Support Policies for RHEL High Availability Clusters - LVM in a Cluster

Updated -

Contents

Overview

Applicable Environments

  • Red Hat Enterprise Linux (RHEL) with the High Availability Add-On

Useful References and Guides

Introduction

This guide offers Red Hat's policies around the usage of LVM within RHEL High Availability clusters. Users of RHEL High Availability clusters should adhere to these policies in order to be eligible for support from Red Hat with the appropriate product support subscriptions.

Policies

Cluster management of shared LVM: Clusters that share block storage devices and manage them with LVM must use one of the following methods for managing shared access to these devices in order to receive support from Red Hat:

  • HA-LVM: Shared LVM volumes are activated by the cluster resource manager using an LVM or LVM-activate resource (with pacemaker) or lvm resource (with rgmanager). When using LVM resource and not managed by clvmd or controld resource, then /etc/lvm.conf attribute should be set tolocking_type=1.
  • lvmlockd in RHEL 8 Resilient Storage clusters with /etc/lvm.conf attribute set to locking_type=1.
  • clvmd in RHEL 6 or 7 Resilient Storage clusters with /etc/lvm.conf attribute set to locking_type=3.

Red Hat may be unable to provide support with cluster environments sharing LVM devices that do not use one of these facilities.

LVM RAID: Red Hat's support of shared LVM RAID volumes - volumes using any "raid" segment type for mirroring - in a High Availability cluster is subject to the following conditions:

  • RHEL 8 with the LVM-activate resource type: Supported by Red Hat
  • RHEL 7 with the LVM or LVM-activate resource type: Supported by Red Hat with resource-agents-3.9.5-54.el7 or later, only when using tagging with a volume_list.
  • RHEL 6 with pacemaker and the LVM resource type: Supported by Red Hat with resource-agents-3.9.5-34.el6 or later, only when using tagging with a volume_list.
  • RHEL 6 with rgmanager and the lvm resource type: Supported by Red Hat with resource-agents-3.9.2-21.el6 or later, only when using tagging with a volume_list.
  • Any RHEL release while using lvmlockd or clvmd from Resilient Storage: Red Hat does not support usage of LVM RAID volumes with these clustered-LVM managers

LVM mirroring (non-RAID): Red Hat supports usage of shared LVM mirorred volumes - those of the segment type mirror - within RHEL High Availability clusters.

Use of cache logical volumes in a cluster: Red Hat supports cache volumes in RHEL 8, in RHEL 7.1 or later and RHEL 6.7 or later.

Cache volumes can only be deployed in non-clustered volume groups - that is, VGs that are managed via an HA-LVM system through an LVM or LVM-activate resource (pacemaker) or lvm resource (rgmanager). Cache volumes are not supported with lvmlockd or clvmd from the the Resilient Storage Add-On.

Cache volumes shared by a cluster must use writethrough cache mode by way of allocation { cache_mode = "writethrough" } in /etc/lvm/lvm.conf.

lvm2-lvmetad: With releases of lvm2 that provide support for lvm2-lvmetad, clusters sharing access to LVM volumes must have lvm2-lvmetad disabled in the configuration and as a service to prevent problems resulting from inconsistent metadata caching throughout the cluster.

  • RHEL 8, 7, and 6: lvm2-lvmetad can be disabled in /etc/lvm/lvm.conf in the global section with use_lvmetad=0.

  • RHEL 8 and 7: lvm2-lvmetad can be disabled and stopped at a service level using systemctl:

    # systemctl disable lvm2-lvmetad
# systemctl stop lvm2-lvmetad

  • RHEL 6: lvm2-lvmetad can be disabled at stopped at a service level using chkconfig and service:

    # chkconfig lvm2-lvmetad off
# service lvm2-lvmetad stop

RHEL 6 ocf:heartbeat:LVM resources with exclusive=true and clvmd: The ocf:heartbeat:LVM resource type in RHEL 6 pacemaker clusters must only be used with non-clustered volume groups if the resource must be exclusively-activated (using attribute exclusive=true). Red Hat does not support the usage of LVM resources with exclusive=true managing a clustered volume group in RHEL 6. Exclusive volume groups must use the tagging method of HA-LVM management, which is achieved by ensuring the volume group does not have the clustered attribute enabled on it.

Related Solutions