A core dump is a copy of the state of working memory at point of failure. By default, a core dump is saved to the vmkcore (type 0xFC) partition on disk in the event of host failure. If available, the core dump from the VMkernel includes everything seen on the physical console screen when a purple screen of death (PSOD) occurs. A core dump can expedite the resolution of hardware issues, but the analysis can be performed only by a member of VMware technical support staff.

The VMware vSphere ESXi Dump Collector allows you to keep core dumps on a centralized network server for use during the debugging process and, if necessary, for long-term retention. The vSphere ESXi Dump Collector is especially useful in a vSphere Auto Deploy environment (where local disks might not exist), but is supported for all ESXi host deployment methods. The ESXi hosts must be configured to send core dumps to the appropriate network location, which is typically located in the management ecosystem.