Search <book_title>...

Veritas InfoScale™ 7.3.1 Virtualization Guide - Linux on ESXi

Last Published: 2017-11-06

Product(s): InfoScale & Storage Foundation (7.3.1)

Section I. Overview
1. Overview of Veritas InfoScale solutions in a VMware environment
Section II. Deploying Veritas InfoScale products in a VMware environment
1. Getting started
Section III. Use cases for Veritas InfoScale product components in a VMware environment
Section IV. Reference
1. Appendix A. Known issues and limitations
  1. Prevention of Storage vMotion
2. Appendix B. Where to find more information

Configuring non-SCSI3 fencing

VMDK files do not currently support SCSI-3 Persistent Reservation and therefore non-SCSI-3 PR fencing must be used. Coordination point (CP) servers provide the required level of server based fencing. At this point in the configuration process, the three CP servers that are to be used with this cluster should be available and the CP service should be up and running.

To configure non-SCSI-3 fencing

If you started at the beginning of the installer process and selected the enable fencing option, you are prompted to configure fencing.
If you chose not to enable fencing at that point, then the cluster configuration is finished. You should now run installsfcfsha61 -fencing to enable fencing in the cluster.
Regardless of how you navigated to the fencing configuration of the installer, select option 1 for Coordination Point client-based fencing.
When prompted if your storage environment supports SCSI-3 PR, select n , since VMDK files do not support SCSI-3 PR.
When prompted if you wnat to configure Non-SCSI-3 fencing, select y.
For production environments, three CP servers are recommended. Enter 3 when prompted for the number of coordination points.
Specify how many interfaces the CP servers will be listening on and the IP address of each interface. If a CP server is reachable via several networks, the best practice is to configure every interface. This allows the SFCFSHA nodes maximum communication flexibility, should a race condition occur.
Enter the host names and VIPs for the other CP servers and review the fencing configuration.
When prompted, select secure mode. All the trusted relationships between cluster nodes and CP servers are automatically set up.
Verify that the cluster information is correct. Each node is registered with each of the CP servers. Once this is done, the installer will restart VCS to apply the fencing configuration. At this point we don´t have any file system configured yet.
When prompted, it is a recommended best practice to configure the Coordination Point Agent on the client, so CP servers are proactively monitored from the cluster. This step completes the fencing configuration.

Once fencing configuration is complete, you can verify if it is correct.

To verify the fencing configuration

Query each of the CP servers to verify each node has been registered.

# CCPS_USERNAME=CPSADM@VCS_SERVICES
# CPS_DOMAINTYPE=vx
[root@cfs01 install]# cpsadm -s cps1v -a list_nodes
ClusterName UUID                                   Hostname(Node ID) Registered
=========== ====================================== ================  ==========
cfs0        {38910d38-1dd2-11b2-a898-f1c7b967fd89} cfs01(0)             1
cfs0        {38910d38-1dd2-11b2-a898-f1c7b967fd89} cfs02(1)             1
cfs0        {38910d38-1dd2-11b2-a898-f1c7b967fd89} cfs03(2)             1
cfs0        {38910d38-1dd2-11b2-a898-f1c7b967fd89} cfs04(3)             1
[root@cfs01 install]# cpsadm -s cps1v -a list_membership -c cfs0
List of registered nodes: 0 1 2 3

Run the same command against the each CP server.
Using the VCS Cluster Explorer screen, we can see that the vxfen service group has been created to monitor CP servers and that it is healthy.