Direct NFS anyone?

#1

I was looking around because of some performance problems with Direct NFS, but it seems nobody
has problems with Direct NFS.

Or is nobody using it?

Thanks,

Gerard

#2

Gerard H. Pille wrote,on my timestamp of 20/12/2010 8:14 PM:

Quote:

I was looking around because of some performance problems with Direct NFS, but
it seems nobody has problems with Direct NFS.

In what OS platform?

Quote:

Or is nobody using it?

Count 1 here.

#3

Noons wrote:

Quote:

Gerard H. Pille wrote,on my timestamp of 20/12/2010 8:14 PM:
I was looking around because of some performance problems with Direct NFS, but
it seems nobody has problems with Direct NFS.

In what OS platform?

Or is nobody using it?

Count 1 here.

It feels great not to be alone. We can form a duo.

We're testing with Oracles Linux on VMWare, database files hosted by NetApp.

Too many variables to locate the bottleneck. When I ran my test yesterday evening, having all
systems to myself, performance was 4 times better than during working hours.

#4

On 20.12.2010 13:57, Gerard H. Pille wrote:

Quote:

Noons wrote:
Gerard H. Pille wrote,on my timestamp of 20/12/2010 8:14 PM:
I was looking around because of some performance problems with Direct
NFS, but
it seems nobody has problems with Direct NFS.

In what OS platform?

Or is nobody using it?

Count 1 here.

It feels great not to be alone. We can form a duo.

We're testing with Oracles Linux on VMWare, database files hosted by
NetApp.

Too many variables to locate the bottleneck. When I ran my test
yesterday evening, having all systems to myself, performance was 4 times
better than during working hours.

netapp has docs on using oracle on their storage. use your support
account to find them.

anyway, I'd expect many people have complaints about nfs (direct or not)
performance with DBMS.

regards

#5

On Tue, 21 Dec 2010 09:15:12 +0100, Hank M. Higher wrote:

Quote:

anyway, I'd expect many people have complaints about nfs (direct or not)
performance with DBMS.

Why?

--
http://mgogala.freehostia.com

#6

Gerard H. Pille wrote,on my timestamp of 20/12/2010 11:57 PM:

Quote:

Or is nobody using it?

Count 1 here.

It feels great not to be alone. We can form a duo.

We're testing with Oracles Linux on VMWare, database files hosted by NetApp.

Too many variables to locate the bottleneck. When I ran my test yesterday
evening, having all systems to myself, performance was 4 times better than
during working hours.

Errr.... I meant: count 1 not using it.
We might have to, if the new hardware ends up being Netapp. But I'm sticking to
iscsi and/or fc for the time being, until someone proves to me d-nfs can do
everything it promises at all times. We all know lab conditions are not real
life ones...

#7

On Tue, 21 Dec 2010 23:05:11 +1100, Noons wrote:

Quote:

Errr.... I meant: count 1 not using it. We might have to, if the new
hardware ends up being Netapp. But I'm sticking to iscsi and/or fc for
the time being, until someone proves to me d-nfs can do everything it
promises at all times. We all know lab conditions are not real life
ones...

I would be really interested to hear somebody's experiences with NFS,
direct or otherwise. I used it once, a long time ago, only for a simple
DW design.

--
http://mgogala.byethost5.com

#8

On Dec 21, 10:56*am, Mladen Gogala <n... (AT) email (DOT) here.invalid> wrote:

Quote:

On Tue, 21 Dec 2010 23:05:11 +1100, Noons wrote:
Errr.... I meant: count 1 not using it. We might have to, if the new
hardware ends up being Netapp. *But I'm sticking to iscsi and/or fc for
the time being, until someone proves to me d-nfs can do everything it
promises at all times. *We all know lab conditions are not real life
ones...

I would be really interested to hear somebody's experiences with NFS,
direct or otherwise. I used it once, a long time ago, only for a simple
DW design.

--http://mgogala.byethost5.com

It can't be all bad (famous last words), as Oracle has internal
support for it in 11gR2.

#9

On Dec 21, 1:05*pm, Noons <wizofo... (AT) yahoo (DOT) com.au> wrote:

Quote:

Gerard H. Pille wrote,on my timestamp of 20/12/2010 11:57 PM:

Or is nobody using it?

Count 1 here.

It feels great not to be alone. We can form a duo.

We're testing with Oracles Linux on VMWare, database files hosted by NetApp.

Too many variables to locate the bottleneck. When I ran my test yesterday
evening, having all systems to myself, performance was 4 times better than
during working hours.

Errr.... I meant: count 1 not using it.
We might have to, if the new hardware ends up being Netapp. *But I'm sticking to
iscsi and/or fc for the time being, until someone proves to me d-nfs can do
everything it promises at all times. *We all know lab conditions are not real
life ones...

<;-)>
You must be one of those "Manly Men" Kevin Closson wrote about 3,5
years ago:
http://kevinclosson.wordpress.com/20...-nfs-is-weird/
http://kevinclosson.wordpress.com/20...1g-direct-nfs/

http://www.nd.edu/~sheridan/Neander%.../Image11..html
<d&r>

--
Jeroen

#10

Noons wrote:

Quote:

Gerard H. Pille wrote,on my timestamp of 20/12/2010 11:57 PM:

Or is nobody using it?

Count 1 here.

It feels great not to be alone. We can form a duo.

We're testing with Oracles Linux on VMWare, database files hosted by NetApp.

Too many variables to locate the bottleneck. When I ran my test yesterday
evening, having all systems to myself, performance was 4 times better than
during working hours.

Errr.... I meant: count 1 not using it.
We might have to, if the new hardware ends up being Netapp. But I'm sticking to iscsi and/or fc
for the time being, until someone proves to me d-nfs can do everything it promises at all times.
We all know lab conditions are not real life ones...

Oh great! One down. And as the other reactions show, no one is using it.