Adrian Crum wrote:
> --- On Sat, 3/13/10, Adam Heath <[email protected]> wrote:
>> Adrian Crum wrote:
>>> --- On Fri, 3/12/10, Adrian Crum <[email protected]>
>> wrote:
>>>> Consistent random foreign key violations. Two
>> records would
>>>> get the same sequence number. Random in that
>> different
>>>> entities would be affected each time, and
>> consistent in that
>>>> it happens nearly every time.
>>>>
>>>> Try running five or more threads against it on a
>>>> multi-processor machine.
>>> Oops, maybe those were primary key violations. Bottom
>> line is, an exception was thrown because two records got the
>> same sequence ID.
>>
>> I've created a test case that runs 100 threads, all trying
>> to get 1000
>> sequence values.  I then verify that I get 100,0000
>> different values,
>> and everything passes.
>>
>> Could it be that your multi-threaded wrapper ends up using
>> different
>> random delegators?  This would say that the locking
>> issues lie in the
>> retry loop, which isn't being run by my multi-threaded
>> code, since the
>> bank in each thread is the same, and synchronized handles
>> that
>> concurrency.
> 
> I modified the demo data loading code to make it multi-threaded, so I believe 
> only one delegator is being used. I have the patch at work, so I can't take a 
> look at it until Monday.

Heh.  I run screen, so when I go home, I can just connect back to
work, and have everything immediately available.

All remote machines I connect to I immediately install screen.  So,
when I have to continue client work from home, I can also resume
wherever I was last at.

My laptop at home has been logged in to work since Feb. 21, without
disconnecting.

Could it be the sub sequence logic that is broken?

Reply via email to