On Fri, Mar 11, 2016 at 7:45 PM, Peter Geoghegan <pg@heroku.com> wrote:
> On Fri, Mar 11, 2016 at 4:30 PM, Jim Nasby <Jim.Nasby@bluetreble.com> wrote:
>> Right, but you still have the option to enable them if you don't want to
>> swamp your IO system. That's why CIC obeys it too. If I was running a
>> consistency check on a production system I'd certainly want the option to
>> throttle it. Without that option, I don't see running this on production
>> systems as being an option. If that's not a goal then fine, but if it is a
>> goal I think it needs to be there.
>>
>> Isn't it just a few extra lines of code to support it?
>
> I see your point.
>
> I'll add that if people like the interface you propose. (Overloading
> the VACUUM cost delay functions to cause a delay for amcheck
> functions, too). Note that the functions already use an appropriate
> buffer access strategy (it avoids blowing shared_buffers, much like
> VACUUM itself).
I don't particularly like that interface. I also suggest that it
would be better to leave throttling to a future commit, and focus on
getting the basic feature in first.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company