deependhulla / proxmox-pve-storage-netapp

proxmox netapp c-mode plugin

proxmox-pve-storage-netapp

proxmox netapp c-mode plugin

description

This module do the integration of netapp san array into proxmox. Template,clones,snapshots, snapshots rollback are done on the netapp array

installation

download libpve-storage-perl_XXX_all.deb in releases, and install it on your proxmox host with

dpkg -i libpve-storage-perl_XXX_all.deb

Netapp requirement

You need to use Cluster Mode You need to have flexvol license, and snaprestore license

proxmox storage config

/etc/pve/storage.cfg

netapp: mynetappstorage
        path /mnt/pve/mymountpoint
        server x.x.x.x (ip of your controller nfs)
        options rw,noatime,nodiratime,noacl,vers=4,rsize=65536,wsize=65536,hard,proto=tcp,timeo=600
        vserver myvservername
        aggregate myaggregatename
        adminserver x.x.x.x (cluster vip admin ip)
        login youradminlogin
        password yourpassword
        content images
        maxfiles 1

you need to create the mountpoint directory on proxmox host

mkdir /mnt/pve/mymountpoint

NETAPP CLUSTER MODE, PHYSICAL VIEW

                (Cluster admin VIP)
           ----------------------------
  |------ |INTERCLUSTER NETWORK (10GB)|------
  |        ----------------------------     |
  |              |           |              |
  |              |           |              |
[NODE1]       [NODE2]     [NODE3]         [NODE4]
  |              |           |              |
----------   ----------   -----------    -----------
AGGREGATE1    AGGREGATE3   AGGREGATE4    AGGREGATE5
[vm1][vm2]    [vm3][vm4]   [vm5][vm6]    [vm7][vm8]
----------   ----------   -----------    ----------
  |
----------
AGGREGATE2
[vm9][vm10]
----------

An aggregate is a raid-dp of x disk.
Each aggregate have differents volumes, 1 by vm.

NETAPP CLUSTER MODE, LOGICAL VIEW
---------------------------------
A virtual server is defined on top of the physical cluster

-----------------------------------
|           VSERVER               |
|        (NFS ips, 1 by node)     |
| [vm1][vm2][vm3][vm4][vm5]...    |
-----------------------------------

Each node is a nfs metadatas server.
You only need to mount nfs on 1 metadataserver to access the whole vserver

NFS VIEW
---------
We only mount the root

mymetadataserver:/ /mnt/pve/netapp-vservername

/
/images/vm1/vm-1-disk-1.raw
/images/vm1/vm-1-disk-2.raw
/images/vm2/
/images/vm3/
/images/vm4/
/images/vm5/
/images/vm6/
/images/vm7/
/images/vm8/
/images/vm9/
/images/vm10/

Eech vm have a volume in the storage.(volume name can't be only numeric, so I prefix them with "vm".vmid.
Each node is limited to 500volumes, that why I don't have 1 volume by disk.
We can move volume in realtime in the cluster, from 1 aggreate to another aggregate. (in netapp interface for now)