vMotion of large size VMs attached to NSX-T segments fails

Products

• VMware NSX
• VMware vSphere ESXi

Issue/Introduction

• Environment has NSX-T 3.0.x deployed. 
• vMotion of small size VMs complete successfully.
• For Large VMs, where a vMotion takes approx. an hour and then fails. 
• On the destination ESXi host log /var/run/log/nsx-syslog, you should see:

nsx-opsagent[2100365]: NSX 2100365 - [nsx@6876 comp="nsx-esx" subcomp="opsagent" s2comp="nsxa" tid="2100670" level="INFO"] [DoVifPortOperation] request=[opId:[########-1813###-auto-#####-h5:700928##-##-##-##-##0e-11] op:[HOSTD_ATTACH_PORT(1)] vif:[90c48f52-####-####-####-########4c3] ls:[11b69099-####-####-####-########240] vmx:[/vmfs/volumes/########-####-########a660/##000005###/vm##.vmx] lp:[]]
... 
nsx-opsagent[2100365]: NSX 2100365 - [nsx@6876 comp="nsx-esx" subcomp="opsagent" s2comp="nsxa" tid="2100668" level="INFO"] [DoVifPortOperation] request=[opId:[sync-detach-0] op:[SYNC_DETACH_PORT(1002)] vif:[90c48f52-####-####-####-########4c3] ls:[11b69099-####-####-####-########240] vmx:[/vmfs/volumes/5fdc4609-####-####-####-########660/##000005###/vm##.vmx] lp:[3ccb1acb-####-####-####-########f94]]

Environment

VMware NSX-T Data Center 3.0.x

 

Cause

On the ESXi host, NSX-T has a resync timer which runs every hour to check for stale or unused ports. If the vMotion does not complete before the next resync timer run, it will consider the port unused and remove it.

Resolution

This issue is resolved in NSX-T Data Center 3.1 available at Support Documents and Downloads (broadcom.com).

Workaround

To allow the vMotion of large size VMs to complete, you can extend the resync timer, see details below:

• Login as root on the ESXi host.
• Make a backup of the file /etc/vmware/nsx-opsagent/nsxa.json

cp /etc/vmware/nsx-opsagent/nsxa.json /etc/vmware/nsx-opsagent/nsxa.json.bak

• Use vi and edit the file as below:

#vi /etc/vmware/nsx-opsagent/nsxa.json

• Change the default value from:

 "resync" : {
      "disabled" : "false",
      "intervalInSec" : 3600
   } 

To:

 "resync" : {
      "disabled" : "false",
      "intervalInSec" : 86400
   }

• Save and quit the file:

#:wq

• Restart opsagent using the command below on ESXi Host: 

/etc/init.d/nsx-opsagent restart

• Repeat the above steps for all impacted Hosts.