asacin

Member
I have noticed that whenever there I am debugging and I happen to hit an infinite loop in a stored procedure, PL/SQL Developer stops responding and needs to be killed in the client. However, this is not all, I have to go to the Oracle server and kill the processes every time. Is there a way that I can stop debugging without having to kill PL/SQL Developer and in addition the generated process in the server?
 
Can you let me know your Oracle Server version and your PL/SQL Developer version?

------------------
Marco Kalter
Allround Automations
 
!!! My place has experienced this too, but not just with the PL/SQL developer program.

We have opted to kill sessions which are not SQL Plus sessions that have been inactive for 30 minutes.

We run a 8.1.7.0.0 client and 9.2.0.3.0 server.

------------------
Kevin Ryan

Senior Computer Programmer / Analyst
 
The versions that we are using here are:
Oracle Server: 8.1.7.4.1
PL/SQL Developer: 5.0.1.480

I hope this gives you a hint on what is going on.
 
I have also found that even when the process gets killed, the stored procedure object remains locked and cannot even be recompiled or deleted by the Oracle DBA until 30 minutes or so have passed!
 
our version number is 5.1.2.682. What is the most current version, and why would you instruct someone to go to 5.0.3 in June when you have at least a 5.1.2xxx version out there?

Also, we have the same problem as "asacin" where even after our process is killed, it doesn't go away for 30+ minutes and on 4 occasions, I had to shut down the database for the process to go away (well, actually abort -shut down immediate just hung).

What version do you have where this is issue is fixed?
 
our version number is 5.1.2.682. What is the most current version, and why would you instruct someone to go to 5.0.3 in June when you have at least a 5.1.2xxx version out there?
The latest version is 5.1.3. I advised asacin to upgrade to 5.0.3 because he was using 5.0.1 and I cannot assume that he is entitled to 5.1, plus the fact that there were no debugger stability improvements between 5.0.3 and 5.1.2.
Also, we have the same problem as "asacin" where even after our process is killed, it doesn't go away for 30+ minutes and on 4 occasions, I had to shut down the database for the process to go away (well, actually abort -shut down immediate just hung).
Please upgrade to 5.1.3 and verify if this helps. The 5.1.3 patch contains a fix for debugger hang-ups, though I'm not sure if this is the same issue.

------------------
Marco Kalter
Allround Automations
 
I'm wondering if this is really a PL/SQL Developer issue.

Asking around a bit, it seems normal for a process to keep going for some time after the database session is killed and I've noticed this myself regardless of the tool I'm using.

To get around this, I have looked up the process id on the UNIX server and killed it so I can re-compile a package that is locked up. That sounds like what others in this discussion are doing anyway, but wish that PL/SQL Developer would handle it. I'm sure that you really want to make sure you have the right process id doing it that way, but it does release the lock on the package right away.
 
Hi asacin,

Are you sure that you only see this behaviour in PL/SQL Developer, and not when you run the same code in SQL*Plus?

We see this problem from time to time (Oracle 8.1.7 for Windows NT), and have to find and kill the thread concerned.

Regards, Mark.
 
I have tried the newest version, 5.1.3.704 and I have to tell you with glee that it the Session->Break does work now. It took about 10 seconds to kill the process.

As a reponse to MarkP I have to say that with SQL* Plus, I did not experience this problem. I was able to recompile after getting in an infinite loop. All I needed to do back then was close SQL* Plus and start again. I have Oracle version 8.1.7.
 
Back
Top