VMware Cloud Community
cyriltam
Contributor
Contributor
Jump to solution

HA Problem

After i setup the HA, it can restart it after the host is failure, but i want to the VM is migrate to the existing Host. the VM no neeed to reboot.

Many THx

Reply
0 Kudos
1 Solution

Accepted Solutions
Yattong
Expert
Expert
Jump to solution

cyriltam,

There is nothing from vmware that will move vm's on host failure without restart.

DRS will move vm's to different hosts when there is too much contention / load on any particular esx host.

HA will move and restart vm's if the esx host fails.

If you are looking for high availability, and zero downtime then you will need to look at MS clustering or Veritas clustering etc...

If you found this or any other answer useful please consider the use of the Helpful or correct buttons to award points ~y

View solution in original post

Reply
0 Kudos
7 Replies
spex
Expert
Expert
Jump to solution

HA always means a reboot for your VM the HA tries to recover from a host error. Migration is only possible if VM is up. If this is the case there is no need for HA.

Regards Spex

Reply
0 Kudos
djciaro
Expert
Expert
Jump to solution

Hi, if i understand your question correctly you would like VMs to be moved to another host without rebooting!

VMware HA is a high availability product that allows Virtual Machines to be started on another ESX server in the case of a hardware failure or host crash. HA is not able to migrate VMs to other hosts it is simply a technology that allows your to bring VMs back online after an outage.

You need VMware DRS and VMware VMotion in order to be able to migrate VMs without shutting them down.

This is a different type of technology that allows us to dynamically balance resource usage across ESX servers, it can help us prevent server instability and reduce the risk of overloading servers but is not a disaster recovery solution, HA is needed in other to bring VMs back online after a physical crash, the use of VMware DRS and HA clusters can provide us with nearly 0 downtime.

Please read these guides for more information: and

If you found this information useful, please consider awarding points for "Correct" or "Helpful". Thanks!
Yattong
Expert
Expert
Jump to solution

If you have a vmotion license and you have vmotion enabled. Then you can put the server in Maintenance Mode.

This will vmotion all the vm's off the host you want to reboot.

If you found this or any other answer useful please consider the use of the Helpful or correct buttons to award points ~y
cyriltam
Contributor
Contributor
Jump to solution

THx Man, Yap I want to when the host failure the VM will migrate to another host automatically without restart. so that i should config in the DR not in HA.

if so, any step i need to take care about it

Reply
0 Kudos
djciaro
Expert
Expert
Jump to solution

The best advice I can offer is to read the resource management guide:

Really the setup of your DRS cluster depends on the number of ESX servers available and the level that you want to set for the migration thresholds (Conservative to aggressive).

don't forget to award rewards points for answers and helpful answers

If you found this information useful, please consider awarding points for "Correct" or "Helpful". Thanks!
Reply
0 Kudos
Yattong
Expert
Expert
Jump to solution

cyriltam,

There is nothing from vmware that will move vm's on host failure without restart.

DRS will move vm's to different hosts when there is too much contention / load on any particular esx host.

HA will move and restart vm's if the esx host fails.

If you are looking for high availability, and zero downtime then you will need to look at MS clustering or Veritas clustering etc...

If you found this or any other answer useful please consider the use of the Helpful or correct buttons to award points ~y
Reply
0 Kudos
depping
Leadership
Leadership
Jump to solution

But MS Clustering isn't supported on 3.5! And VMware will be releasing Continous Availability later this year, which will give you a cluster alike HA feature....

But Indeed DRS doesn't work like that unfortunately.

Duncan

My virtualisation blog:

Reply
0 Kudos