Re: Single stack optimistic lock exception?
Re: Single stack optimistic lock exception?
- Subject: Re: Single stack optimistic lock exception?
- From: Ken Anderson <email@hidden>
- Date: Fri, 24 Sep 2010 13:53:16 -0400
On Sep 24, 2010, at 12:58 PM, Miguel Arroz wrote:
> Hi!
>
> I never worked with Oracle before. Anyway:
>
> On 2010/09/24, at 17:50, Ken Anderson wrote:
>
>> The data type of the trans_id column is Number (12,0)
>
> Is that an integer or a float? If it's float, then there's your problem. :)
No, been around WO since 1995 - it's an integer.
>
>> It's Oracle RAC, so yes, a cluster.
>
> Could it be that you are writing a row to one server and reading the same row immediately after from another server that still didn't receive the update?
I'm pretty sure the way Oracle RAC works, I have a single connection, but it is a very interesting point. If the first update hit instance 1, then I hit instance 2 with my second update before the first one got sent over to sync up, that would cause this exact problem.
>
>> Yes, no way for someone else to change those rows, since we audit the tables with a trigger, and they weren't modified by anything else.
>
> Could it be that your trigger is somehow running after the transaction commits and doing something nasty to the transaction_id?
The trigger only copies rows into the audit table, so no.
Ken
>
> Regards,
>
> Miguel Arroz
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Webobjects-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden