Re: Scalability in postgres

On Thu, May 28, 2009 at 12:50 PM, Fabrix <fabrixio1(at)gmail(dot)com> wrote:
>
> HI.
>
> Someone had some experience of bad performance with postgres in some server
> with many processors?

Seems to depend on the processors and chipset a fair bit.

> I have a server with 4 CPUS dual core  and gives me a very good performance
> but I have experienced problems with another server that has 8 CPUS quad
> core (32 cores). The second one only gives me about 1.5 of performance of
> the first one.

What model CPUs and chipset on the mobo I wonder?

> Monitoring (nmon, htop, vmstat) see that everything is fine (memory, HD,
> eth, etc) except that processors regularly climb to 100%.
>
> I can see that the processes are waiting for CPU time:
>
> vmstat 1
>
> procs -----------memory---------- ---swap-- -----io---- --system--
> -----cpu------
>  r  b   swpd   free   buff  cache   si   so    bi    bo   in   cs us sy id
> wa st
>  0  0      0 47123020 117320 17141932    0    0     8   416 1548 2189  1  1
> 99  0  0
>  0  0      0 47121904 117328 17141940    0    0     8   148 1428 2107  1  1
> 98  0  0
>  0  0      0 47123144 117336 17141956    0    0     8   172 1391 1930  1  0
> 99  0  0
>  0  0      0 47124756 117352 17141940    0    0     8   276 1327 2171  1  1
> 98  0  0
>  0  0      0 47118556 117360 17141956    0    0     0   100 1452 2254  1  1
> 98  0  0
>  2  0      0 47120364 117380 17141952    0    0     8   428 1526 2477  1  0
> 99  0  0
>  1  0      0 47119372 117388 17141972    0    0     0   452 1581 2662  1  1
> 98  0  0
>  0  0      0 47117948 117396 17141988    0    0    16   468 1705 3243  1  1
> 97  0  0
>  0  0      0 47116708 117404 17142020    0    0     0   268 1610 2115  1  1
> 99  0  0
>  0  0      0 47119688 117420 17142044    0    0     0   200 1545 1810  1  1
> 98  0  0
> 318  0      0 47116464 117440 17142052    0    0     0   532 1416 2396  1  0
> 99  0  0
> 500  0      0 47115224 117440 17142052    0    0     0     0 1118 322144 91
> 5  4  0  0
> 440  0      0 47114728 117440 17142044    0    0     0     0 1052 333137 90
> 5  5  0  0
> 339  0      0 47114484 117440 17142048    0    0     0     0 1061 337528 85
> 4 11  0  0
> 179  0      0 47114112 117440 17142048    0    0     0     0 1066 312873 71
> 4 25  0  0
>  5  1      0 47122180 117468 17142028    0    0   192  3128 1958 136804 23
> 2 75  1  0
>  3  0      0 47114264 117476 17142968    0    0   608  5828 2688 4684  7  2
> 89  2  0
>  0  1      0 47109940 117484 17142876    0    0   512  5084 2248 3727  3  1
> 94  2  0
>  0  1      0 47119692 117500 17143816    0    0   520  4976 2231 2941  2  1
> 95  2  0
>
>
> Have postgres problems of lock or degradation of performance with many
> CPU's?
> Any comments?

Looks like a context switch storm, which was pretty common on older
Xeon CPUs. I imagine with enough pg processes running on enough CPUs
it could still be a problem.