How to fix a VMWare disk of a broken machine that has snapshots

You may be out of luck but if you don't have a backup it may be work taking a structured approach to this.

With the VM powered off make a copy of all the files in the VM folder and put them away somewhere safe, they may be damaged but you want to have a complete copy in case you want to start over.

From an SSH shell or with an SCP utility (Veeam FastSCP or Tunnellier are handy if your using Windows) locate the VM folder and find all of the files with .vmdk extensions. There should be two for the base disk (a small descriptor file called VMname1.vmdk and a much larger file that contains the actual disk data called VMname1-flat.vmdk) and a similar pair for each of your snapshot delta's (a small descriptor called VMname1-00001.vmdk and a larger file called VMname1-00001-delta.vmdk that contains the changes made to the base disk during the lifetime of this snapshot). These are also referred to in the snapshot management file (the VMname.vmsd file) that contains all of the metadata about all active snapshots on a VM. The Virtual Machine itself has all its settings stored in the .VMX file including (importantly) the reference to the location of the current active VMDK for it's disk(s) - this will point to the latest snapshot delta.

VMware then uses internal ID's (called CID's) and a parameter called the ParentFileNameHint within the descriptor VMDK's to link each delta back to its parent (which may itself be a delta) recursively until it gets to the basedisk. A -delta on its own is not a complete disk it needs to be linked to an unmodified copy of its parent in order to work. If the parent file has been modified you are almost certainly out of luck.

You can manually move the delta files and their associated descriptor and fix scenarios where things have gotten lost but this will only work if no changes have been made to the base disks or any intermediate deltas. The descriptor files can all be edited with a text editor.

Once you have gotten your head around this you should read SANBarrow's CID chain repair guide that will help you figure out how to stitch things back together if they can be stitched back together.

Finally - never rely on Snapshots as Backups. They are not suitable for that purpose and they cause a significant additional IO overhead when active.

The snapshot file contains all new data written to disk since the snapshot was took. If you do not have the last snapshot you will lose all data written since the last snapshot. You will need to edit the vmdk file (the small text file, not the *flat.vmdk) and remove the snapshot reference.