#NFS Config checklist?

1 messages · Page 1 of 1 (latest)

stoic harness
#

Hi all,

In the next month or two I will be configuring my first NetApp boxes from scratch (AFF-A150) to serve NFS for vCenter datastores. I have undergone a good amount of self learning including Neil Anderson’s ONTAP 9 course as well as NetApp’s Cluster Administration course, and I have been managing our NetApps that my predecessor configured and installed for a few months.

I feel like I know what I need to do in regards to what actually needs to be configured/created on the NetApp, however what I’m struggling with a little bit is a solid “order of operations” so to speak. I know that certain things require other things to be configured first in order to be configured (my data SVM created before my data LIFs, my aggregates before my volumes etc.) however I still feel like there are many ways to skin the cat regarding what order you configure things in.

I was wondering if any of you experienced folks had insight or advice on what has worked for you in the past, or things that are easily forgotten to make sure to add to the list. Thanks in advance!

noble kraken
thin stratus
#

I have a setup doc for VMware. If you want, I can share

old gust
#

I encourage you to thoroughly read through this as it has a ton of great information: https://docs.netapp.com/us-en/ontap-apps-dbs/vmware/vmware-vsphere-overview.html

Many awesome tips in there. Especially things like setting the protocol in the export-policy to 'nfs' for VAAI integration (hardware acceleration) even when using NFSv3 (The VAAI copy offload feature requires the NFSv4 protocol to function, even if the data protocol is NFSv3). This has caught many people out. Also enabling the "-vstorage" option in your NFS Server.

Assuming the cluster is setup and aggregates and networking configured (ports, broadcast domains, interface groups, etc), then my high-level order is usually (for a NAS SVM to serve NFS to VMware):

  1. Create a new SVM
  2. Create a SVM Root LS Mirror Volume and create and initialise a LS SnapMirror (with a 15 minute schedule)
  3. Create a NFS LIF on each node that will host datastore FlexVols
  4. Create a read-only rule in the Default Export Policy for the SVM for the NFS VMkernel IPs of all ESXi hosts, allowing 'nfs' and superuser 'sys'
  5. Create a new export-policy that will be applied to datastore FlexVols (this must have the same rule as the Default policy, but allows read/write)
  6. Create a NFS Server and enable v3.0 and vstorage options
  7. Create the required Datastore FlexVols with the appropriate new export-policy, mount path, snapshot policy, and percent-snapshot-space (snapshot reserve)

You then should be ok to start mounting the datastores. Remember to mount the datastore via the LIF IP Address on the same node where each FlexVol resides. If you the FlexVol is on Node2, mount that datastores via the NFS LIF on Node2.

That's usually my high-level workflow.

VMware vSphere with ONTAP

ONTAP has been a leading storage solution for VMware vSphere environments for almost two decades and continues to add innovative capabilities to simplify man...

stoic harness
#

Thank you very much for your responses! I will definitely look through these docs.