docs: move provisioned layers out of Later

The keyed-layer design shipped: provision --spec hashes the spec with
its base reference and --layer names the entry for run --layer. The
row was describing the past as the future; what remains is eviction.
This commit is contained in:
vhaudiquet committed 2026-10-30 15:30:22 +01:00
1 parent 3d999b7be0
commit e632fa2125
1 file changed
+4 -7
+4 -7
View File
@@ -279,13 +279,10 @@ build is deleted.
correct-and-slow v1 posture graduates into this. correct-and-slow v1 posture graduates into this.
- **Guest agent, long-lived machines.** Exec without a boot; the shape - **Guest agent, long-lived machines.** Exec without a boot; the shape
to grow into when per-command VMs stop being right. to grow into when per-command VMs stop being right.
- **Provisioned layers, keyed.** `provision` today bakes one entry - **Layer eviction.** Provisioned layers are already keyed: a spec
per image; the next step keys entries by (base digest, hash of the file (base, packages, run) hashed with its base reference, named by
provisioning spec) so consumers each get their own layer over a `--layer` for `run --layer`. Nothing prunes the derived entries yet;
shared base, the spec file living in the repo and the artifact eviction (LRU or an explicit prune) keeps them bounded.
reproducible from it. Marker-hit semantics then skip the command
only for the identical spec, and eviction (LRU or an explicit
prune) keeps derived entries bounded.
- **passt** (`krun_set_passt_fd`) if TSI throughput or UDP listening - **passt** (`krun_set_passt_fd`) if TSI throughput or UDP listening
ever matters. ever matters.
- **An agent in the guest** over port publishing: an http backend - **An agent in the guest** over port publishing: an http backend