You are unable to successfully configure a dedicated local coredump partition on an ESXi host. When attempting to create or copy the diagnostic dump file to the local disk, the configuration fails and you receive the following error message in the command line interface:
WARNING: Start offset less than header size!
This issue occurs when the local disk selected for the coredump partition contains pre-existing partitions. Because the disk is not blank and the existing data cannot be safely verified or overwritten, the ESXi host rejects it for use as a dedicated diagnostic partition.
If a suitable, unpartitioned local disk is not available, you can resolve this by configuring a coredump file on a shared network file system (VMFS datastore) instead of using a dedicated local partition.
Follow these steps to configure and activate a core dump file on a VMFS datastore:
Open a command-line session (SSH) to the affected ESXi host.
Navigate to the disks directory:
cd /dev/disks
List the available storage file systems to identify the specific UUID of the shared network storage datastore you intend to use:
esxcli storage filesystem list
Add the coredump file to the selected datastore using the UUID identified in the previous step:
esxcli system coredump file add -d <UUID>
Verify that the core dump file has been successfully created:
esxcli system coredump file list
Activate the core dump file for the host. Using the smart selection algorithm will automatically enable the newly created file:
esxcli system coredump file set --smart --enable true
Verify that the core dump file is active and configured. Both the "Active" and "Configured" columns should display true:
esxcli system coredump file listFor further details on configuring and managing core dump files and partitions, please refer to the following official documentation: