Daemon News Ezine BSD News BSD Mall BSD Support Forum BSD Advocacy BSD Updates

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Horrible PostgreSQL performance with NFS

--- Slawek Zak <slawek.zak@xxxxxxxxx> wrote:
> On 1/13/06, Arne Woerner <arne_woerner@xxxxxxxxx> wrote:
> 40MB/s. CPU load negligible. I don't have
> an exact number, as this machine has other
> processes running. But overall, the system
> load didn't exceed 5%.
Looks good...

> I saturated fast ethernet on the host
> with this test. Filer is connected with
> Gb and can spew around 70MB/s easily.
> CPU load on the host didn't exceed 4%.
Looks even better... :-)

> > 3. test the NIC performance with
> Filer doesn't respond to large icmp packets.
Ok... This isn't so important, since NFS speed is higher than
local disc speed.

> > My theory would be, that your NICs need a
> > lot of CPU time, while your local discs
> > dont need so much CPU time. :-)
> I don't think so. Drivers account for system
> time. It doesn't exceed 20% of overall load.
> The postgres processes are very busy doing
> almost nothing. Semops is most of the work
> they seem to do.

But why does switching from local disc to NFS makes the PostgreSQL
performance so bad?

The semops of postgres are most likely the same, since postgres
would not check if it runs on NFS or local filesystems, would it?

Were there any other changes?

Did you do those "dd" tests with small block sizes (like 1byte:
bs=1), like somebody on one of those lists suggests, too? Then we
could see, if there is a high latency that ruins everything...


Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around