Hi,
I am not lucky with preventing duplicate bugs in IdBroker by
synchronisation, because I am not sure whether this works in distributed
environments (several clients accessing the same database). But I am not
sure about this, I did not look into the code to see whether IdBroker is
supposed to work in distributed environments at all.
So I would like to find out how the duplicate ids can happen before I apply
any patch. However, if you do not use more than one client, synchronizing
the method should be fine.
Thomas
David Ware <[EMAIL PROTECTED]> schrieb am 20.05.2005 03:10:27:
> Hi,
>
> I've been having problems with IDBroker giving duplicate keys in long
> running transactions. To make sure I made a stress test situation where
> in a loop I'd simply create a new row object and save it. With a loop
> size of 200000 this would usually result in the error occuring. Using
> P6Spy I verified that it was trying to insert the same ID on 2 different
> objects sequentially.
> Here's an some info I got from spy.log from the last time I ran it, this
> shows the last 2 times ID_TABLE was accessed/changed, and which ID was
> duplicated:
>
> ID_TABLE updates: QUANTITY = 39496
> NEXT_ID = 62913
> primary key of next insert : 23417
> ...
> ...
> ID_TABLE updates: QUANTITY = 39496
> NEXT_ID = 102409
> primary key of next insert : 28676
> ...
> ...
> attempts to insert primary key 62912 twice.
>
> This is using hypersonic database. This seems to be solved by
> synchronizing the 'storeIDs' method of IDBroker as suggested in issue
> TRQS296.
>
> David Ware
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
>
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
