VMWare High Availability

[EnChong]

Hi,

I am new to VMWare and hope someone can advise me regarding the VMWare configuration.

Please refer to the attached word document named Inquiry_VMware.docx

Page 1, I have 2 hosts and 3 virtual machines. Host 1 has no virtual machine. Host 2 has 3 virtual machines.

Page 2, It show that vCenter HA is not configured, does it mean if any of my virtual machine failed, it will not failover from the master host to the slave host ?

Page 3, It show that no of vSphere HA heartbeat datastores for host is 1 less than required 2, how to rectify this and does it effect anything ? So far my virtual machines are running fine.

Page 4, May i know the failure of "Guest not heartbeating", does it mean vCenter unable to detect the host ? Does warning in page 3 will affect this ?

Page 5, Currently my application and database are located in virtual machine named "EGSC-MFS". There is one time that the VM swap to the slave host due to master host breakdown, there is some downtime for my application before the slave host started up the VM and took over the master role. I would like to use Fault Tolerance in order to have zero downtime, may i know how should i configure this Fault Tolerance properly ?

Thank you so much for your time.

Regards,

En Chong.

[GayathriS]

Hi

Find answers to your questions:

Page 2, It show that vCenter HA is not configured, does it mean if any of my virtual machine failed, it will not fail over from the master host to the slave host ?

A: Yes you are right, if HA is not configured, in case of host failure VM will not be failed over to another host.

There are requirements to configure HA check doc:

HA cluster configuration: Requirements and steps

------------------------------------

Page 3, It show that no of vSphere HA heartbeat datastores for host is 1 less than required 2, how to rectify this and does it effect anything ? So far my virtual machines are running fine.

A: This explain you about what this issue is about :

The number of vSphere HA heartbeat datastores for this host is 1

---------------------------------------

Page 4, May i know the failure of "Guest not heart beating", does it mean vCenter unable to detect the host ? Does warning in page 3 will affect this ?

A: This condition can occur because the VMware Tools heartbeat or the application heartbeat status turned red of guest machines.

--------------------------------------

Page 5, Currently my application and database are located in virtual machine named "EGSC-MFS". There is one time that the VM swap to the slave host due to master host breakdown, there is some downtime for my application before the slave host started up the VM and took over the master role. I would like to use Fault Tolerance in order to have zero downtime, may i know how should i configure this Fault Tolerance properly ?

A:  Below blog helps you in configuring FT :

How to Set Up and Use vSphere Fault Tolerance

I dont see document attached.

Please mark this answer as correct and helpful if this helps you.

regards

Gayathri

[adgate]

I found these guides useful for the case:

1. Configure Heartbeat Datastores

2. add a secondary heartbeat datastore. VMware Knowledge Base

What is the storage used for HA cluster?

[Duncan Epping]

Page 1, I have 2 hosts and 3 virtual machines. Host 1 has no virtual machine. Host 2 has 3 virtual machines.
Page 2, It show that vCenter HA is not configured, does it mean if any of my virtual machine failed, it will not failover from the master host to the slave host ?
Page 3, It show that no of vSphere HA heartbeat datastores for host is 1 less than required 2, how to rectify this and does it effect anything ? So far my virtual machines are running fine.
Page 4, May i know the failure of "Guest not heartbeating", does it mean vCenter unable to detect the host ? Does warning in page 3 will affect this ?
Page 5, Currently my application and database are located in virtual machine named "EGSC-MFS". There is one time that the VM swap to the slave host due to master host breakdown, there is some downtime for my application before the slave host started up the VM and took over the master role. I would like to use Fault Tolerance in order to have zero downtime, may i know how should i configure this Fault Tolerance properly ?

You are confusing two things: vCenter HA and vSphere HA. vCenter HA is an availability solution for vCenter Server. vSphere HA is an availability solution for your VMs.

Page 2: Correct, but your VMs are protected

Page 3: Heartbeat datastores are used to determine what has happened when your network fails. I suspect you have a single datastore as this is a small environment, you can safely ignore this message

Page 4: It could be that you don't have VMware tools installed in the VMs that are running on the host. This does not impact vSphere HA itself as you do not have "VM monitoring enabled", you can ignore it

Page 5: As mentioned above, FT configuration can be easily done by following the VMware documentation.

I would recommend the following:

Bookmarks: VMware Docs Home

It has ALL the basic documents one could need. If you like to have a deep knowledge of HA I would recommend to download the ebook I wrote, it is free: https://pages.rubrik.com/clustering-deep-dive-ebook.html

[NathanosBlightc]

First of all, you should know there are two different concepts:

1. vSphere High Availability (vSphere HA) is a cluster-level feature that provides availability for the VMs in the virtual infrastructure. in the case of a host failure, HA will migrate its VMs to another healthy hosts belongs to that cluster.
2. vCenter HA is a new feature that is introduced in the release of vSphere 6.5 that provides availability for only the vCenter Server (that is enabled just for VCSA). This is a great feature, but because of a specific clustering structure of VCHA, it requires at least 3 hosts to deploy Active, Passive and Witness nodes.

For the 2nd Q: vSphere HA needs to detect host failures intelligently, So it must decide based on many investigation items include Heartbeats that are generated by all hosts belong to that cluster. It chooses at least two shared datastore for the heartbeating, then the cluster can device what is the reason for the host failure... it is just a simple network segmentation or there is a really host failure. If the vCenter doesn't have any heartbeat from that host, So consider it as a failed host, then HA will migrate VMs of that host to another host in the cluster.

For the 3rd Q: For using FT it's better to dedicate separate physical network uplinks to keep a good and reliable synchronization (because of high network transmission). FT has some limitations in each version of vSphere, So be careful if you want to use it. Please read this guide: https://docs.vmware.com/en/VMware-vSphere/6.7/vsphere-esxi-vcenter-server-67-availability-guide.pdf