Remediating an ESXi 5.x or 6.0 host fails with the error: The host returns esxupdate error code:15. The package manager transaction is not successful

 Symptoms

You cannot remediate an ESXi 5.x or 6.0 host using vCenter Update Manager.
Remediating ESXi 5.x or 6.0 hosts fails.
A package is to be updated on the host, particularly when VMware_locker_tools-light* is corrupt.

You see the error:

error code:15. The package manager transaction is not successful. Check the Update Manager log files and esxupdate log files for more details.
Cause

This issue occurs if the package files for floppies in the /locker/packages/Version/ folder is corrupt or full.

For example:
    In ESXi 5.0 systems – /locker/packages/5.0.0/
    In ESXi 5.1 systems – /locker/packages/5.1.0/
    In ESXi 5.5 systems – /locker/packages/5.5.0/
    In ESXi 6.0 systems – /locker/packages/6.0.0/

Resolution

To resolve this issue, recreate the/locker/packages/version/ folder, where version is:
    ESXi 5.0 – /locker/packages/5.0.0/
    ESXi 5.1 – /locker/packages/5.1.0/
    ESXi 5.5 – /locker/packages/5.5.0/
    ESXi 6.0 – /locker/packages/6.0.0/

To verify the store folders contents and symbolic link:

 Connect to the ESXi host using an SSH session.
Check for information in the /store folder by running this command:
    ls /store

 This folder must contain packages and var folder.
 Run this command to verify that the symbolic link is valid:
    ls -l /

 The /store folder should be linked to /locker and appear as:
    locker  -> /store

 If that link is not displayed, run this command to add the symbolic link:
    ln -s /store /locker

To recreate the/locker/packages/version/ folder:
    Put the host in the maintenance mode.
    Navigate to the /locker/packages/version/ folder on the host.
    Rename /locker/packages/version/ folder to /locker/packages/version.old.
    Remediate the host using Update Manager.

The /locker/packages/version/ folder is recreated and the remediation should now be successful.

Note: Verify if you can change to the other folders in /locker/packages/version/. If not, rename all the three folders including floppies.

An alternative resolution for ESXi:
    Put the host in the maintenance mode.
    Navigate to the /locker/packages/version/ folder on the host.

Rename the folder to:
    /locker/packages/ version.old

Run this command as the root user to recreate the folder:
 mkdir / locker/packages/ version/

For example:

 In ESXi 5.0:
    mkdir / locker/packages/5.0.0/

In ESXi 5.1:
    mkdir / locker/packages/5.1.0/

In ESXi 5.5:
mkdir / locker/packages/5.5.0/

In ESXi 6.0:
 mkdir / locker/packages/6.0.0/

 Use WinSCP to copy the folders and files from the / locker/packages/ version/ directory on a working host to the affected host.

If the preceding methods do not resolve the issue:

 Verify and ensure that there is sufficient free space on root folder by running this command
    vdf -h

Check the locker location by running this command:
    ls -ltr /

 If the locker is not pointing to a datastore:
Rename the old locker file by running this command:
    mv /locker /locker.old

Recreate the symbolic link by running this command:
    ln -s /store /locker