NFS datastores with NetApp vfilers - anyone using this?

[mlubinski]

I would like to know if there is anyone here using this config together with VMware?

I wanted to ask about network setup to match best practicies.

If you found this or any other answer useful please consider the use of the Helpful or correct buttons to award points

[DSTAVERT]

Have a look at. http://www.vmware.com/resources/techresources/10096

From Netapp http://blogs.netapp.com/virtualstorageguy/2009/07/vsphere-on-netapp-best-practices-released.html

[mlubinski]

Hi,

Thanks for your answer, but I read these infos plenty of times. They do not contain information I need (about vfilers and their config). Here is description of my question:

My FAS has 10Gbit single VIFconfigured as

trunk. I have vfilers configured (each has different vlan assigned (like

nfs-vif100, nfs-vif200), and also IP assigned to vfiler). So let's say:

vfiler1:

vlan 100: 192.168.100.1 (interface assigned nfs-vif100)

/vol/vfiler1/vol1

/vol/vfiler1/vol2

/vol/vfiler1/vol3

/vol/vfiler1/vol4

vfiler2:

vlan 200: 192.168.200.1 (interface assigned nfs-vif200)

/vol/vfiler2/vol1

/vol/vfiler2/vol2

/vol/vfiler2/vol3

/vol/vfiler2/vol4

etc

Now:

I want to setup my esx hosts to follow up best practicies (with NFS).

Each ESX host connects to 2 1gbit switches, but now I have only one

vmkernel port confitgured, so it goes over one link only, and second one

is only "redundant". Best practicies from NetApp say that I should have

2 vmkernel ports configured on my ESX host and each vmkernel should be

in different IP subnet. Also Netapp should have 2 "target" IPs

configured. Then please tell me how should I proceed with that? Should I

just for each ESX vfiler perform following:

ifconfig nfs-vif100 alias

192.168.101.1 netmask 255.255.255.0

ifconfig nfs-vif200 alias

192.168.201.1 netmask 255.255.255.0

and then on each ESX host add second

vmkernel port in vlan 100/200 with IPs from range 192.168.101.x and

192.168.201.x and then remove all volumes, and remount them as

following:

NFS 192.168.100.1 -> /vol/vfiler1/vol1

NFS

192.168.101.1 -> /vol/vfiler1/vol2

NFS 192.168.100.1 ->

/vol/vfiler1/vol3

NFS 192.168.101.1 -> /vol/vfiler1/vol4

the same for

second vfiler?

this way I would use only one vlan number on ESX side (as

well as on NetApp) and use ip alias to configure second target IP from

different subnet. This is the way? Or should something be changed in

that? Is this the best solution in the meaning of performance to volumes

over NFS? If I misconfigure something by using IP alias, can it be

easily removed? or after mistake it stays as "unconfigured" and can't be

removed without a reboot? I can't afford any reboots, so I prefer not

to make any changes, than implement any that require reboot.

Thank you in

advance for your answer.

If you found this or any other answer useful please consider the use of the Helpful or correct buttons to award points

[markymark007]

Hi

Just wondering if you got this questions answered?

I am also looking to deploy many vfilers and deploy ESX datastores on them. Because the vfilers will be on different networks this brings up the question of best practices to use on the ESX vmkernel configuration. Eg. Do you add a vmkernel port for every vfiler? This would clearly mean that the ESX hosts would have lots of vmkernel ports on many networks.

Any info you have on this config would be very helpful

[mlubinski]

Naah, my questions were partially answered, but I think I will follow up these steps:

- on each vfiler will add 2-3 additional VLANs and assign 2-3 IPs from these vlans

- on ESX host (that will be using ONLY this vfiler) will assign 2-3 vmkernel ports with IPs from range specified above

- re-mount volumes using different IPs (from vfiler)

that should make this fine solution

If you found this or any other answer useful please consider the use of the Helpful or correct buttons to award points