Symptoms
NSX-T upgrade (3.0.0 -> 3.0.1) via SDDC Manager fails at "NSX T HOST CLUSTER UPGRADE COODINATOR" phase.SDDC Manager UI shows the following:
Description: Check if there are any failures for Host Cluster Upgrade Coordinator ChecksHealth Status: REDError Description: Failed to execute check for sufficient free space on /tmp partition of all ESXi hosts.Impact: High: Do not perform upgrade without addressing the issue. NSX-T Upgrade Coordinator Checks Failed with an error.Remediation: Please remediate the component precheck failures
In theupgrade-coordinator.log file, you see the entries similar to:
2020-09-18T04:32:00.197Z INFO pool-21-thread-1 UpgradeServiceImpl - SYSTEM [nsx@6876 comp="nsx-manager" level="INFO" subcomp="upgrade-coordinator"] Executing check Check for sufficient free space on tmp partition of all ESXi hosts2020-09-18T04:32:00.197Z INFO pool-21-thread-1 UpgradeMessagingServiceImpl - SYSTEM [nsx@6876 comp="nsx-manager" level="INFO" subcomp="upgrade-coordinator"] Sending RPC to localhost for 40d07bfc-2d84-484f-b739-594d5b6bd4f6: [com.vmware.nsx.sfhc.SfhcMessage] message_type: MT_RUN_COMMANDoperation_id: "b1dbf8c0-2fd2-44ce-be1a-ce32f350098d"service_fabric_object { fabric_object_identifier: "d089712d-5818-4259-9242-31cc3f199afd" runCommand { command: "vdf -p | grep tmp | sed \'s/ \\+/ /g\' | awk \'{print $3, $6}\'" }}return_same_correlationid: true Each ESXi host has additional tmp space. # vdf -p | grep tmptmp 2048 262144 7108 0 255036 8104amsd-tmp 0 256000 0 0 256000 0 <=====sut-tmp 0 256000 1340 0 254660 1340 <=====
Resolution
This issue is fixed in NSX-T 3.0.2 available at VMware Downloads.
Workaround
1. Upgrade manually via NSX-T Manager first.
Login to NSX-T manager using admin user. Select [System] -> [Upgrade] Select [CONTINUE WITH UPGRADE] [2. Edge] Select [NEXT] [3. Host] Select [START]
2. After the upgrade works fine via NSX-T Manager UI, return to SDDC Manager and select [RETRY UPDATE].
