[automated] docs: fix incorrect claim about compressed layer deduplication in containerd image store

[dvdksn]

🤖 Generated with Claude Code

Summary

The "Disk space usage" section incorrectly stated that the containerd image store stores per-image copies of shared compressed layers. This is factually wrong.

• The containerd content store is content-addressed: compressed blobs are stored once per unique digest and shared across all images that reference them — the same deduplication behavior as legacy storage drivers.
• The actual source of increased disk usage is dual storage: containerd keeps both the compressed blob (content store, by digest) and the unpacked snapshot (snapshotter), whereas legacy drivers discarded the compressed form after unpacking.
• The overhead is therefore per unique layer, not per image referencing that layer.

The fix replaces the incorrect paragraph with an accurate description of both deduplication behavior and the real source of overhead.

Closes #24476

[netlify]

✅ Deploy Preview for docsdocker ready!

Name	Link
🔨 Latest commit	1c906d9
🔍 Latest deploy log	https://app.netlify.com/projects/docsdocker/deploys/69c3c2effb19760009437993
😎 Deploy Preview	https://deploy-preview-24496--docsdocker.netlify.app
📱 Preview on mobile	Toggle QR Code... Use your smartphone camera to open QR code link.

---

To edit notification comments on pull requests, go to your Netlify project configuration.

[docker-agent]

Assessment: 🟢 APPROVE

This PR correctly fixes a factual inaccuracy in the documentation. The original text incorrectly stated that the containerd image store duplicates compressed layers across images ("each image stores its own compressed version of shared layers").

The corrected text accurately explains that:

• Compressed layers ARE deduplicated by digest in the content store (same as legacy drivers)
• The overhead comes from dual storage (compressed + unpacked), not from duplication

The fix improves technical accuracy and aligns with how containerd's content-addressed storage actually works. No documentation quality issues found.

[mike-sul]

I suggest removing the paragraph starting with "This difference is particularly noticeable with multiple images sharing the same base layers" altogether, as the opening sentence contradicts the explanation. Since shared layers are deduplicated, there’s nothing particularly noticeable here - the real difference is the dual storage (compressed and unpacked) of each unique layer, which is already covered in the previous paragraphs.

[dvdksn]

Good catch, @mike-sul — removed the paragraph entirely. The opening sentence was indeed self-defeating: if deduplication works the same as legacy drivers, there's nothing particularly noticeable about the shared-layer case. The dual-storage explanation in the preceding paragraph is sufficient.