Part 8 – Highly Available iSCSI Multi Pathing Storage for VME 9.1

On ESXi you enable the software iSCSI adapter and it generates the IQN. Here, cluster provisioning does that step iSCSI initiator name assignment in its iSCSI phase. Every host already has its initiator ready, which is why we could read it from the file before adding any target.

At the time of writing, VME 9.1 doesn’t show the host’s iSCSI initiator name (IQN) anywhere in the UI. Read it on each host instead, then add it to the SAN’s ACL so the host can log in

I’ve written a detailed blog in the past to cover Multipathing on ISCSI in VMware but at this moment, I’m just extending my Linux SAN to mount few datastores and check VME but it’s very easy to setup one from Windows.

We added a 200 GB disk from our block array

In the past chapter we had added these two virtual switches ISCSI A and ISCSI B and this offer adding multipathing to the storage creates a HA solution that will enable you to test most features on the VME.

ISCSI path A runs on VLAN 1617

ISCSI path B runs on VLAN 1618

The next step is to add a connection to the target SAN Array and this is also known as “target discovery” so click Add ISCSI

Add the IP Address of the SPA ISCSI Port .Thus, we will have two connections to our storage array

Add the IP Address of the SPB ISCSI Port as well. Thus, we will have two connections to our storage array

Hit Rescan and check the path status again and you will be able to identify the “Logged in” status of one of the luns.

Verification of multipathing

The GUI is bit unpolished here so we cant see the active paths as one would expect from vSphere but here is the view from Host CLI to show its working

multipath -ll
hvmcli storage multipath status
iscsiadm -m session
sudo hvmcli storage multipath validate

This shows we have 2 Active paths.

Creating Datastores

To create a datastore from the ISCSI lun, hover over to the storage section and select Add

Give the datastore a name, choose type as shared lun and pick the 200 GB lun we mapped and Click Save

Our shared datastore is ready

Status changes to provisioning

After few minutes its ready for consumption

Do a write check on the path to verify if volume is writable.

The filesystem is GFS2 (Global File System 2). It’s Red Hat’s clustered filesystem and it plays the role VMFS plays on ESX . That’s it and we successfully configured multipathing and high availability in our iSCSI Storage Network and created a datastore from it.

Creating Heartbeat Datastore

Without HA, failed host’s VMs won’t restart elsewhere.

Go to Cluster → Storage → Data Stores, click on shared-ds-01

Enable HeartBeat Target

Summary → Quorum panel should list shared-ds-01 as the heartbeat datastore.

On a host, we can see the heartbeat files being written every 20 seconds and HA Is now configured

Migrating the vme-mgmr VM to Shared Storage

The Morpheus manager is a VM and that was installed to the first node, so now as our shared storage is set should move the vme-mgmr to a managed state