Subject | Re: [firebird-support] Lost transactions that causes deadlocks |
---|---|
Author | Ann W. Harrison |
Post date | 2005-06-22T18:21:31Z |
Peter Ypenburg wrote:
different problem from two users trying to change a normal record.
You've created a metadata object and want to replace it, but it's
visible to other users. That's one reason why creating metadata objects
as part of normal application processing is a bad idea.
with a different name OPEN_ITEMS_7A for example. Firebird has no
concept of "private" metadata - which is what you're tying to create -
and any change to let you drop procedures that others can see is going
to allow other, more serious problems.
persistence is not going away. Though we might try to add a better message.
Regards,
Ann
>OK. That's a bad (well, misunderstood) error message, and quite a
>
> The problem is that sometimes from creating the stored proc to selecting
> data from it something goes wrong on the client side (e.g. user lost
> connection). Now when he opens the program again and redoes the process and
> at the point the code tries to create the stored proc again the following
> error gets thrown:
> Lock conflict on no wait transaction unsuccessful metadata update object
> OPEN_ITEMS_7 is in use.
different problem from two users trying to change a normal record.
You've created a metadata object and want to replace it, but it's
visible to other users. That's one reason why creating metadata objects
as part of normal application processing is a bad idea.
>One possibility is to recognize the error and create a new procedure
> This error (I think) is because there is still a transaction alive
> implicating the stored proc and thus the error occurs. The question is about
> how do we fix this.
with a different name OPEN_ITEMS_7A for example. Firebird has no
concept of "private" metadata - which is what you're tying to create -
and any change to let you drop procedures that others can see is going
to allow other, more serious problems.
> To answer your question, yes we are seeing some other behavior. I read onMaybe, maybe not. The problem of metadata visibility and the need for
> the FB list that there is a timeout for transactions build into FB 2, and
> that's the kind of thing that will really make a big difference in our
> lives.
persistence is not going away. Though we might try to add a better message.
Regards,
Ann