The installation and upgrade software enables you to identify problems on the host machine that can cause an installation, upgrade, or migration to fail.

For interactive installations, upgrades, and migrations, the errors or warnings are displayed on the final panel of the installer, where you are asked to confirm or cancel the installation or upgrade. For scripted installations, upgrades, or migrations, the errors or warnings are written to the installation log file. You can also consult the product release notes for known problems.

vSphere Update Manager provides custom messages for these errors or warnings. To see the original errors and warnings returned by the precheck script during an Update Manager host upgrade scan, review the Update Manager log file vmware-vum-server-log4cpp.log.

The vSphere Ugrade guide describes how to use VMware products and their features. If you encounter problems or error situations that are not described in this guide, you may find a solution in VMware Knowledge Base. You can also use VMware Community Forums to find others with same problem or ask for help, or you can open Support Request to get help from VMware service professional.

Collect Installation Logs for vCenter Server Appliance

You can collect installation log files and check these files to identify the source of a failure if vCenter Server appliance stops responding during initial startup.

Procedure

  1. Access the vCenter Server appliance shell.
    Option Description
    If you have direct access to the appliance Press Alt+F1.
    To connect remotely Use SSH or another remote console connection to start a session to the appliance.
  2. Enter a user name and password that the appliance recognizes.
  3. In the appliance shell, run the pi shell command to access the Bash shell.
  4. In the Bash shell, run the vc-support.sh script to generate a support bundle.
    This command generates a .tgz file in /storage/log.
  5. Export the generated support bundle to the user@x.x.x.x:/tmp folder.
    scp /var/tmp/vc-etco-vm-vlan11-dhcp-63-151.eng.vmware.com-2014-02-28--21.11.tgz user@x.x.x.x:/tmp
  6. Determine which firstboot script failed.
    cat /var/log/firstboot/firstbootStatus.json

What to do next

To identify potential causes of the failure, examine the log file of the firstboot script that failed.

Errors and Warnings Returned by the Installation and Upgrade Precheck Script

The installation and upgrade precheck script runs tests to identify problems on the host machine that can cause an installation, upgrade, or migration to fail.

For interactive installations, upgrades, and migrations, the errors or warnings are displayed on the final screen of the GUI installer, where you are asked to confirm or cancel the installation or upgrade. For scripted installations, upgrades, or migrations, the errors and warnings are written to the installation log file.

vSphere Update Manager provides custom messages for these errors or warnings. To see the original errors and warnings returned by the precheck script during an Update Manager host upgrade scan, review the Update Manager log file vmware-vum-server-log4cpp.log.

Table 1. Error and Warning Codes That Are Returned by the Installation and Upgrade Precheck Script
Error or Warning Description
64BIT_LONGMODESTATUS The host processor must be 64-bit.
COS_NETWORKING Warning. An IPv4 address was found on an enabled service console virtual NIC that has no corresponding address in the same subnet in the VMkernel. A separate warning appears for each such occurrence.
CPU_CORES The host must have at least two cores.
DISTRIBUTED_VIRTUAL_SWITCH If the Cisco Virtual Ethernet Module (VEM) software is found on the host, the test checks that the upgrade also contains the VEM software. The test also determines whether the upgrade supports the same version of the Cisco Virtual Supervisor Module (VSM) as the existing version on the host. If the software is missing or is compatible with a different version of the VSM, the test returns a warning. The result indicates which version of the VEM software was expected on the upgrade ISO and which versions, if any, were found. You can use ESXi Image Builder CLI to create a custom installation ISO that includes the appropriate version of the VEM software.
HARDWARE_VIRTUALIZATION Warning. If the host processor does not have hardware virtualization, or if hardware virtualization is not turned on in the host BIOS, host performance reduces. You can enable hardware virtualization in the host machine boot options panel. See your hardware vendor's documentation.
MD5_ROOT_PASSWORD This test checks that the root password is encoded in MD5 format. If a password is not encoded in MD5 format, it might be significant only to eight characters. In this case, any characters after the first eight are no longer authenticated after the upgrade, which can create a security problem. To work around this problem, see Knowledge Base article KB 1024500.
MEMORY_SIZE The host requires the specified amount of memory to upgrade.
PACKAGE_COMPLIANCE vSphere Update Manager upgrade to vSphere Lifecycle Manager only. This test checks the existing software on the host against the software contained on the upgrade ISO to determine whether the host has been successfully upgraded. If any of the packages are missing or are an older version than the package on the upgrade ISO, the test returns an error. The test results indicate which software was found on the host and which software was found on the upgrade ISO.
PARTITION_LAYOUT You can upgrade or migrate software only if at most one VMFS partition on the disk is being upgraded. The VMFS partition must begin after sector 1843200.
POWERPATH This test checks for the installation of EMC PowerPath software, consisting of a CIM module and a kernel module. If either of these components is found on the host, the test checks that matching components, such as CIM or the VMkernel and module, also exist in the upgrade. If they do not exist, the test returns a warning that indicates which PowerPath components were expected on the upgrade ISO and which, if any, were found.
PRECHECK_INITIALIZE This test checks that the precheck script can be run.
SANE_ESX_CONF The /etc/vmware/esx.conf file must exist on the host.
SPACE_AVAIL_ISO vSphere Update Manager only. The host disk must have enough free space to store the contents of the installer CD or DVD.
SPACE_AVAIL_CONFIG vSphere Update Manager upgrade to vSphere Lifecycle Manager only. The host disk must have enough free space to store the legacy configuration between reboots.
SUPPORTED_ESX_VERSION You can upgrade or migrate to ESXi 8.0 only from version 6.0 ESXi hosts.
TBOOT_REQUIRED This message applies only to vSphere Update Manager upgrade to vSphere Lifecycle Manager only. The upgrade fails with this error when the host system is running in trusted boot mode (tboot), but the ESXi upgrade ISO does not contain any tboot VIBs. This test prevents an upgrade that can make the host less secure.
UNSUPPORTED_DEVICES Warning. This test checks for unsupported devices. Some PCI devices are not supported in ESXi 8.0.
UPDATE_PENDING

This test checks the host for VIB installations that require a reboot. This test fails if one or more such VIBs are installed, but the host has not yet been rebooted. In these conditions, the precheck script is unable to determine reliably which packages are currently installed on the host. If this test fails, it may not be safe to rely on the rest of the precheck tests to determine whether an upgrade is safe.

If you encounter this error, restart the host and retry the upgrade.

Upgrade Issues with vCenter Server Containing Host Profiles

Most common issues that might occur during a vCenter Server upgrade to version 8.0, that contains host profiles.

  • For issues occurring during a vCenter Server upgrade or ESXi upgrade, see Troubleshooting a vSphere Upgrade.
  • If upgrading vCenter Server 6.7 or 7.0, containing host profiles with version earlier than 6.7, results with a failure, see KB 52932.
  • For error There is no suitable host in the inventory as reference host for the profile Host Profile. The profile does not have any associated reference host, see KB 2150534.
  • If an error occurs when you import a host profile to an empty vCenter Server inventory, see vSphere Host Profiles for Reference Host is Unavailable.
  • If a host profile compliance check fails for NFS datastore, see vSphere Host Profiles for Host Profile without NFS Datastore.
  • If compliance check fails with an error for the UserVars.ESXiVPsDisabledProtocols option, when an ESXi host upgraded to version 8.0 is attached to a host profile with version 6.7, see VMware vSphere 8.0 Release Notes.

Roll Back a vCenter Server Instance on Windows When vCenter Server Upgrade Fails

You can roll back or restore a vCenter Server instance on Windows when an upgrade of vCenter Server with an external Platform Services Controller fails after the export stage and the legacy environment has been uninstalled.

Prerequisites

The roll back or restore of vCenter Server applies when all of the following conditions apply:

  • You must have access to the vCenter Server for Windows machine.
  • The vCenter Server instance is attached to an external Platform Services Controller.
  • The upgrade of the vCenter Server instance attached to the Platform Services Controller instance is in a failed state after the export stage and uninstallation of the legacy vCenter Server.
  • Ensure that vCenter Server rollback happened properly in case of upgrade failure and that no stale failed upgrade log entries remain.

For Rollback Method 1:

  • To clean up vCenter Server appliance 7.0 data from the legacy Platform Services Controller, see KB 2106736.
  • Use a vCenter Server database snapshot taken before the start of the vCenter Server upgrade.

For Rollback Method 2:

  • Use a powered off snapshot of the vCenter Server before the vCenter Server upgrade.
  • Use a Platform Services Controller snapshot taken after the Platform Services Controller node upgrade and before the start of the vCenter Server upgrade.
  • Use a vCenter Server snapshot taken after the Platform Services Controller upgrade and before the start of the vCenter Server upgrade.

Procedure

  • You can restore the legacy vCenter Server using Rollback Method 1 or Rollback Method 2.
    • Use Rollback Method 1.
      1. Manually clean up vCenter Server appliance data from Platform Services Controller.
      2. Restore the legacy vCenter Server database from a backup which was taken before the upgrade.
      3. Point the legacy vCenter Server instance to the Platform Services Controller, and also point to the database with the restored data.
      4. Ensure that the vCenter Server services are up and running.
    • Use Rollback Method 2.
      1. Restore the Platform Services Controller instance from a snapshot to the point where you were about to start vCenter Server upgrade. You can use a backup for a Windows configuration or use another backup and restore approach to revert the snapshot.
      2. Restore the vCenter Server instance from a snapshot.
      3. Restore the vCenter Server database from a snapshot.
      4. Ensure that the vCenter Server services are up and running.
    For Rollback Method 2, you will lose all data written to Platform Services Controller after the vCenter Server upgrade has been started when you restore from the Platform Services Controller snapshot taken before that point in time.

Collect Logs to Troubleshoot ESXi Hosts

You can collect installation or upgrade log files for ESXi. If an installation or upgrade fails, checking the log files can help you identify the source of the failure.

Solution

  1. Enter the vm-support command in the ESXi Shell or through SSH.
  2. Navigate to the /var/tmp/ directory.
  3. Retrieve the log files from the .tgz file.