Re: Row Locking on Oracle 8

Thanks a lot. We wanted to check if we missed out something but we see that that is the only way. Again, thaks for the answer.
Carlos.

Jeremiah Wilton wrote:

> Juan Carlos Alonso Lafuente wrote:
>
> > What we would like to do is to show the user some kind of message window telling the user
> > that such and such person is editing the document (but without blocking him/her.) In this
> > case, we can not use the waiters sql (we have already taken a look of the oracle's
> > scripts :utlockt.sql and catblock.sql)
>
> I understand the problem you are up against. You use "...for update nowait;" statements to
> prevent locking pileups, but that gives you no ability to detect the identity of the blocker.
>
> Why not get rid of the nowait clause, and program the application with a timeout that checks
> for blocking (and even other types of waits). This could be performed in a separate process
> or thread, and might improve lock success rates because it gives the blocker a few seconds to
> get out of the way before the waiter fails. After the timeout period, the timeout checker
> thread could identify the blocker's identity and print an appropriate message to the
> application UI.
> --
> Jeremiah
Received on Wed Sep 22 1999 - 12:27:05 CDT