fix: dkg use different ports and adjust timeout (#435)

* fix: dkg use different ports and adjust timeout

* fix: change shutdown delay to 30s

Author: iamquang95

.github/workflows/dkg-runner.yml

@@ -105,8 +105,11 @@ jobs:
           CHARON_BIN: ${{ github.workspace }}/bin/charon
           PLUTO_BIN: ${{ github.workspace }}/target/debug/pluto
           TIMEOUT: ${{ inputs.timeout || '120' }}
+          SHUTDOWN_DELAY: "30s"
+          NODE_EXIT_TIMEOUT: "180"
           RUN_SMOKE_VERIFY: "1"
-          SMOKE_SECONDS: "3"
+          SMOKE_SECONDS: "8"
+          SMOKE_PORT_BASE: "19000"
         run: ./scripts/dkg-runner/run.sh
 
       - name: Upload work dir on failure

scripts/dkg-runner/README.md

@@ -60,13 +60,13 @@ All variables are optional. Set them in the environment before calling any scrip
 | `FEE_RECIPIENT` | `0xDeaDBeef…` | Fee recipient address for the cluster |
 | `WITHDRAWAL_ADDR` | `0xDeaDBeef…` | Withdrawal address for the cluster |
 | `TIMEOUT` | `120` | Seconds to wait before declaring the ceremony failed |
-| `SHUTDOWN_DELAY` | `30s` | Graceful shutdown delay passed to each node via `--shutdown-delay` |
-| `NODE_EXIT_TIMEOUT` | `90` | Seconds to wait for node processes to exit cleanly after artifacts appear |
+| `SHUTDOWN_DELAY` | `120s` | Graceful shutdown delay passed to each node via `--shutdown-delay` |
+| `NODE_EXIT_TIMEOUT` | `180` | Seconds to wait for node processes to exit cleanly after artifacts appear |
 | `PLUTO_BIN` | `./target/debug/pluto` | Path to the Pluto binary (only required when `PLUTO_NODES > 0`) |
 | `CHARON_BIN` | `charon` | Path to the Charon binary |
 | `RUN_SMOKE_VERIFY` | `1` | Smoke-start the collected node dirs with `charon run` after output collection |
-| `SMOKE_SECONDS` | `8` | Seconds the smoke-started nodes must stay alive |
-| `SMOKE_PORT_BASE` | `39000` | First local port used by smoke verification |
+| `SMOKE_SECONDS` | `8` | Seconds to wait for smoke validator APIs to become ready |
+| `SMOKE_PORT_BASE` | `19000` | First local port used by smoke verification |
 | `WORK_DIR` | `/tmp/dkg-run` | Scratch directory — wiped at the start of every run |
 | `KEEP_NODES` | `0` | Leave node processes running after a successful ceremony when set to `1`/`true`/`yes`/`on` |
 | `CI` | _(unset)_ | When truthy, suppresses per-node tee to stdout; logs go to `WORK_DIR/node-*/node.log` only |
@@ -83,7 +83,7 @@ All variables are optional. Set them in the environment before calling any scrip
 | 4 | `wait-node-exits.sh` | Waits for each node process to exit with status `0` unless `KEEP_NODES` is enabled |
 | 5 | `collect.sh` | Copies keystores and `cluster-lock.json` to `WORK_DIR/output/`; prints a summary |
 | 6 | `ci/verify-output-semantic.sh` | Validates the collected output is internally consistent across nodes |
-| 7 | `ci/verify-run-smoke.sh` | Starts the collected node dirs with `charon run` and checks they stay up through the smoke window |
+| 7 | `ci/verify-run-smoke.sh` | Starts the collected node dirs with `charon run` and checks every validator API reaches readiness |
 
 On success, outputs are under `$WORK_DIR/output/`. On failure or timeout, partial outputs are still collected and `WORK_DIR` is preserved for inspection. `run.sh` never deletes `WORK_DIR`; use `./scripts/dkg-runner/reset.sh` when you're done.
 

scripts/dkg-runner/ci/verify-run-smoke.sh

@@ -8,22 +8,22 @@
 #   SMOKE_SECONDS  seconds allowed for monitoring endpoints to become ready
 #                  (default: 8)
 #   SMOKE_PORT_BASE
-#                  first local port used by this check (default: 39000)
+#                  first local port used by this check (default: 19000)
 #
 # This verifies the generated full node data dirs are loadable by a later
 # Charon/Pluto-style runtime: cluster lock, p2p key, validator keystores, and
 # passwords are all usable enough for the process to start.
 #
 # It does not prove real beacon duties. It uses Charon simnet mocks and kills
-# the processes after the smoke window.
+# the processes after every validator API reaches readiness.
 
 set -euo pipefail
 
 WORK_DIR="${WORK_DIR:-/tmp/dkg-run}"
 NODES="${NODES:-4}"
 CHARON_BIN="${CHARON_BIN:-charon}"
 SMOKE_SECONDS="${SMOKE_SECONDS:-8}"
-SMOKE_PORT_BASE="${SMOKE_PORT_BASE:-39000}"
+SMOKE_PORT_BASE="${SMOKE_PORT_BASE:-19000}"
 SMOKE_DIR="${WORK_DIR}/run-smoke"
 
 fail() {

scripts/dkg-runner/config.sh

@@ -11,15 +11,15 @@
 : "${CHARON_NODES:=2}"
 : "${RELAY_URL:=https://0.relay.obol.tech}"
 : "${TIMEOUT:=120}"
-: "${SHUTDOWN_DELAY:=5s}"
-: "${NODE_EXIT_TIMEOUT:=90}"
+: "${SHUTDOWN_DELAY:=120s}"
+: "${NODE_EXIT_TIMEOUT:=180}"
 : "${PLUTO_BIN:=./target/debug/pluto}"
 : "${CHARON_BIN:=charon}"
 : "${WORK_DIR:=/tmp/dkg-run}"
 : "${KEEP_NODES:=0}"
 : "${RUN_SMOKE_VERIFY:=1}"
 : "${SMOKE_SECONDS:=8}"
-: "${SMOKE_PORT_BASE:=39000}"
+: "${SMOKE_PORT_BASE:=19000}"
 : "${NETWORK:=holesky}"
 : "${FEE_RECIPIENT:=0xDeaDbeefdEAdbeefdEadbEEFdeadbeEFdEaDbeeF}"
 : "${WITHDRAWAL_ADDR:=0xDeaDbeefdEAdbeefdEadbEEFdeadbeEFdEaDbeeF}"

scripts/dkg-runner/run.sh

@@ -12,18 +12,20 @@
 #   RELAY_URL=https://0.relay.obol.tech
 #                        Relay ENR endpoint used by the DKG nodes.
 #   TIMEOUT=120          Seconds to wait for all nodes before aborting.
-#   NODE_EXIT_TIMEOUT=90 Seconds to wait for nodes to exit after completion.
+#   SHUTDOWN_DELAY=120s Graceful shutdown delay passed to each node.
+#   NODE_EXIT_TIMEOUT=180
+#                        Seconds to wait for nodes to exit after completion.
 #   PLUTO_BIN=./target/debug/pluto
 #                        Path to the Pluto binary.
 #   CHARON_BIN=charon    Path to the Charon binary.
 #   WORK_DIR=/tmp/dkg-run
 #                        Scratch directory for the run (wiped on every call).
 #   KEEP_NODES=0         Leave nodes running after a successful ceremony when
 #                        set to 1/true/yes/on.
-#   RUN_SMOKE_VERIFY=0   Smoke-start generated node dirs with charon run after
+#   RUN_SMOKE_VERIFY=1   Smoke-start generated node dirs with charon run after
 #                        successful output collection.
-#   SMOKE_SECONDS=8      Seconds smoke-started nodes must stay alive.
-#   SMOKE_PORT_BASE=39000
+#   SMOKE_SECONDS=8      Seconds to wait for smoke validator APIs to become ready.
+#   SMOKE_PORT_BASE=19000
 #                        First local port used by runtime smoke verification.
 #   NETWORK=holesky      Ethereum network for the cluster definition.
 #   FEE_RECIPIENT=0xDeaD...
@@ -110,6 +112,7 @@ log_info "  CHARON_NODES = ${CHARON_NODES}"
 log_info "  RELAY_URL    = ${RELAY_URL}"
 log_info "  NETWORK      = ${NETWORK}"
 log_info "  TIMEOUT      = ${TIMEOUT}s"
+log_info "  SHUTDOWN_DELAY = ${SHUTDOWN_DELAY}"
 log_info "  NODE_EXIT_TIMEOUT = ${NODE_EXIT_TIMEOUT}s"
 log_info "  PLUTO_BIN    = ${PLUTO_BIN}"
 log_info "  CHARON_BIN   = ${CHARON_BIN}"