## TL;DR
A missing parent in a snapshot chain orphans every delta built on it, because deltas store only differences. First freeze the VM to prevent more writes, search all datastores and backups for the parent file, and only if it is truly gone, plan a restore from backup or a specialist chain repair using descriptor edits as a last resort with full copies taken first.

```text
snapshot chain broken missing parent
```

## Use this when
- The VM fails to power on citing a missing parent disk or CID mismatch after manual file moves.
- A datastore cleanup deleted a base VMDK that had snapshots.
- Descriptor files reference a parent filename that does not exist.

## Not for this skill when
- All chain files exist but consolidation fails. That is a merge problem.
- A single standalone disk is missing with no snapshots. That is simple data loss.

## Steps
1. Power off the VM and copy its descriptor and delta files to safe storage before touching anything. Verify: forensic copies of the chain metadata exist
2. Search every datastore, backup and replica for the exact parent filename and any copy of that disk. Verify: the parent is recovered or confirmed absent everywhere
3. If a parent copy is found, restore it to the original location and check the chain CID values align in the descriptors. Verify: the chain describes a consistent lineage
4. If the parent is gone, restore the VM from the newest backup instead of attempting to fabricate the parent. Verify: the restored VM boots and data is current to the backup point
5. If no backup exists, engage specialist recovery with the preserved descriptors and accept partial data recovery expectations. Verify: a recovery plan with stated limits is agreed before work starts

## Variant phrasings
### snapshot chain missing parent vmdk recover
The recovery phrasing.
### vmware parent disk missing cid mismatch
The CID variant.
### broken snapshot chain deleted base disk
The deletion wording.

Compatibility: VMware vSphere 7 and 8 snapshot chains and Hyper-V checkpoint differencing disks, VMFS, NFS and vSAN datastores.

## Why it happens
Differencing disks are not self-contained: each stores changes against its parent, so the parent is part of the data. Manual datastore cleanups, partial copies and mistaken deletes remove what looks like an unused old file while it is actually the foundation of the running VM. Descriptor parent hints and CID values exist to detect exactly this breakage at power-on.

## Edge cases / pitfalls
- Never edit CID values to force a boot without understanding the mismatch. A wrong parent silently corrupts data.
- Datastore cleanup scripts must check for snapshot children before deleting any VMDK.
- Flat and descriptor files are a pair. Recovering only one half still leaves the chain broken.

## Provenance

Resolved from the public thread: https://vectle.com/posts/pst_obBL7RFNrCfmHGe9MldYGw