2008-06-30 09:58:28

by Krishna Kumar2

[permalink] [raw]
Subject: Re: NFS performance degradation of local loopback FS.

[email protected] wrote on 06/27/2008 07:36:44 PM:

> > But loopback is better than actual network traffic.
>
> What precisely do you mean by that?

Sorry I was not clear. I meant that the loopback will be better than
actual traffic between different server/client.

> You are testing with the client and server on the same machine. Is
> the loopback mount over the lo interface, but you mount the machine's
> actual IP address for the "network" test?

Actually isn't that the same? I am using localhost in any case.

> It would be interesting to compare a network-only performance test
> (like iPerf) for loopback and for going through the NIC.

iperf (one thread, 64K I/O size, 30 secs):
NIC: 445 MB/s
Loopback: 735 MB/s

In retrospect, for disk I/O:
/local: 39 MB/s
/nfs (loopback): 29 MB/s (25.5% drop)
/nfs (from a real server): 27 MB/s (30.5% drop, only point is that
this is a different disk on a different system and it doesn't
make
much sense to compare this to /local).

Thanks,

- KK



2008-06-30 15:25:41

by Chuck Lever

[permalink] [raw]
Subject: Re: NFS performance degradation of local loopback FS.

On Mon, Jun 30, 2008 at 5:57 AM, Krishna Kumar2 <[email protected]> wrote:
> [email protected] wrote on 06/27/2008 07:36:44 PM:
>
>> > But loopback is better than actual network traffic.
>>
>> What precisely do you mean by that?
>
> Sorry I was not clear. I meant that the loopback will be better than
> actual traffic between different server/client.
>
>> You are testing with the client and server on the same machine. Is
>> the loopback mount over the lo interface, but you mount the machine's
>> actual IP address for the "network" test?
>
> Actually isn't that the same? I am using localhost in any case.

As I understand it, "lo" is effectively a virtualized network device
with point-to-point routing. Looping back through a real NIC can, in
many cases, go all the way down to the network hardware and back, and
is likely subject to routing decisions in your system's network layer.
So I would expect them to be different in most cases.

>> It would be interesting to compare a network-only performance test
>> (like iPerf) for loopback and for going through the NIC.
>
> iperf (one thread, 64K I/O size, 30 secs):
> NIC: 445 MB/s
> Loopback: 735 MB/s

--
Chuck Lever