Support Policies for RHEL Resilient Storage - Exporting gfs2 contents via other protocols

Updated -

Contents

Overview

Applicable Environments

  • Red Hat Enterprise Linux (RHEL) with the Resilient Storage Add-On

Useful References and Guides

Introduction

This guide describes Red Hat's policies around the exporting of gfs2 filesystems using other protocols. Users of gfs2 should adhere to these policies in order to be eligible for support from Red Hat with the appropriate product support subscriptions.

Policies

Supported gfs2 export protocols: Red Hat provides support for gfs2 file systems that have their contents shared or exported only with the following protocols:

  • NFS
  • CIFS
  • CTDB (clustered CIFS export)

Red Hat's support for gfs2 exported by these protocols is subject to the conditions laid out in other policies in this document.

Conditions for exporting gfs2 via NFS or single-host CIFS: When exporting gfs2 by way of NFS or single-host CIFS (not CTDB), the following conditions must be met in order to receive support from Red Hat for that gfs2 filesystem:

  • The filesystem can only be exported by a single protocol. No exporting of the same gfs2 filesystem via both NFS and CIFS, for example.
  • Only a single node of the cluster may mount a gfs2 filesystem that is being exported.
  • Any gfs2 filesystem that is to be exported must be mounted with the localflocks mount option.
    • NOTE: Data is at risk of corruption if mounted on multiple nodes while this mandatory-option is in use.
  • Any gfs2 directory that is exported cannot be accessed by any local process on the cluster node that mounts the filesystem. All access to the filesystem must come through the export protocol.

Conditions for exporting gfs2 via CTDB:

  • Note: As of RHEL 8.4, support for certain types of CTDB deployments in a RHEL Resilient Storage cluster is deprecated. See Support Policies for RHEL High Availability - ctdb General Policies for additional details.
  • The gfs2 filesystem should be mounted without the localflocks option specified if it is to be mounted by multiple cluster nodes.
  • Any gfs2 directory that is exported cannot be accessed by any local process on any cluster node. All access to the filesystem must come through the CTDB export.

Exporting gfs2 via other protocols or with other conditions: Red Hat does not provide support for gfs2 filesystems exported by other protocols besides those allowed by the policies in this guide, or with the allowed protocols with other conditions besides those listed here.

Lack of support by Red Hat for the file system does not automatically disqualify the export from being supported by Red Hat - the conditions of support for the export protocol may be defined by policies specific to that protocol. However even if Red Hat can provide support for the export protocol itself, any concerns or issues relating to the underlying file system behind that export would fall outside Red Hat's scope of coverage, if gfs2 is in use and the conditions laid out in this guide are not met.

No support for leases over CIFS exports backed by gfs2: gfs2 does not support the concept of leases, as provided by samba / CIFS and may be possible with other filesystems. Requests for a lease from a CIFS export of gfs2 contents will result in an error being returned to the issuer of the request.

Comments