Re: BUG #12071: Stat collector went crasy (50MB/s constant writes)

On Wed, Nov 26, 2014 at 4:15 PM, Alvaro Herrera <alvherre@2ndquadrant.com>
wrote:

> maxim.boguk@gmail.com wrote:
>
> > Hi,
> >
> > On one of my servers stat collector went crasy on idle system.
> > It constantly write around 50MB/s on idle database (no activity at all)=
.
> >
> > What I tried:
> > 1)stop/start the database - no effect
> > 2)update to 9.2.9 from 9.2.7 - no effect
>
> Is the server time correct?  What is autovacuum doing


Server time look correct for me.
=D0=90utovacuum doing nothing since no activity on server at all (I checked=
 it
twice).

After careful examination of strace output I found one strange thing:
constant flow of:
16:22:33.202523 read(3, 0x7fff4148a5b0, 16) =3D -1 EAGAIN (Resource
temporarily unavailable)
16:22:33.297432 read(3, 0x7fff4148a5b0, 16) =3D -1 EAGAIN (Resource
temporarily unavailable)
16:22:33.322471 read(3, 0x7fff4148a5b0, 16) =3D -1 EAGAIN (Resource
temporarily unavailable)
16:22:33.417873 read(3, 0x7fff4148a5b0, 16) =3D -1 EAGAIN (Resource
temporarily unavailable)
16:22:33.456801 read(3, 0x7fff4148a5b0, 16) =3D -1 EAGAIN (Resource
temporarily unavailable)
But I not sure that is actual error signal or normal behaviour.

--=20
Maxim Boguk
Senior Postgresql DBA
http://www.postgresql-consulting.ru/ <http://www.postgresql-consulting.com/=
>