Change Canton synchronization to use the healthcheck endpoint (#13463)

changelog_begin
changelog_end

bazel_tools/client_server/runner_with_health_check/BUILD.bazel

@@ -0,0 +1,26 @@
+# Copyright (c) 2022 Digital Asset (Switzerland) GmbH and/or its affiliates. All rights reserved.
+# SPDX-License-Identifier: Apache-2.0
+
+load("//bazel_tools:haskell.bzl", "da_haskell_binary")
+
+da_haskell_binary(
+    name = "runner_with_health_check",
+    srcs = ["Main.hs"],
+    hackage_deps = [
+        "base",
+        "extra",
+        "typed-process",
+        "process",
+        "async",
+        "http-client",
+        "http-conduit",
+        "http-types",
+        "text",
+        "safe",
+        "split",
+        "monad-loops",
+        "safe-exceptions",
+    ],
+    visibility = ["//visibility:public"],
+    deps = ["//libs-haskell/da-hs-base"],
+)

bazel_tools/client_server/runner_with_health_check/Main.hs

@@ -0,0 +1,40 @@
+-- Copyright (c) 2022 Digital Asset (Switzerland) GmbH and/or its affiliates. All rights reserved.
+-- SPDX-License-Identifier: Apache-2.0
+
+module Main(main) where
+
+import Data.Maybe (isJust)
+import System.Environment
+import Control.Exception.Safe
+import Control.Monad.Loops (untilJust)
+import System.Process.Typed
+import Data.List.Split (splitOn)
+import Control.Monad (guard, when)
+import Network.HTTP.Client (parseUrlThrow)
+import qualified Network.HTTP.Simple as HTTP
+import Control.Concurrent
+import System.Info.Extra
+import System.Process (terminateProcess)
+
+main :: IO ()
+main = do
+  [clientExe, clientArgs, serverExe, serverArgs, runnerArgs] <- getArgs
+  let splitArgs = filter (/= "") . splitOn " "
+  let serverProc = proc serverExe (splitArgs serverArgs)
+  healthPort : _ :: [Int] <- pure $ reverse $ map read $ splitArgs runnerArgs
+  withProcessTerm serverProc $ \ph -> do
+    waitForHealthcheck (threadDelay 500000) healthPort
+    runProcess_ (proc clientExe (splitArgs clientArgs))
+    -- See the comment on DA.Daml.Helper.Util.withProcessWait_'
+    when isWindows (terminateProcess $ unsafeProcessHandle ph)
+
+waitForHealthcheck :: IO () -> Int -> IO ()
+waitForHealthcheck sleep port = do
+    request <- parseUrlThrow $ "http://localhost:"  <> show port <> "/health"
+    untilJust $ do
+      r <- tryJust (\e -> guard (isIOException e || isHttpException e)) $ HTTP.httpNoBody request
+      case r of
+        Left _ -> sleep *> pure Nothing
+        Right _ -> pure $ Just ()
+  where isIOException e = isJust (fromException e :: Maybe IOException)
+        isHttpException e = isJust (fromException e :: Maybe HTTP.HttpException)

ledger/ledger-api-test-tool-on-canton/BUILD.bazel

@@ -60,6 +60,7 @@ conformance_test(
         "@jq_dev_env//:jq",
         "@bazel_tools//tools/jdk",
     ],
+    extra_runner_args = ["7000"],
     lf_versions = [
         "default",
         "latest",
@@ -70,7 +71,7 @@ conformance_test(
         5031,
         5041,
     ],
-    runner = "@//bazel_tools/client_server/runner_with_port_check",
+    runner = "@//bazel_tools/client_server/runner_with_health_check",
     server = ":canton-test-runner-with-dependencies",
     server_args = [],
     test_tool_args = [
@@ -121,6 +122,7 @@ conformance_test(
         "@jq_dev_env//:jq",
         "@bazel_tools//tools/jdk",
     ],
+    extra_runner_args = ["7000"],
     lf_versions = [
         "default",
         "latest",
@@ -131,7 +133,7 @@ conformance_test(
         5031,
         5041,
     ],
-    runner = "@//bazel_tools/client_server/runner_with_port_check",
+    runner = "@//bazel_tools/client_server/runner_with_health_check",
     server = ":canton-test-runner-with-dependencies",
     test_tool_args = [
         "--verbose",

ledger/ledger-api-test-tool-on-canton/canton-test-runner.sh

@@ -15,29 +15,6 @@ CANTON_COMMAND=(
   "--bootstrap=$(rlocation com_github_digital_asset_daml/ledger/ledger-api-test-tool-on-canton/bootstrap.canton)"
 )
 
-PARTICIPANT_1_HOST=localhost
-PARTICIPANT_1_MONITORING_PORT=7000
-
-TIMEOUT=60
-
-function wait_until() {
-  local start
-
-  start="$(date +%s)"
-  while true; do
-    if [[ "$(("$(date +%s)" - start))" -gt "$TIMEOUT" ]]; then
-      echo >&2 "Timed out after ${TIMEOUT} seconds."
-      return 1
-    fi
-
-    if "$@" >&/dev/null; then
-      return 0
-    fi
-
-    sleep 1
-  done
-}
-
 command=("${CANTON_COMMAND[@]}" "$@")
 
 export UNIQUE_CONTRACT_KEYS="$(rlocation com_github_digital_asset_daml/ledger/ledger-api-test-tool-on-canton/unique-contract-keys.conf)"
@@ -77,8 +54,4 @@ function stop() {
 
 trap stop EXIT INT TERM
 
-wait_until curl -fsS "http://${PARTICIPANT_1_HOST}:${PARTICIPANT_1_MONITORING_PORT}/health"
-
-echo >&2 'Canton is up and running.'
-
 wait "$pid"

ledger/ledger-api-tests/conformance.bzl

@@ -18,6 +18,7 @@ def conformance_test(
         test_tool_args = [],
         tags = [],
         runner = "@//bazel_tools/client_server/runner_with_port_check",
+        extra_runner_args = [],
         lf_versions = ["default"],
         flaky = False):
     for lf_version in lf_versions_aggregate(lf_versions):
@@ -27,7 +28,7 @@ def conformance_test(
             client_server_test(
                 name = test_name,
                 runner = runner,
-                runner_args = ["%s" % port for port in ports],
+                runner_args = ["%s" % port for port in ports] + extra_runner_args,
                 timeout = "long",
                 client = "//ledger/ledger-api-tests/tool:tool-%s" % lf_version,
                 client_args = test_tool_args + ["localhost:%s" % port for port in ports],