Virtual machine auto migrate to host other in HA cluster.

[hongchuong]

Hi everyone !

I have 3 hosts ESXi in a HA cluster. I was enable vMotion for HA cluster this.

But, When I was migrate a virtual machine to a  new host by normal then a day it auto migrate to old host.

P/S: 3 hosts no shutdown, We running

So, Can you help me fix the case this.

Thanks .

[Anjani_Kumar]

I believe you are running out of resource in the host you are migrating. Also check below things.

1. DPM is not enabled.

2.check the DRS effect level.

3. Check if any affinity or anti affinity rules are there.

[SavkoorSuhas]

Hello,

For that cluster, please check if DRS is enabled.

Right click cluster and edit settings.

if DRS is enabled check the DRS level, is it set to manual or partially automated or fully automated along with the migration threshold.

If DRS is not required, you can uncheck it.

If DRS is required, change the migration threshold to lower value, or change the level to manual, where it will ask you if migration of VM is required to balance to resource load in cluster.

[hongchuong]

Thank Anjani_Kumar‌

I was enable Admission Control in vSphere HA.

I was config box "Percentage of cluster resources reserved as failover spare capacity " are CPU 50% and Memory 50%.

When  i view performance of 3 hosts this.

Host     CPU      Memory

1          1%             55%

2          7%             62%

3          1%             2%

But, when i was migrate a virtual machine from host 2 to host 3 by normal then a day virtual machine this be auto to host 2,

In host 3 haven't any virtual machine.

[hongchuong]

Hi SavkoorSuhas‌

I don't enable DRS because license VM of me not support

Thank you.

[Anjani_Kumar]

Well, in your scenario Where no DRS/DPM is there, so no automigration and intiial placement is possible.

After this only HA failure is the reason which can restart the vm on other host. (i am sure this is not the reason else you get the downtime on that vm.)

I will suggest you to the check the vm event/tasks which are intiated on it. This will make it to clear more in depth.

[hongchuong]

Hi Anjani_Kumar‌

Now, I was disable the option Admisson Control on vSphere HA  and migrate by normal a virtual machine to host other.

Tomorrow, I will check it has auto migrate to old host.

I think, i wasn't understand to config the value in Admisson Control.

Thank you

[hongchuong]

Thanks Anjani_KumarAnjani_KumarAnjani_Kumar

After I disable option "Admission Control " on vSphere HA of " setting HA cluster " and migrate normal the virtual machine to host other.

But, after 2 days. The virtual machine this auto migrate to old host.

Now, Have a host with RAM 98 %, host other with RAM 0% and haven't any Virtual Machines.

I don't understand, why is this.

Thanks .

[npadmani]

as Anjani suggested earlier in one of the post, check event/log entries, it will give you an idea what is happening.

your explanation tells, this is failover event. your host failed/restarted where some VMs were running, due to HA, VMs failed over to old host. and the host which got restarted is not having any VMs now. if you had fully auto. DRS enabled, it would have tried doing load balancing but since only HA is in use so one of the host is heavily loaded while the other one is not in use.

So next step would be to find out reason of this unexpected reboot of your host and try to fix it.

[hongchuong]

Hi npadmani‌

I don't enable option DRS because my license not support.

I was migrate virtual machine by normal to new host when it running.

But, after a day, it be auto migrate to old host , although ram of host this is 96%.

[npadmani]

I understand license limitation you have, I wasn't asking you to enable it.

I was just trying to explain, like what if you had DRS, that's all.

what is happening in your case is, as you said your have ESXi1 and ESXi2, and VM1 along with some more VMs running on ESXi1.

you are migrating this VM1 from ESXi1 to ESXi2.

HA is in place with these to ESXi  host.

let's say ESXi2 get's rebooted for some reason overnight. in that case, VM1 which was running on ESXi2 goes down, due to HA, it gets failed over to ESXi1 and that's how you are able to see that VM on ESXi1 again. See if this helps.

[hongchuong]

Dear npadmani‌

As you said, I was migrate VM1 from ESXi 1 to ESX2 by normal but after a day VM1 be auto migrate from ESX2 to ESX1 ( two hosts running ).

[Anjani_Kumar]

I suggested you to do verify the logs why its migrated and by whom. The logs are quite self elaborative itselfs. i must say have to look in to logs why its moved.

[hongchuong]

Dear Anjani_Kumar‌

So, Can you guide me export file log for case this ?

Because, I don't see any alarm on ESXi and HA cluster.

Thanks .

[Anjani_Kumar]

You can check your vmware*.log files on the datastore for the specific VM for a quick review of the time, executor, and host migration. etc.

[hongchuong]

Thank Anjani_Kumar‌

after check the log file host ESXi

I think it auto reboot. Can you view log it.

[Anjani_Kumar · Accepted solution]

Here is the whole story. Your Host is rebooting due to some reasons and thats why its failover the vms to the others host available.

AS all said by logs here.

HA is sending those vms to other host via HA failover. hope you got your doubt clear here.

2015-08-04T10:00:01Z heartbeat: up 2d16h20m56s, 8 VMs; [[359230 vmx 12576068kB]  [359735 vmx 16319076kB] [358441 vmx 18670668kB]] [[35901 sfcb-vmware_raw 6%max]                                                                                         [651271 sfcb-vmware_bas 16%max] [35378 sfcb-pycim 19%max]]

2015-08-04T11:00:01Z heartbeat: up 2d17h20m56s, 8 VMs; [[359230 vmx 12576048kB]                                                                                         [359735 vmx 16319012kB] [358441 vmx 18779196kB]] [[35901 sfcb-vmware_raw 6%max]                                                                                         [35378 sfcb-pycim 19%max] [661513 sfcb-vmware_bas 19%max]]

2015-08-04T11:42:55Z bootstop: Host has booted

2015-08-04T11:42:55Z bootstop: partition core dump found

2015-08-04T12:00:01Z heartbeat: up 0d0h18m22s, 0 VMs; [[37143 sfcb-vmware_bas 15                                                                                        192kB] [34709 vpxa-worker 16844kB] [34001 hostd-worker 40092kB]] [[35899 sfcb-vm                                                                                        ware_raw 6%max] [35367 sfcb-pycim 19%max] [37143 sfcb-vmware_bas 19%max]]

2015-08-04T13:00:01Z heartbeat: up 0d1h18m22s, 0 VMs; [[35367 sfcb-pycim 15032kB 

[hongchuong]

Thank Anjani_Kumar‌

I will check the physical server.

Thank you .