Re: [COMMITTERS] pgsql: Provide much better wait information in pg_stat_activity.

On Thu, Mar 10, 2016 at 5:07 PM, Robert Haas <robertmhaas(at)gmail(dot)com> wrote:
> On Thu, Mar 10, 2016 at 5:05 PM, Robert Haas <robertmhaas(at)gmail(dot)com> wrote:
>> On Thu, Mar 10, 2016 at 4:51 PM, Pavel Stehule <pavel(dot)stehule(at)gmail(dot)com> wrote:
>>> Maybe it be clear from attached text file
>>
>> Uh, yikes, that looks messed up, but I wouldn't have thought this
>> commit would have changed anything there one way or the other. The
>> current query is reported by pgstat_report_activity(), which I didn't
>> touch. I think.
>
> I just tried this on 9.5 - changing the query only to "select pid,
> state, query from pg_stat_activity" and doing everything else the
> same - and I see the same behavior there. So it looks like this is a
> preexisting bug.

Or ... maybe this is intentional behavior? Now that I think about it,
doesn't each backend cache this info the first time its transaction
reads the data?

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company