Why vMotion halt virutal machines for few seconds

[chen_zhe]

Currently, we used vsphere 6.0u2. We have made some tests on vMotion, and found a strange problem that virtual machine will be in unresponsive status for few seconds during the migration process. Normally, we expect that vMotion will only make the virtual machine lost in 200ms~300ms. Stuck in few seconds is not acceptable for our app due to it is sensitive to the network status.

After check the virtual machine log, we get the content as:

2017-07-13T09:48:22.693Z| vcpu-0| I120: MigrateSetState: Transitioning from state 11 to 12.

2017-07-13T09:48:22.693Z| vcpu-0| I120: Migrate: Restoring original message callbacks.

2017-07-13T09:48:22.694Z| vcpu-0| I120: Migrate_SetSuccess: Reopening config file.

2017-07-13T09:48:22.694Z| vcpu-0| A115: ConfigDB: Setting config.readOnly = "FALSE"

2017-07-13T09:48:24.777Z| vmx| I120: MigrateESXVmkEventHandler: migration not in progress, not handling message 6.

2017-07-13T09:48:24.777Z| vcpu-0| I120: SVGA: Registering MemSpace at 0xec000000(0xec000000) and 0xfe000000(0xfe000000)

2017-07-13T09:48:24.778Z| vcpu-0| I120: SCSI: switching scsi0 to push completion mode

2017-07-13T09:48:24.778Z| vcpu-0| I120: Vix: [4135076 mainDispatch.c:4292]: VMAutomation_ReportPowerOpFinished: statevar=0, newAppState=1872, success=1 additionalError=0

2017-07-13T09:48:24.778Z| vcpu-0| I120: Vix: [4135076 mainDispatch.c:4210]: VMAutomationReportPowerStateChange: Reporting power state change (opcode=0, err=0).

2017-07-13T09:48:24.778Z| vcpu-0| I120: Vix: [4135076 mainDispatch.c:4210]: VMAutomationReportPowerStateChange: Reporting power state change (opcode=2, err=0).

2017-07-13T09:48:24.778Z| vcpu-0| I120: Transitioned vmx/execState/val to poweredOn

2017-07-13T09:48:24.778Z| vcpu-0| I120: Migrate: cleaning up migration state.

2017-07-13T09:48:24.778Z| vcpu-0| I120: Migrate: Final status reported through Vigor.

2017-07-13T09:48:24.778Z| vcpu-0| I120: MigrateSetState: Transitioning from state 12 to 0.

It seem the migration encounter some problem, but we can not found the detail about why. Does any one know about this and how to avoid vMotion take a long time halt?

[Harishtoy]

How about VMtools installed on the VM? hope you have latest tools installed.

[chen_zhe]

VMtools has been installed. Version is 10246.

[faherne_CTI]

Is your storage network physically separate from your vMotion and VM production traffic network? e.g. Do you use a different switching infrastructure to separate this traffic?

[rmigliore]

I would like to add to faherne_CTI comment, What is the config for the vMotion network? The speed and configuration of your vMotion network coupled with the size of the VM being moved could cause the performance issues.