26 Posts

July 10th, 2014 05:00

They connect via SmartConnect.

Each node has a network address.

The smartconnect policy setting is Round Robin

It seems set up correctly but obviously not!

2 Intern

 • 

300 Posts

July 10th, 2014 05:00

Hi Vince,

no that's not normal.

You should check:

- how do the clients connect: if they connect via IP, they only connect to the node who serves this IP.

- Subnet Setting: do you have network interfaces of every node in your IP Adress pool?

- SmartConnect Policy settings

26 Posts

July 10th, 2014 06:00

Yes if I ping the ZoneName I get a different IP each time.

I can browse through to each ip address and I can see all the fileshares from each one.

9 Legend

 • 

20.4K Posts

July 10th, 2014 06:00

use nslookup "smartzonename" multiple times.

2 Intern

 • 

300 Posts

July 10th, 2014 06:00

if you ping the SmartConnect ZoneName from the client 5 Times do you get 5 different IPs back?

can you connect to every ip from the client?

how do the "pool members" of your subnet look like?

26 Posts

July 10th, 2014 07:00

I get different IP’s each time

9 Legend

 • 

20.4K Posts

July 10th, 2014 08:00

are these CIFS or NFS connections ?

26 Posts

July 10th, 2014 08:00

CIFS

9 Legend

 • 

20.4K Posts

July 10th, 2014 11:00

are you licensed for SmartConnect Advances, if yes could you change policy to Connection Count instead of Round Robin ?

26 Posts

July 11th, 2014 01:00

It won’t affect current connections will it?

9 Legend

 • 

20.4K Posts

July 11th, 2014 04:00

it will not impact current connections

26 Posts

July 11th, 2014 05:00

I’ve changed to Connection Count, but still on the connections are going through node 1

4 Operator

 • 

1.2K Posts

July 11th, 2014 06:00

As is it always node 1... are you mounting from the SmartConnect zone name or from the

SmartConnect service address (or name) ? The latter always operates on the

node with the lowest number, which apparently happens to be---node 1.

If you do use the SmartConnect zone name, then the first IP returned might

had been massively cached, breaking all rules (TTL time to live = 0 secs!).

Check your DNS server for cache content and cache settings;

Windows servers cache for at least one second, but should not be

allowed for longer caching.

Addition caching might occur on the clients; don't rely on ping tests

as different caching might be in effect for mounts. (Google for

SMB DNS caching and your particular client OS, and your DNS server.)

26 Posts

July 11th, 2014 07:00

They currently connect through the SmartConnect service name

9 Legend

 • 

20.4K Posts

July 11th, 2014 07:00

VinceNoir wrote:

They currently connect through the SmartConnect service name

that's your problem, should be connecting to SmartConnect zone name, good catch Peter.

No Events found!

Top