If an ESXi-host crashes and shows a “Purple Screen Of Death” usually a coredump is created. This article shows how to save the coredump, if you need to send it to VMware for further analysis.

Preparation

  1. Remember the “Partition-ID” of the coredump shown in the “Purple Screen Of Death”
    Purple Screen of Death shows partition ID it saved the coredump to
  2. After the crash restart the ESXi-host and wait till it has properly booted.
  3. Login as “root” into the ESXi host by SSH or through the console.
  4. Find the coredump partition with one of these two commands 
    1. esxcli system coredump partition list
      You’ll get a list of multiple partitions. Copy the path of active one (= it has “true” in column “Active”). We will need the partitions “Name”.
      List of local partitions coredumps might be saved to 
    2. esxcfg-dumppart -t
      You’ll only get the Name of the active coredump partition without it’s path.
      Command to show the partition ID coredumps are saved to

Save the coredump

  1. Find a datastore mounted to the ESXi-host that has sufficient space for the coredump file and chose or create a directory you want to save the file to. The coredump file can be several Gigabyte in size.
  2. Save the coredump with the following command to the chosen datastore: 
    esxcfg-dumppart --copy --devname "/vmfs/devices/disks/<Partition-ID>" --zdumpname /vmfs/volumes/<datastore_name>/<directory><esxi-name>_<date>.zdump
    e.g. this will save the coredump to “/vmfs/volumes/datastore01/extracted-corefiles/myesxi01-2019-05-22.zdump” 
    esxcfg-dumppart --copy \
                --devname "/vmfs/devices/disks/naa.6006016007603800181bd37a7f85e611:9" \
                --zdumpname /vmfs/volumes/datastore01/extracted-corefiles/myesxi01-2019-05-22.zdump
  3. Now download the file to your local PC by SCP or the datastore browser and - if requested - upload it to VMware support.