Subject | Re: [IBO] deadlock error on IBO /FB 1.02 - what may be the reason ? |
---|---|
Author | Helen Borrie |
Post date | 2003-12-04T15:43:45Z |
At 03:34 PM 4/12/2003 +0000, you wrote:
on the transaction instead of a hard commit?
row. Note *transaction*, not *user*. A single user can get lock conflicts
when attempting to update the same row from two different transactions.
Helen
>Hi, all !Do you actually COMMIT the transaction?
>
>I have this huge problem with a rewritten application. I'll try to
>explain
>what's happening and what's done inside the application:
>
>1. Logging on to server
>2. Checking table USER_SETTINGS to check it's a valid user, and if
>found
>valid:
> 2.b) Load user settings.
> 2.c) Write Logindata (machine, 'Now') to this record in an UPDATE
>statement.
>3. Close transaction
>4. Open TIBODataSet in a grid (only first XXX rows are shown. AHow do you verify that the commit occurs? Are you *really* calling Close
>ReadOnly
>dataset.
>5. (User uses application)
>6. When application is closing, LoginData is written, now with NULL
>for all
>values written upon startup.
>
>The error:
>When I start 2 instances of this application with the same user name,
>I get
>the "deadlock" error when closing app nr. 2. No matter if it was
>started
>first or last. I am fairly sure all writings to table USER_SETTINGS is
>committed immediately.
on the transaction instead of a hard commit?
>I know the problem is within writing the last table, because the errorIf you call Commit and another transaction has work pending on the same
>doesn't appear if I write to another record here.
>
>Questions:
>- How can this happen ?
>- Is there a way I can *not* have the Commit; actually _commit_ when
>called
>explicitly ?
row. Note *transaction*, not *user*. A single user can get lock conflicts
when attempting to update the same row from two different transactions.
Helen