NFS Support

HPE Ezmeral Runtime Enterprise supports the NFSv3 service in MFS pods.

NOTE NFSv4 is not currently supported.

Enabling or Disabling NFSv3

To enable NFSv3, add the nfs: true entry to the dataplatform YAML file. The full.yaml and simple.yaml example files include this entry by default. 

gateways:
  nfs: true
  mast: true
  objectstore:
    image: objectstore-2.0.0:202101050329C
    zones:
      - name: zone1
        count: 1
        sshport: 5010
        size: 5Gi
        fspath: ""
        hostports:
           - hostport: 31900
        requestcpu: "1000m"
        limitcpu: "1000m"
        requestmemory: 2Gi
        limitmemory: 2Gi
        requestdisk: 20Gi
        limitdisk: 30Gi
        loglevel: INFO

NFSv3 support is disabled if either:

  • You specify nfs: false.
  • The nfs: < true | false > entry is not present.

Considerations

When working with NFSv3 in HPE Ezmeral Runtime Enterprise:

  • The default share is /mapr. To change the default share, see Customizing NFS, below.
  • If the MFS pod hosting the NFSv3 service stops and then restarts on a different node, then the NFSv3 mounts will stop working.
  • The default NFS server port is 2049. Do not change the default port.
  • You must mount the NFS drive manually; auto-mount is not supported.

Known Issues

The following issues are known to exist with maprcli commands for NFS:

  • maprcli setloglevel nfs -loglevel or maprcli trace info or tracelevel: DEBUG - Command does not work or returns an exception. The default trace levels are:
    • NFSD : INFO
    • NFSDProfile : INFO
    • NFSExport : ERROR
    • NFSHandle : ERROR
  • maprcli nfsmgmt - No response from the NFS server.
  • maprcli node services -nfs start|stop|restart|enable|disable - Command has no effect.
  • maprcli alarm list - An alarm is generated that indicates the wrong NFS version.

Customizing NFS

You can customize NFS behavior by modifying the exports and nfsserver.conf sections of the config map (template-mfs-cm.yaml), which is located at bootstrap/customize/templates-dataplatform/template-mfs-cm.yaml. Any configuration changes must be made before bootstrapping. For example, to change the share name, you must change the /mapr (rw): entry in the exports: section of the config map:

exports: |
    # Sample Exports file

    # for /mapr exports
    # <Path> <exports_control>

    #access_control -> order is specific to default
    # list the hosts before specifying a default for all
    #  a.b.c.d,1.2.3.4(ro) d.e.f.g(ro) (rw)
    #  enforces ro for a.b.c.d & 1.2.3.4 and everybody else is rw

    # special path to export clusters in mapr-clusters.conf. To disable exporting,
    # comment it out. to restrict access use the exports_control
    #
    /mapr (rw)

    #to export only certain clusters, comment out the /mapr & uncomment.
    #/mapr/clustername (rw)
    .
    .
    .
nfsserver.conf: |
    # Configuration for nfsserver

    #
    # The system defaults are in the comments
    #

    # Default compression is true
    #Compression = true

    # chunksize is 64M
    #ChunkSize = 67108864

    # Number of threads for compression/decompression: default=2
    #CompThreads = 2

    #Mount point for the ramfs file for mmap
    #RamfsMntDir = /ramfs/mapr
    .
    .
    .

Please refer to Managing the HPE Ezmeral Data Fabric NFS Service for information about managing the NFS service (link opens an external website in a new browser tab/window).

Configuring Client Access

To configure clients to connect to NFS on the Kubernetes cluster network, locate the node on which the MFS pod is running, such as the mfs-group1-0 pod shown below. The network location where NFS is running is the NODE value for the MFS pod:



In this example, the command to mount your share on the data-fabric cluster in Kubernetes is:

user-vbox2:~$ m2-sm2028-15-n4.mip.storage.hpecorp.net:/mapr