Skip to content

cut(e2e): envdsmoke moves to sandbox-operator; sandbox keeps no envd knowledge #263

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
CMGS merged 2 commits into from
Sep 28, 2026
Merged

cut(e2e): envdsmoke moves to sandbox-operator; sandbox keeps no envd knowledge #263

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
62650a1
cut(e2e): envdsmoke moves to sandbox-operator; sandbox keeps no envd …
CMGS Sep 28, 2026
2a718b9
cut(e2e): portsmoke's comments name no guest daemon
CMGS Sep 28, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 0 additions & 1 deletion .gitignore
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3,7 +3,6 @@ dist/
boot/init/target/
silkd/target/
e2e/demo
e2e/envdsmoke
e2e/guestserver
e2e/portsmoke
e2e/smoke
Expand Down
2 changes: 1 addition & 1 deletion README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -59,7 +59,7 @@ performance) — source in
`smoke`, `meshsmoke`, `crossnode`, `coldproof`, `egresssmoke`,
`interceptsmoke`, `sockssmoke`, `volumesmoke`, `lifecycle` (idle→hibernate→archive),
`androidsmoke`, `browsersmoke`, `ringsmoke` (output ring cap, exec as a user),
`portsmoke` (the guest-port relay), `envdsmoke` (the e2b flavor's envd through it),
`portsmoke` (the guest-port relay),
and the `pullbench`/`pushbench`/`rpcbench`/`qaab` perf drivers
- `boot/kernel/` — kernel version pin (`VERSION` + matching tarball `SHA256`,
bump both together) + config fragment (amd64: over `x86_64_defconfig` +
Expand Down
24 changes: 7 additions & 17 deletions docs/e2b.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -130,23 +130,13 @@ window where that lock is absent.

## Proving it

`scripts/envd-e2e.sh` runs the guest half on a node: it builds the pool with the
warmup gate above and then drives `envdsmoke`, which claims a sandbox and
reaches the real `envd` through the relay — health, `GET`/`POST /files`, a
ConnectRPC unary, and a `process.Process/Start` whose output comes back over the
Connect server stream — while asserting silkd still answers on the same VM.

```bash
K=<kit> TEMPLATE=ghcr.io/cocoonstack/sandbox/e2b-rt:24.04 bash scripts/envd-e2e.sh
K=<kit> TEMPLATE=ghcr.io/cocoonstack/sandbox/e2b-ci:24.04 SIZE=medium CODE_INTERPRETER=1 bash scripts/envd-e2e.sh
```

Both passes check the `user` account; `CODE_INTERPRETER=1` also gates warmup on
49999, curls its `/health` inside the guest, and runs `1+1` through the relay
the way the SDK's `runCode` does.

`envdsmoke -hold` then keeps that sandbox alive for an out-of-tree harness that
puts a real edge proxy in front of it.
The flavor's daemon is proved from the consumer's side: sandbox-operator's
`scripts/envd-e2e.sh` builds the pool with the warmup gate above and drives
`test/envdsmoke`, which claims a sandbox and reaches the real `envd` through
the relay (health, the version the compat API reports, files, ConnectRPC, the
`user` account, HTTP/1.1 only, and the code-interpreter API with
`CODE_INTERPRETER=1`). sandboxd itself knows nothing of `envd`: the pool's
warmup is a shell command like any other, and the relay carries bytes.

The suite does **not** cover the guard's drop. Reaching 49983 from off the
guest means lifting sandboxd's tap lock, which no e2e run should do. Measured
Expand Down
Loading
Loading