Merge #11789: [tests] [travis-ci] Combine logs on failure

ff8a9b0 [tests] Add combinedlogslen argument to test_runner.py (John Newbery)
dba94ea [tests] [travis-ci] Move Travis functional test log post processing to test_runner (John Newbery)
bba1c54 [tests] Improve logging shutdown and add hint for combine_logs (John Newbery)

Pull request description:

  Replaces #11779 . Notes from that PR:

  > Currently, when a functional test fails, the debug logs are printed sequentially to the travis log. This makes debugging race conditions based on the travis log hard. Instead, all logs events should be combined and sorted by their timestamp, then appended to the travis log.

  @MarcoFalke

Tree-SHA512: 56c80067d6a2c92f7e6a35e3ae5160637a0de052d9da593c7be6e02233544a93c66d62456f903f85e2edc09e31ab4bdafd1aed1d9897ae48c634f82631f856f7

3 files changed, 26 insertions, 29 deletions

diff --git a/.travis.yml b/.travis.yml
index a643d2ff5d..d3dd37e76c 100644
--- a/.travis.yml
+++ b/.travis.yml
@@ -18,7 +18,6 @@ env:
     - CCACHE_COMPRESS=1
     - BASE_OUTDIR=$TRAVIS_BUILD_DIR/out
     - SDK_URL=https://bitcoincore.org/depends-sources/sdks
-    - PYTHON_DEBUG=1
     - WINEDEBUG=fixme-all
   matrix:
 # ARM
@@ -79,7 +78,7 @@ script:
     - export LD_LIBRARY_PATH=$TRAVIS_BUILD_DIR/depends/$HOST/lib
     - if [ "$RUN_TESTS" = "true" ]; then travis_wait 30 make $MAKEJOBS check VERBOSE=1; fi
     - if [ "$TRAVIS_EVENT_TYPE" = "cron" ]; then extended="--extended --exclude pruning,dbcrash"; fi
-    - if [ "$RUN_TESTS" = "true" ]; then test/functional/test_runner.py --coverage --quiet ${extended}; fi
+    - if [ "$RUN_TESTS" = "true" ]; then test/functional/test_runner.py --combinedlogslen=4000 --coverage --quiet ${extended}; fi
 after_script:
     - echo $TRAVIS_COMMIT_RANGE
     - echo $TRAVIS_COMMIT_LOG
diff --git a/test/functional/test_framework/test_framework.py b/test/functional/test_framework/test_framework.py
index 4590b4c650..54fe689686 100755
--- a/test/functional/test_framework/test_framework.py
+++ b/test/functional/test_framework/test_framework.py
@@ -4,7 +4,6 @@
 # file COPYING or http://www.opensource.org/licenses/mit-license.php.
 """Base class for RPC testing."""
 
-from collections import deque
 from enum import Enum
 import logging
 import optparse
@@ -149,32 +148,19 @@ class BitcoinTestFramework():
             shutil.rmtree(self.options.tmpdir)
         else:
             self.log.warning("Not cleaning up dir %s" % self.options.tmpdir)
-            if os.getenv("PYTHON_DEBUG", ""):
-                # Dump the end of the debug logs, to aid in debugging rare
-                # travis failures.
-                import glob
-                filenames = [self.options.tmpdir + "/test_framework.log"]
-                filenames += glob.glob(self.options.tmpdir + "/node*/regtest/debug.log")
-                MAX_LINES_TO_PRINT = 1000
-                for fn in filenames:
-                    try:
-                        with open(fn, 'r') as f:
-                            print("From", fn, ":")
-                            print("".join(deque(f, MAX_LINES_TO_PRINT)))
-                    except OSError:
-                        print("Opening file %s failed." % fn)
-                        traceback.print_exc()
 
         if success == TestStatus.PASSED:
             self.log.info("Tests successful")
-            sys.exit(TEST_EXIT_PASSED)
+            exit_code = TEST_EXIT_PASSED
         elif success == TestStatus.SKIPPED:
             self.log.info("Test skipped")
-            sys.exit(TEST_EXIT_SKIPPED)
+            exit_code = TEST_EXIT_SKIPPED
         else:
             self.log.error("Test failed. Test logging available at %s/test_framework.log", self.options.tmpdir)
-            logging.shutdown()
-            sys.exit(TEST_EXIT_FAILED)
+            self.log.error("Hint: Call {} '{}' to consolidate all logs".format(os.path.normpath(os.path.dirname(os.path.realpath(__file__)) + "/../combine_logs.py"), self.options.tmpdir))
+            exit_code = TEST_EXIT_FAILED
+        logging.shutdown()
+        sys.exit(exit_code)
 
     # Methods to override in subclass test scripts.
     def set_test_params(self):
diff --git a/test/functional/test_runner.py b/test/functional/test_runner.py
index d953e1585c..8c5654a85d 100755
--- a/test/functional/test_runner.py
+++ b/test/functional/test_runner.py
@@ -15,6 +15,7 @@ For a description of arguments recognized by test scripts, see
 """
 
 import argparse
+from collections import deque
 import configparser
 import datetime
 import os
@@ -174,6 +175,7 @@ def main():
                                      epilog='''
     Help text and arguments for individual test script:''',
                                      formatter_class=argparse.RawTextHelpFormatter)
+    parser.add_argument('--combinedlogslen', '-c', type=int, default=0, help='print a combined log (of length n lines) from all test nodes and test framework to the console on failure.')
     parser.add_argument('--coverage', action='store_true', help='generate a basic coverage report for the RPC interface')
     parser.add_argument('--exclude', '-x', help='specify a comma-separated-list of scripts to exclude.')
     parser.add_argument('--extended', action='store_true', help='run the extended test suite in addition to the basic tests')
@@ -266,9 +268,9 @@ def main():
     if not args.keepcache:
         shutil.rmtree("%s/test/cache" % config["environment"]["BUILDDIR"], ignore_errors=True)
 
-    run_tests(test_list, config["environment"]["SRCDIR"], config["environment"]["BUILDDIR"], config["environment"]["EXEEXT"], tmpdir, args.jobs, args.coverage, passon_args)
+    run_tests(test_list, config["environment"]["SRCDIR"], config["environment"]["BUILDDIR"], config["environment"]["EXEEXT"], tmpdir, args.jobs, args.coverage, passon_args, args.combinedlogslen)
 
-def run_tests(test_list, src_dir, build_dir, exeext, tmpdir, jobs=1, enable_coverage=False, args=[]):
+def run_tests(test_list, src_dir, build_dir, exeext, tmpdir, jobs=1, enable_coverage=False, args=[], combined_logs_len=0):
     # Warn if bitcoind is already running (unix only)
     try:
         if subprocess.check_output(["pidof", "bitcoind"]) is not None:
@@ -314,7 +316,7 @@ def run_tests(test_list, src_dir, build_dir, exeext, tmpdir, jobs=1, enable_cove
     max_len_name = len(max(test_list, key=len))
 
     for _ in range(len(test_list)):
-        test_result, stdout, stderr = job_queue.get_next()
+        test_result, testdir, stdout, stderr = job_queue.get_next()
         test_results.append(test_result)
 
         if test_result.status == "Passed":
@@ -325,6 +327,14 @@ def run_tests(test_list, src_dir, build_dir, exeext, tmpdir, jobs=1, enable_cove
             print("\n%s%s%s failed, Duration: %s s\n" % (BOLD[1], test_result.name, BOLD[0], test_result.time))
             print(BOLD[1] + 'stdout:\n' + BOLD[0] + stdout + '\n')
             print(BOLD[1] + 'stderr:\n' + BOLD[0] + stderr + '\n')
+            if combined_logs_len and os.path.isdir(testdir):
+                # Print the final `combinedlogslen` lines of the combined logs
+                print('{}Combine the logs and print the last {} lines ...{}'.format(BOLD[1], combined_logs_len, BOLD[0]))
+                print('\n============')
+                print('{}Combined log for {}:{}'.format(BOLD[1], testdir, BOLD[0]))
+                print('============\n')
+                combined_logs, _ = subprocess.Popen([os.path.join(tests_dir, 'combine_logs.py'), '-c', testdir], universal_newlines=True, stdout=subprocess.PIPE).communicate()
+                print("\n".join(deque(combined_logs.splitlines(), combined_logs_len)))
 
     print_results(test_results, max_len_name, (int(time.time() - time0)))
 
@@ -389,13 +399,15 @@ class TestHandler:
             log_stdout = tempfile.SpooledTemporaryFile(max_size=2**16)
             log_stderr = tempfile.SpooledTemporaryFile(max_size=2**16)
             test_argv = t.split()
-            tmpdir = ["--tmpdir=%s/%s_%s" % (self.tmpdir, re.sub(".py$", "", test_argv[0]), portseed)]
+            testdir = "{}/{}_{}".format(self.tmpdir, re.sub(".py$", "", test_argv[0]), portseed)
+            tmpdir_arg = ["--tmpdir={}".format(testdir)]
             self.jobs.append((t,
                               time.time(),
-                              subprocess.Popen([self.tests_dir + test_argv[0]] + test_argv[1:] + self.flags + portseed_arg + tmpdir,
+                              subprocess.Popen([self.tests_dir + test_argv[0]] + test_argv[1:] + self.flags + portseed_arg + tmpdir_arg,
                                                universal_newlines=True,
                                                stdout=log_stdout,
                                                stderr=log_stderr),
+                              testdir,
                               log_stdout,
                               log_stderr))
         if not self.jobs:
@@ -404,7 +416,7 @@ class TestHandler:
             # Return first proc that finishes
             time.sleep(.5)
             for j in self.jobs:
-                (name, time0, proc, log_out, log_err) = j
+                (name, time0, proc, testdir, log_out, log_err) = j
                 if os.getenv('TRAVIS') == 'true' and int(time.time() - time0) > 20 * 60:
                     # In travis, timeout individual tests after 20 minutes (to stop tests hanging and not
                     # providing useful output.
@@ -422,7 +434,7 @@ class TestHandler:
                     self.num_running -= 1
                     self.jobs.remove(j)
 
-                    return TestResult(name, status, int(time.time() - time0)), stdout, stderr
+                    return TestResult(name, status, int(time.time() - time0)), testdir, stdout, stderr
             print('.', end='', flush=True)
 
 class TestResult():