Re: bzr commit into mysql-5.1-bugteam branch (davi:2728) Bug#40536

[Davi Arnaut]

On 1/15/09 6:50 AM, Sergei Golubchik wrote:
> Hi, Davi!
>
> On Jan 13, Davi Arnaut wrote:
>> On 1/13/09 3:57 PM, Sergei Golubchik wrote:
>>> On Jan 06, Davi Arnaut wrote:
>>>>    2728 Davi Arnaut	2009-01-06
>>>>         Bug#40536: SELECT is blocked by INSERT DELAYED waiting on
>>>>                    upgrading lock, even with low_pri
>>>>
>>>>         The problem is that there is no mechanism to control whether a
>>>>         delayed insert takes a high or low priority lock on a table.
>>>>
>>>>         The solution is to add a new global variable which specifies
>>>>         whether a delayed insert thread should take high or low priority
>>>>         locks. The name of the variable is low_priority_delayed_updates.
>>> I think this new feature is uncalled for.
>>> Why don't you use low_priority_updates ?
>> Because the session value of the variable wouldn't be meaningful, which
>
> Not necessarily.
> When a delayed insert is created it could take the value of
> low_priority_updates from the thread that issued INSERT DELAYED.
>
> I think this is a desired behaviour. But my preference isn't too strong,
> and if it'd be difficult to implement, feel free to use the global value
> of low_priority_updates.

It's not that difficult as we could pass it via the Delayed_insert 
class, but I'm not fond of exposing racy behavior. I've sticked to the 
global value of low_priority_updates.

-- Davi