#!/usr/bin/env python # smoke.py: run some mongo tests. # Bugs, TODOs: # 0 Some tests hard-code pathnames relative to the mongo repository, # so the smoke.py process and all its children must be run with the # mongo repo as current working directory. That's kinda icky. # 1 The tests that are implemented as standalone executables ("test"), # don't take arguments for the dbpath, but unconditionally use # "/tmp/unittest". # 2 mongod output gets intermingled with mongo output, and it's often # hard to find error messages in the slop. Maybe have smoke.py do # some fancier wrangling of child process output? # 3 Some test suites run their own mongods, and so don't need us to # run any mongods around their execution. (It's harmless to do so, # but adds noise in the output.) # 4 Running a separate mongo shell for each js file is slower than # loading js files into one mongo shell process. Maybe have runTest # queue up all filenames ending in ".js" and run them in one mongo # shell at the "end" of testing? # 5 Right now small-oplog implies master/slave replication. Maybe # running with replication should be an orthogonal concern. (And # maybe test replica set replication, too.) # 6 We use cleanbb.py to clear out the dbpath, but cleanbb.py kills # off all mongods on a box, which means you can't run two smoke.py # jobs on the same host at once. So something's gotta change. from datetime import datetime from itertools import izip import glob import logging from optparse import OptionParser import os import pprint import re import shlex import signal import socket import stat from subprocess import (PIPE, Popen, STDOUT) import sys import time import threading import traceback from pymongo import MongoClient from pymongo.errors import OperationFailure from pymongo import ReadPreference import cleanbb import utils try: import cPickle as pickle except ImportError: import pickle try: from hashlib import md5 # new in 2.5 except ImportError: from md5 import md5 # deprecated in 2.5 try: import json except: try: import simplejson as json except: json = None # Get relative imports to work when the package is not installed on the PYTHONPATH. if __name__ == "__main__" and __package__ is None: sys.path.append(os.path.dirname(os.path.dirname(os.path.abspath(os.path.realpath(__file__))))) from buildscripts.resmokelib.core import pipe # TODO clean this up so we don't need globals... mongo_repo = os.getcwd() #'./' failfile = os.path.join(mongo_repo, 'failfile.smoke') test_path = None mongod_executable = None mongod_port = None shell_executable = None continue_on_failure = None file_of_commands_mode = False start_mongod = True temp_path = None clean_every_n_tests = 1 clean_whole_dbroot = False tests = [] winners = [] losers = {} fails = [] # like losers but in format of tests # For replication hash checking replicated_collections = [] lost_in_slave = [] lost_in_master = [] screwy_in_slave = {} smoke_db_prefix = '' small_oplog = False small_oplog_rs = False test_report = { "results": [] } report_file = None # This class just implements the with statement API class NullMongod(object): def start(self): pass def stop(self): pass def __enter__(self): self.start() return self def __exit__(self, type, value, traceback): self.stop() return not isinstance(value, Exception) def dump_stacks(signal, frame): print "======================================" print "DUMPING STACKS due to SIGUSR1 signal" print "======================================" threads = threading.enumerate(); print "Total Threads: " + str(len(threads)) for id, stack in sys._current_frames().items(): print "Thread %d" % (id) print "".join(traceback.format_stack(stack)) print "======================================" def buildlogger(cmd, is_global=False): # if the environment variable MONGO_USE_BUILDLOGGER # is set to 'true', then wrap the command with a call # to buildlogger.py, which sends output to the buidlogger # machine; otherwise, return as usual. if os.environ.get('MONGO_USE_BUILDLOGGER', '').lower().strip() == 'true': if is_global: return [utils.find_python(), 'buildscripts/buildlogger.py', '-g'] + cmd else: return [utils.find_python(), 'buildscripts/buildlogger.py'] + cmd return cmd def clean_dbroot(dbroot="", nokill=False): # Clean entire /data/db dir if --with-cleanbb, else clean specific database path. if clean_whole_dbroot and not (small_oplog or small_oplog_rs): dbroot = os.path.normpath(smoke_db_prefix + "/data/db") if os.path.exists(dbroot): print("clean_dbroot: %s" % dbroot) cleanbb.cleanup(dbroot, nokill) class mongod(NullMongod): def __init__(self, **kwargs): self.kwargs = kwargs self.proc = None self.auth = False self.job_object = None self._inner_proc_pid = None self._stdout_pipe = None def ensure_test_dirs(self): utils.ensureDir(smoke_db_prefix + "/tmp/unittest/") utils.ensureDir(smoke_db_prefix + "/data/") utils.ensureDir(smoke_db_prefix + "/data/db/") def check_mongo_port(self, port=27017): sock = socket.socket() sock.setsockopt(socket.IPPROTO_TCP, socket.TCP_NODELAY, 1) sock.settimeout(1) sock.connect(("localhost", int(port))) sock.close() def is_mongod_up(self, port=mongod_port): if not start_mongod: return False try: self.check_mongo_port(int(port)) return True except Exception,e: print >> sys.stderr, e return False def did_mongod_start(self, port=mongod_port, timeout=300): while timeout > 0: time.sleep(1) is_up = self.is_mongod_up(port) if is_up: return True timeout = timeout - 1 print >> sys.stderr, "timeout starting mongod" return False def start(self): global mongod_port global mongod if self.proc: print >> sys.stderr, "probable bug: self.proc already set in start()" return self.ensure_test_dirs() dir_name = smoke_db_prefix + "/data/db/sconsTests/" self.port = int(mongod_port) self.slave = False if 'slave' in self.kwargs: dir_name = smoke_db_prefix + '/data/db/sconsTestsSlave/' srcport = mongod_port self.port += 1 self.slave = True clean_dbroot(dbroot=dir_name, nokill=self.slave) utils.ensureDir(dir_name) argv = [mongod_executable, "--port", str(self.port), "--dbpath", dir_name] # These parameters are always set for tests # SERVER-9137 Added httpinterface parameter to keep previous behavior argv += ['--setParameter', 'enableTestCommands=1', '--httpinterface'] if self.kwargs.get('small_oplog'): if self.slave: argv += ['--slave', '--source', 'localhost:' + str(srcport)] else: argv += ["--master", "--oplogSize", "511"] if self.kwargs.get('storage_engine'): argv += ["--storageEngine", self.kwargs.get('storage_engine')] if self.kwargs.get('wiredtiger_engine_config_string'): argv += ["--wiredTigerEngineConfigString", self.kwargs.get('wiredtiger_engine_config_string')] if self.kwargs.get('wiredtiger_collection_config_string'): argv += ["--wiredTigerCollectionConfigString", self.kwargs.get('wiredtiger_collection_config_string')] if self.kwargs.get('wiredtiger_index_config_string'): argv += ["--wiredTigerIndexConfigString", self.kwargs.get('wiredtiger_index_config_string')] params = self.kwargs.get('set_parameters', None) if params: for p in params.split(','): argv += ['--setParameter', p] if self.kwargs.get('small_oplog_rs'): argv += ["--replSet", "foo", "--oplogSize", "511"] if self.kwargs.get('no_journal'): argv += ['--nojournal'] if self.kwargs.get('no_preallocj'): argv += ['--nopreallocj'] if self.kwargs.get('auth'): argv += ['--auth', '--setParameter', 'enableLocalhostAuthBypass=false'] authMechanism = self.kwargs.get('authMechanism', 'SCRAM-SHA-1') if authMechanism != 'SCRAM-SHA-1': argv += ['--setParameter', 'authenticationMechanisms=' + authMechanism] self.auth = True if self.kwargs.get('keyFile'): argv += ['--keyFile', self.kwargs.get('keyFile')] if self.kwargs.get('use_ssl'): argv += ['--sslMode', "requireSSL", '--sslPEMKeyFile', 'jstests/libs/server.pem', '--sslCAFile', 'jstests/libs/ca.pem', '--sslAllowConnectionsWithoutCertificates'] if self.kwargs.get('rlp_path'): argv += ['--basisTechRootDirectory', self.kwargs.get('rlp_path')] print "running " + " ".join(argv) self.proc = self._start(buildlogger(argv, is_global=True)) # If the mongod process is spawned under buildlogger.py, then the first line of output # should include the pid of the underlying mongod process. If smoke.py didn't create its own # job object because it is already inside one, then the pid is used to attempt to terminate # the underlying mongod process. first_line = self.proc.stdout.readline() match = re.search("^\[buildlogger.py\] pid: (?P[0-9]+)$", first_line.rstrip()) if match is not None: self._inner_proc_pid = int(match.group("pid")) else: # The first line of output didn't include the pid of the underlying mongod process. We # write the first line of output to smoke.py's stdout to ensure the message doesn't get # lost since it's possible that buildlogger.py isn't being used. sys.stdout.write(first_line) logger = logging.Logger("", level=logging.DEBUG) handler = logging.StreamHandler(sys.stdout) handler.setFormatter(logging.Formatter(fmt="%(message)s")) logger.addHandler(handler) self._stdout_pipe = pipe.LoggerPipe(logger, logging.INFO, self.proc.stdout) self._stdout_pipe.wait_until_started() if not self.did_mongod_start(self.port): raise Exception("Failed to start mongod") if self.slave: local = MongoClient(port=self.port, read_preference=ReadPreference.SECONDARY_PREFERRED).local synced = False while not synced: synced = True for source in local.sources.find({}, ["syncedTo"]): synced = synced and "syncedTo" in source and source["syncedTo"] def _start(self, argv): """In most cases, just call subprocess.Popen(). On Windows, this method also assigns the started process to a job object if a new one was created. This ensures that any child processes of this process can be killed with a single call to TerminateJobObject (see self.stop()). """ creation_flags = 0 if os.sys.platform == "win32": # Create a job object with the "kill on job close" # flag; this is inherited by child processes (ie # the mongod started on our behalf by buildlogger) # and lets us terminate the whole tree of processes # rather than orphaning the mongod. import win32job import win32process # Don't create a job object if the current process is already inside one. if not win32job.IsProcessInJob(win32process.GetCurrentProcess(), None): self.job_object = win32job.CreateJobObject(None, '') job_info = win32job.QueryInformationJobObject( self.job_object, win32job.JobObjectExtendedLimitInformation) job_info['BasicLimitInformation']['LimitFlags'] |= \ win32job.JOB_OBJECT_LIMIT_KILL_ON_JOB_CLOSE win32job.SetInformationJobObject( self.job_object, win32job.JobObjectExtendedLimitInformation, job_info) # Magic number needed to allow job reassignment in Windows 7 # see: MSDN - Process Creation Flags - ms684863 creation_flags |= win32process.CREATE_BREAKAWAY_FROM_JOB proc = Popen(argv, creationflags=creation_flags, stdout=PIPE, stderr=None, bufsize=0) if self.job_object is not None: win32job.AssignProcessToJobObject(self.job_object, proc._handle) return proc def stop(self): if not self.proc: print >> sys.stderr, "probable bug: self.proc unset in stop()" return try: if os.sys.platform == "win32" and self.job_object is not None: # If smoke.py created its own job object, then we clean up the spawned processes by # terminating it. import win32job win32job.TerminateJobObject(self.job_object, -1) # Windows doesn't seem to kill the process immediately, so give it some time to die time.sleep(5) elif os.sys.platform == "win32": # If smoke.py didn't create its own job object, then we attempt to clean up the # spawned processes by terminating them individually. import win32api import win32con import win32event import win32process import winerror def win32_terminate(handle): # Adapted from implementation of Popen.terminate() in subprocess.py of Python # 2.7 because earlier versions do not catch exceptions. try: win32process.TerminateProcess(handle, -1) except win32process.error as err: # ERROR_ACCESS_DENIED (winerror=5) is received when the process has # already died. if err.winerror != winerror.ERROR_ACCESS_DENIED: raise return_code = win32process.GetExitCodeProcess(handle) if return_code == win32con.STILL_ACTIVE: raise # Terminate the mongod process underlying buildlogger.py if one exists. if self._inner_proc_pid is not None: # The PROCESS_TERMINATE privilege is necessary to call TerminateProcess() and # the SYNCHRONIZE privilege is necessary to call WaitForSingleObject(). See # https://msdn.microsoft.com/en-us/library/windows/desktop/ms684880(v=vs.85).aspx # for more details. required_access = win32con.PROCESS_TERMINATE | win32con.SYNCHRONIZE inner_proc_handle = win32api.OpenProcess(required_access, False, self._inner_proc_pid) try: win32_terminate(inner_proc_handle) win32event.WaitForSingleObject(inner_proc_handle, win32event.INFINITE) finally: win32api.CloseHandle(inner_proc_handle) win32_terminate(self.proc._handle) elif hasattr(self.proc, "terminate"): # This method added in Python 2.6 self.proc.terminate() else: os.kill(self.proc.pid, 15) except Exception, e: print >> sys.stderr, "error shutting down mongod" print >> sys.stderr, e self.proc.wait() if self._stdout_pipe is not None: self._stdout_pipe.wait_until_finished() sys.stderr.flush() sys.stdout.flush() # Fail hard if mongod terminates with an error. That might indicate that an # instrumented build (e.g. LSAN) has detected an error. For now we aren't doing this on # windows because the exit code seems to be unpredictable. We don't have LSAN there # anyway. retcode = self.proc.returncode if os.sys.platform != "win32" and retcode != 0: raise(Exception('mongod process exited with non-zero code %d' % retcode)) def wait_for_repl(self): print "Awaiting replicated (w:2, wtimeout:5min) insert (port:" + str(self.port) + ")" MongoClient(port=self.port).testing.smokeWait.insert({}, w=2, wtimeout=5*60*1000) print "Replicated write completed -- done wait_for_repl" class Bug(Exception): def __str__(self): return 'bug in smoke.py: ' + super(Bug, self).__str__() class TestFailure(Exception): pass class TestExitFailure(TestFailure): def __init__(self, *args): self.path = args[0] self.status=args[1] def __str__(self): return "test %s exited with status %d" % (self.path, self.status) class TestServerFailure(TestFailure): def __init__(self, *args): self.path = args[0] self.status = -1 # this is meaningless as an exit code, but # that's the point. def __str__(self): return 'mongod not running after executing test %s' % self.path def check_db_hashes(master, slave): # Need to pause a bit so a slave might catch up... if not slave.slave: raise(Bug("slave instance doesn't have slave attribute set")) master.wait_for_repl() # FIXME: maybe make this run dbhash on all databases? for mongod in [master, slave]: client = MongoClient(port=mongod.port, read_preference=ReadPreference.SECONDARY_PREFERRED) mongod.dbhash = client.test.command("dbhash") mongod.dict = mongod.dbhash["collections"] global lost_in_slave, lost_in_master, screwy_in_slave, replicated_collections replicated_collections += master.dict.keys() for coll in replicated_collections: if coll not in slave.dict and coll not in lost_in_slave: lost_in_slave.append(coll) mhash = master.dict[coll] shash = slave.dict[coll] if mhash != shash: mTestDB = MongoClient(port=master.port).test sTestDB = MongoClient(port=slave.port, read_preference=ReadPreference.SECONDARY_PREFERRED).test mCount = mTestDB[coll].count() sCount = sTestDB[coll].count() stats = {'hashes': {'master': mhash, 'slave': shash}, 'counts':{'master': mCount, 'slave': sCount}} try: mDocs = list(mTestDB[coll].find().sort("_id", 1)) sDocs = list(sTestDB[coll].find().sort("_id", 1)) mDiffDocs = list() sDiffDocs = list() for left, right in izip(mDocs, sDocs): if left != right: mDiffDocs.append(left) sDiffDocs.append(right) stats["docs"] = {'master': mDiffDocs, 'slave': sDiffDocs } except Exception, e: stats["error-docs"] = e; screwy_in_slave[coll] = stats if mhash == "no _id_ index": oplog = "oplog.$main" if small_oplog_rs: oplog = "oplog.rs" mOplog = mTestDB.connection.local[oplog]; oplog_entries = list(mOplog.find({"$or": [{"ns":mTestDB[coll].full_name}, \ {"op":"c"}]}).sort("$natural", 1)) print "oplog for %s" % mTestDB[coll].full_name for doc in oplog_entries: pprint.pprint(doc, width=200) for db in slave.dict.keys(): if db not in master.dict and db not in lost_in_master: lost_in_master.append(db) def ternary( b , l="true", r="false" ): if b: return l return r # Blech. def skipTest(path): basename = os.path.basename(path) parentPath = os.path.dirname(path) parentDir = os.path.basename(parentPath) if small_oplog or small_oplog_rs: # For tests running in parallel if basename in ["cursor8.js", "indexh.js", "dropdb.js", "dropdb_race.js", "connections_opened.js", "opcounters_write_cmd.js", "dbadmin.js", # Should not run in repl mode: "read_after_optime.js", ## Capped tests "capped_max1.js", "capped_convertToCapped1.js", "rename.js"]: return True if auth or keyFile: # For tests running with auth # Skip any tests that run with auth explicitly if parentDir.lower() == "auth" or "auth" in basename.lower(): return True if parentPath == mongo_repo: # Skip client tests return True if parentDir == "tool": # SERVER-6368 return True if parentDir == "dur": # SERVER-7317 return True if parentDir == "disk": # SERVER-7356 return True authTestsToSkip = [("jstests", "drop2.js"), # SERVER-8589, ("jstests", "killop.js"), # SERVER-10128 ("sharding", "sync3.js"), # SERVER-6388 for this and those below ("sharding", "parallel.js"), ("sharding", "copydb_from_mongos.js"), # SERVER-13080 ("jstests", "bench_test1.js"), ("jstests", "bench_test2.js"), ("jstests", "bench_test3.js"), ("core", "bench_test1.js"), ("core", "bench_test2.js"), ("core", "bench_test3.js"), ] if os.path.join(parentDir,basename) in [ os.path.join(*test) for test in authTestsToSkip ]: return True return False legacyWriteRE = re.compile(r"jstests[/\\]multiVersion") def setShellWriteModeForTest(path, argv): swm = shell_write_mode if legacyWriteRE.search(path): swm = "legacy" argv += ["--writeMode", swm] def runTest(test, result): # result is a map containing test result details, like result["url"] # test is a tuple of ( filename , usedb ) # filename should be a js file to run # usedb is true if the test expects a mongod to be running (path, usedb) = test (ignore, ext) = os.path.splitext(path) test_mongod = mongod() mongod_is_up = test_mongod.is_mongod_up(mongod_port) result["mongod_running_at_start"] = mongod_is_up; if file_of_commands_mode: # smoke.py was invoked like "--mode files --from-file foo", # so don't try to interpret the test path too much if os.sys.platform == "win32": argv = [path] else: argv = shlex.split(path) path = argv[0] # if the command is a python script, use the script name if os.path.basename(path) in ('python', 'python.exe'): path = argv[1] elif ext == ".js": argv = [shell_executable, "--port", mongod_port] setShellWriteModeForTest(path, argv) if not usedb: argv += ["--nodb"] if small_oplog or small_oplog_rs: argv += ["--eval", 'testingReplication = true;'] if use_ssl: argv += ["--ssl", "--sslPEMKeyFile", "jstests/libs/client.pem", "--sslCAFile", "jstests/libs/ca.pem", "--sslAllowInvalidCertificates"] argv += [path] elif ext in ["", ".exe"]: # Blech. if os.path.basename(path) in ["dbtest", "dbtest.exe"]: argv = [path] # default data directory for dbtest is /tmp/unittest if smoke_db_prefix: dir_name = smoke_db_prefix + '/unittests' argv.extend(["--dbpath", dir_name] ) if storage_engine: argv.extend(["--storageEngine", storage_engine]) if wiredtiger_engine_config_string: argv.extend(["--wiredTigerEngineConfigString", wiredtiger_engine_config_string]) if wiredtiger_collection_config_string: argv.extend(["--wiredTigerCollectionConfigString", wiredtiger_collection_config_string]) if wiredtiger_index_config_string: argv.extend(["--wiredTigerIndexConfigString", wiredtiger_index_config_string]) # more blech elif os.path.basename(path) in ['mongos', 'mongos.exe']: argv = [path, "--test"] else: argv = [test_path and os.path.abspath(os.path.join(test_path, path)) or path, "--port", mongod_port] else: raise Bug("fell off in extension case: %s" % path) mongo_test_filename = os.path.basename(path) # sys.stdout.write() is more atomic than print, so using it prevents # lines being interrupted by, e.g., child processes sys.stdout.write(" *******************************************\n") sys.stdout.write(" Test : %s ...\n" % mongo_test_filename) sys.stdout.flush() # FIXME: we don't handle the case where the subprocess # hangs... that's bad. if ( argv[0].endswith( 'mongo' ) or argv[0].endswith( 'mongo.exe' ) ) and not '--eval' in argv : evalString = 'TestData = new Object();' + \ 'TestData.storageEngine = "' + ternary( storage_engine, storage_engine, "" ) + '";' + \ 'TestData.wiredTigerEngineConfigString = "' + ternary( wiredtiger_engine_config_string, wiredtiger_engine_config_string, "" ) + '";' + \ 'TestData.wiredTigerCollectionConfigString = "' + ternary( wiredtiger_collection_config_string, wiredtiger_collection_config_string, "" ) + '";' + \ 'TestData.wiredTigerIndexConfigString = "' + ternary( wiredtiger_index_config_string, wiredtiger_index_config_string, "" ) + '";' + \ 'TestData.testName = "' + re.sub( ".js$", "", os.path.basename( path ) ) + '";' + \ 'TestData.setParameters = "' + ternary( set_parameters, set_parameters, "" ) + '";' + \ 'TestData.setParametersMongos = "' + ternary( set_parameters_mongos, set_parameters_mongos, "" ) + '";' + \ 'TestData.noJournal = ' + ternary( no_journal ) + ";" + \ 'TestData.noJournalPrealloc = ' + ternary( no_preallocj ) + ";" + \ 'TestData.auth = ' + ternary( auth ) + ";" + \ 'TestData.keyFile = ' + ternary( keyFile , '"' + str(keyFile) + '"' , 'null' ) + ";" + \ 'TestData.keyFileData = ' + ternary( keyFile , '"' + str(keyFileData) + '"' , 'null' ) + ";" + \ 'TestData.authMechanism = ' + ternary( authMechanism, '"' + str(authMechanism) + '"', 'null') + ";" # this updates the default data directory for mongod processes started through shell (src/mongo/shell/servers.js) evalString += 'MongoRunner.dataDir = "' + os.path.abspath(smoke_db_prefix + '/data/db') + '";' evalString += 'MongoRunner.dataPath = MongoRunner.dataDir + "/";' if temp_path: evalString += 'TestData.tmpPath = "' + temp_path + '";' if os.sys.platform == "win32": # double quotes in the evalString on windows; this # prevents the backslashes from being removed when # the shell (i.e. bash) evaluates this string. yuck. evalString = evalString.replace('\\', '\\\\') if auth and usedb: evalString += 'jsTest.authenticate(db.getMongo());' argv = argv + [ '--eval', evalString] if argv[0].endswith( 'dbtest' ) or argv[0].endswith( 'dbtest.exe' ): if no_preallocj : argv = argv + [ '--nopreallocj' ] if temp_path: argv = argv + [ '--tempPath', temp_path ] sys.stdout.write(" Command : %s\n" % ' '.join(argv)) sys.stdout.write(" Date : %s\n" % datetime.now().ctime()) sys.stdout.flush() os.environ['MONGO_TEST_FILENAME'] = mongo_test_filename t1 = time.time() proc = Popen(buildlogger(argv), cwd=test_path, stdout=PIPE, stderr=STDOUT, bufsize=0) first_line = proc.stdout.readline() # Get suppressed output URL m = re.search(r"\s*\(output suppressed; see (?P.*)\)" + os.linesep, first_line) if m: result["url"] = m.group("url") sys.stdout.write(first_line) sys.stdout.flush() while True: # print until subprocess's stdout closed. # Not using "for line in file" since that has unwanted buffering. line = proc.stdout.readline() if not line: break; sys.stdout.write(line) sys.stdout.flush() proc.wait() # wait if stdout is closed before subprocess exits. r = proc.returncode t2 = time.time() del os.environ['MONGO_TEST_FILENAME'] timediff = t2 - t1 # timediff is seconds by default scale = 1 suffix = "seconds" # if timediff is less than 10 seconds use ms if timediff < 10: scale = 1000 suffix = "ms" # if timediff is more than 60 seconds use minutes elif timediff > 60: scale = 1.0 / 60.0 suffix = "minutes" sys.stdout.write(" %10.4f %s\n" % ((timediff) * scale, suffix)) sys.stdout.flush() result["exit_code"] = r is_mongod_still_up = test_mongod.is_mongod_up(mongod_port) if start_mongod and not is_mongod_still_up: print "mongod is not running after test" result["mongod_running_at_end"] = is_mongod_still_up; raise TestServerFailure(path) result["mongod_running_at_end"] = is_mongod_still_up; if r != 0: raise TestExitFailure(path, r) print "" def run_tests(tests): # FIXME: some suites of tests start their own mongod, so don't # need this. (So long as there are no conflicts with port, # dbpath, etc., and so long as we shut ours down properly, # starting this mongod shouldn't break anything, though.) # The reason we want to use "with" is so that we get __exit__ semantics # but "with" is only supported on Python 2.5+ master = NullMongod() slave = NullMongod() try: if start_mongod: master = mongod(small_oplog_rs=small_oplog_rs, small_oplog=small_oplog, no_journal=no_journal, storage_engine=storage_engine, wiredtiger_engine_config_string=wiredtiger_engine_config_string, wiredtiger_collection_config_string=wiredtiger_collection_config_string, wiredtiger_index_config_string=wiredtiger_index_config_string, set_parameters=set_parameters, no_preallocj=no_preallocj, auth=auth, authMechanism=authMechanism, keyFile=keyFile, rlp_path=rlp_path, use_ssl=use_ssl) master.start() if small_oplog: slave = mongod(slave=True, small_oplog=True, small_oplog_rs=False, storage_engine=storage_engine, wiredtiger_engine_config_string=wiredtiger_engine_config_string, wiredtiger_collection_config_string=wiredtiger_collection_config_string, wiredtiger_index_config_string=wiredtiger_index_config_string, set_parameters=set_parameters) slave.start() elif small_oplog_rs: slave = mongod(slave=True, small_oplog_rs=True, small_oplog=False, no_journal=no_journal, storage_engine=storage_engine, wiredtiger_engine_config_string=wiredtiger_engine_config_string, wiredtiger_collection_config_string=wiredtiger_collection_config_string, wiredtiger_index_config_string=wiredtiger_index_config_string, set_parameters=set_parameters, no_preallocj=no_preallocj, auth=auth, authMechanism=authMechanism, keyFile=keyFile, rlp_path=rlp_path, use_ssl=use_ssl) slave.start() primary = MongoClient(port=master.port); primary.admin.command({'replSetInitiate' : {'_id' : 'foo', 'members' : [ {'_id': 0, 'host':'localhost:%s' % master.port}, {'_id': 1, 'host':'localhost:%s' % slave.port,'priority':0}]}}) # Wait for primary and secondary to finish initial sync and election ismaster = False while not ismaster: result = primary.admin.command("ismaster"); ismaster = result["ismaster"] if not ismaster: print "waiting for primary to be available ..." time.sleep(.2) secondaryUp = False sConn = MongoClient(port=slave.port, read_preference=ReadPreference.SECONDARY_PREFERRED); while not secondaryUp: result = sConn.admin.command("ismaster"); secondaryUp = result["secondary"] if not secondaryUp: print "waiting for secondary to be available ..." time.sleep(.2) if small_oplog or small_oplog_rs: master.wait_for_repl() for tests_run, test in enumerate(tests): tests_run += 1 # enumerate from 1, python 2.5 compatible test_result = { "start": time.time() } (test_path, use_db) = test if test_path.startswith(mongo_repo + os.path.sep): test_result["test_file"] = test_path[len(mongo_repo)+1:] else: # user could specify a file not in repo. leave it alone. test_result["test_file"] = test_path try: if skipTest(test_path): test_result["status"] = "skip" print "skipping " + test_path else: fails.append(test) runTest(test, test_result) fails.pop() winners.append(test) test_result["status"] = "pass" test_result["end"] = time.time() test_result["elapsed"] = test_result["end"] - test_result["start"] test_report["results"].append( test_result ) if small_oplog or small_oplog_rs: master.wait_for_repl() # check the db_hashes if isinstance(slave, mongod): check_db_hashes(master, slave) check_and_report_replication_dbhashes() elif use_db: # reach inside test and see if "usedb" is true if clean_every_n_tests and (tests_run % clean_every_n_tests) == 0: # Restart mongod periodically to clean accumulated test data # clean_dbroot() is invoked by mongod.start() master.stop() master = mongod(small_oplog_rs=small_oplog_rs, small_oplog=small_oplog, no_journal=no_journal, storage_engine=storage_engine, wiredtiger_engine_config_string=wiredtiger_engine_config_string, wiredtiger_collection_config_string=wiredtiger_collection_config_string, wiredtiger_index_config_string=wiredtiger_index_config_string, set_parameters=set_parameters, no_preallocj=no_preallocj, auth=auth, authMechanism=authMechanism, keyFile=keyFile, rlp_path=rlp_path, use_ssl=use_ssl) master.start() except TestFailure, f: test_result["end"] = time.time() test_result["elapsed"] = test_result["end"] - test_result["start"] test_result["error"] = str(f) test_result["status"] = "fail" test_report["results"].append( test_result ) try: print f # Record the failing test and re-raise. losers[f.path] = f.status raise f except TestServerFailure, f: return 2 except TestFailure, f: if not continue_on_failure: return 1 if isinstance(slave, mongod): check_db_hashes(master, slave) finally: slave.stop() master.stop() return 0 def check_and_report_replication_dbhashes(): def missing(lst, src, dst): if lst: print """The following collections were present in the %s but not the %s at the end of testing:""" % (src, dst) for db in lst: print db missing(lost_in_slave, "master", "slave") missing(lost_in_master, "slave", "master") if screwy_in_slave: print """The following collections have different hashes in the master and slave:""" for coll in screwy_in_slave.keys(): stats = screwy_in_slave[coll] # Counts are "approx" because they are collected after the dbhash runs and may not # reflect the states of the collections that were hashed. If the hashes differ, one # possibility is that a test exited with writes still in-flight. print "collection: %s\t (master/slave) hashes: %s/%s counts (approx): %i/%i" % (coll, stats['hashes']['master'], stats['hashes']['slave'], stats['counts']['master'], stats['counts']['slave']) if "docs" in stats: if (("master" in stats["docs"] and len(stats["docs"]["master"]) == 0) and ("slave" in stats["docs"] and len(stats["docs"]["slave"]) == 0)): print "All docs matched!" else: print "Different Docs" print "Master docs:" pprint.pprint(stats["docs"]["master"], indent=2) print "Slave docs:" pprint.pprint(stats["docs"]["slave"], indent=2) if "error-docs" in stats: print "Error getting docs to diff:" pprint.pprint(stats["error-docs"]) return True if (small_oplog or small_oplog_rs) and not (lost_in_master or lost_in_slave or screwy_in_slave): print "replication ok for %d collections" % (len(replicated_collections)) return False def report(): print "%d tests succeeded" % len(winners) num_missed = len(tests) - (len(winners) + len(losers.keys())) if num_missed: print "%d tests didn't get run" % num_missed if losers: print "The following tests failed (with exit code):" for loser in losers: print "%s\t%d" % (loser, losers[loser]) test_result = { "start": time.time() } if check_and_report_replication_dbhashes(): test_result["end"] = time.time() test_result["elapsed"] = test_result["end"] - test_result["start"] test_result["test_file"] = "/#dbhash#" test_result["error"] = "dbhash mismatch" test_result["status"] = "fail" test_report["results"].append( test_result ) if report_file: f = open( report_file, "wb" ) f.write( json.dumps( test_report ) ) f.close() if losers or lost_in_slave or lost_in_master or screwy_in_slave: raise Exception("Test failures") # Keys are the suite names (passed on the command line to smoke.py) # Values are pairs: (filenames, ) suiteGlobalConfig = {"js": ("core/*.js", True), "quota": ("quota/*.js", True), "jsPerf": ("perf/*.js", True), "disk": ("disk/*.js", True), "noPassthroughWithMongod": ("noPassthroughWithMongod/*.js", True), "noPassthrough": ("noPassthrough/*.js", False), "parallel": ("parallel/*.js", True), "concurrency": ("concurrency/*.js", True), "clone": ("clone/*.js", False), "repl": ("repl/*.js", False), "replSets": ("replsets/*.js", False), "dur": ("dur/*.js", False), "auth": ("auth/*.js", False), "sharding": ("sharding/*.js", False), "tool": ("tool/*.js", False), "aggregation": ("aggregation/*.js", True), "multiVersion": ("multiVersion/*.js", True), "failPoint": ("fail_point/*.js", False), "ssl": ("ssl/*.js", True), "sslSpecial": ("sslSpecial/*.js", True), "jsCore": ("core/*.js", True), "mmap_v1": ("mmap_v1/*.js", True), "gle": ("gle/*.js", True), "rocksDB": ("rocksDB/*.js", True), "slow1": ("slow1/*.js", True), "serial_run": ("serial_run/*.js", True), } def get_module_suites(): """Attempts to discover and return information about module test suites Returns a dictionary of module suites in the format: { "" : "", ... } This means the values of this dictionary can be used as "glob"s to match all jstests in the suite directory that don't start with an underscore The module tests should be put in 'src/mongo/db/modules///*.js' NOTE: This assumes that if we have more than one module the suite names don't conflict """ modules_directory = 'src/mongo/db/modules' test_suites = {} # Return no suites if we have no modules if not os.path.exists(modules_directory) or not os.path.isdir(modules_directory): return {} module_directories = os.listdir(modules_directory) for module_directory in module_directories: test_directory = os.path.join(modules_directory, module_directory, "jstests") # Skip this module if it has no "jstests" directory if not os.path.exists(test_directory) or not os.path.isdir(test_directory): continue # Get all suites for this module for test_suite in os.listdir(test_directory): test_suites[test_suite] = os.path.join(test_directory, test_suite, "[!_]*.js") return test_suites def expand_suites(suites,expandUseDB=True): """Takes a list of suites and expands to a list of tests according to a set of rules. Keyword arguments: suites -- list of suites specified by the user expandUseDB -- expand globs (such as [!_]*.js) for tests that are run against a database (default True) This function handles expansion of globs (such as [!_]*.js), aliases (such as "client" and "all"), detection of suites in the "modules" directory, and enumerating the test files in a given suite. It returns a list of tests of the form (path_to_test, usedb), where the second part of the tuple specifies whether the test is run against the database (see --nodb in the mongo shell) """ globstr = None tests = [] module_suites = get_module_suites() for suite in suites: if suite == 'all': return expand_suites(['dbtest', 'jsCore', 'jsPerf', 'mmap_v1', 'noPassthroughWithMongod', 'noPassthrough', 'clone', 'parallel', 'concurrency', 'repl', 'auth', 'sharding', 'slow1', 'serial_run', 'tool'], expandUseDB=expandUseDB) if suite == 'dbtest' or suite == 'test': if os.sys.platform == "win32": program = 'dbtest.exe' else: program = 'dbtest' (globstr, usedb) = (program, False) elif suite == 'mongosTest': if os.sys.platform == "win32": program = 'mongos.exe' else: program = 'mongos' tests += [(os.path.join(mongo_repo, program), False)] elif os.path.exists( suite ): usedb = True for name in suiteGlobalConfig: if suite in glob.glob( "jstests/" + suiteGlobalConfig[name][0] ): usedb = suiteGlobalConfig[name][1] break tests += [ ( os.path.join( mongo_repo , suite ) , usedb ) ] elif suite in module_suites: # Currently we connect to a database in all module tests since there's no mechanism yet # to configure it independently usedb = True paths = glob.glob(module_suites[suite]) paths.sort() tests += [(path, usedb) for path in paths] else: try: globstr, usedb = suiteGlobalConfig[suite] except KeyError: raise Exception('unknown test suite %s' % suite) if globstr: if usedb and not expandUseDB: tests += [ (suite,False) ] else: if globstr.endswith('.js'): loc = 'jstests/' else: loc = '' globstr = os.path.join(mongo_repo, (os.path.join(loc, globstr))) globstr = os.path.normpath(globstr) paths = glob.glob(globstr) paths.sort() tests += [(path, usedb) for path in paths] return tests def add_exe(e): if os.sys.platform.startswith( "win" ) and not e.endswith( ".exe" ): e += ".exe" return e def set_globals(options, tests): global mongod_executable, mongod_port, shell_executable, continue_on_failure global small_oplog, small_oplog_rs global no_journal, set_parameters, set_parameters_mongos, no_preallocj, storage_engine, wiredtiger_engine_config_string, wiredtiger_collection_config_string, wiredtiger_index_config_string global auth, authMechanism, keyFile, keyFileData, smoke_db_prefix, test_path, start_mongod global rlp_path global use_ssl global file_of_commands_mode global report_file, shell_write_mode, use_write_commands global temp_path global clean_every_n_tests global clean_whole_dbroot start_mongod = options.start_mongod if hasattr(options, 'use_ssl'): use_ssl = options.use_ssl #Careful, this can be called multiple times test_path = options.test_path mongod_executable = add_exe(options.mongod_executable) if not os.path.exists(mongod_executable): raise Exception("no mongod found in this directory.") mongod_port = options.mongod_port shell_executable = add_exe( options.shell_executable ) if not os.path.exists(shell_executable): raise Exception("no mongo shell found in this directory.") continue_on_failure = options.continue_on_failure smoke_db_prefix = options.smoke_db_prefix small_oplog = options.small_oplog if hasattr(options, "small_oplog_rs"): small_oplog_rs = options.small_oplog_rs no_journal = options.no_journal storage_engine = options.storage_engine wiredtiger_engine_config_string = options.wiredtiger_engine_config_string wiredtiger_collection_config_string = options.wiredtiger_collection_config_string wiredtiger_index_config_string = options.wiredtiger_index_config_string set_parameters = options.set_parameters set_parameters_mongos = options.set_parameters_mongos no_preallocj = options.no_preallocj auth = options.auth authMechanism = options.authMechanism keyFile = options.keyFile rlp_path = options.rlp_path clean_every_n_tests = options.clean_every_n_tests clean_whole_dbroot = options.with_cleanbb if auth and not keyFile: # if only --auth was given to smoke.py, load the # default keyFile from jstests/libs/authTestsKey keyFile = os.path.join(mongo_repo, 'jstests', 'libs', 'authTestsKey') if keyFile: f = open(keyFile, 'r') keyFileData = re.sub(r'\s', '', f.read()) # Remove all whitespace f.close() os.chmod(keyFile, stat.S_IRUSR | stat.S_IWUSR) else: keyFileData = None # if smoke.py is running a list of commands read from a # file (or stdin) rather than running a suite of js tests file_of_commands_mode = options.File and options.mode == 'files' # generate json report report_file = options.report_file temp_path = options.temp_path use_write_commands = options.use_write_commands shell_write_mode = options.shell_write_mode def file_version(): return md5(open(__file__, 'r').read()).hexdigest() def clear_failfile(): if os.path.exists(failfile): os.remove(failfile) def run_old_fails(): global tests try: f = open(failfile, 'r') state = pickle.load(f) f.close() except Exception: try: f.close() except: pass clear_failfile() return # This counts as passing so we will run all tests if ('version' not in state or state['version'] != file_version()): print "warning: old version of failfile.smoke detected. skipping recent fails" clear_failfile() return testsAndOptions = state['testsAndOptions'] tests = [x[0] for x in testsAndOptions] passed = [] try: for (i, (test, options)) in enumerate(testsAndOptions): # SERVER-5102: until we can figure out a better way to manage # dependencies of the --only-old-fails build phase, just skip # tests which we can't safely run at this point path, usedb = test if not os.path.exists(path): passed.append(i) winners.append(test) continue filename = os.path.basename(path) if filename in ('dbtest', 'dbtest.exe') or filename.endswith('.js'): set_globals(options, [filename]) oldWinners = len(winners) run_tests([test]) if len(winners) != oldWinners: # can't use return value due to continue_on_failure passed.append(i) finally: for offset, i in enumerate(passed): testsAndOptions.pop(i - offset) if testsAndOptions: f = open(failfile, 'w') state = {'version':file_version(), 'testsAndOptions':testsAndOptions} pickle.dump(state, f) else: clear_failfile() report() # exits with failure code if there is an error def add_to_failfile(tests, options): try: f = open(failfile, 'r') testsAndOptions = pickle.load(f)["testsAndOptions"] except Exception: testsAndOptions = [] for test in tests: if (test, options) not in testsAndOptions: testsAndOptions.append( (test, options) ) state = {'version':file_version(), 'testsAndOptions':testsAndOptions} f = open(failfile, 'w') pickle.dump(state, f) def main(): global mongod_executable, mongod_port, shell_executable, continue_on_failure, small_oplog global no_journal, set_parameters, set_parameters_mongos, no_preallocj, auth, storage_engine, wiredtiger_engine_config_string, wiredtiger_collection_config_string, wiredtiger_index_config_string global keyFile, smoke_db_prefix, test_path, use_write_commands, rlp_path try: signal.signal(signal.SIGUSR1, dump_stacks) except AttributeError: print "Cannot catch signals on Windows" parser = OptionParser(usage="usage: smoke.py [OPTIONS] ARGS*") parser.add_option('--mode', dest='mode', default='suite', help='If "files", ARGS are filenames; if "suite", ARGS are sets of tests (%default)') # Some of our tests hard-code pathnames e.g., to execute, so until # that changes we don't have the freedom to run from anyplace. # parser.add_option('--mongo-repo', dest='mongo_repo', default=None, parser.add_option('--test-path', dest='test_path', default=None, help="Path to the test executables to run, " "currently only used for 'client' (%default)") parser.add_option('--mongod', dest='mongod_executable', default=os.path.join(mongo_repo, 'mongod'), help='Path to mongod to run (%default)') parser.add_option('--port', dest='mongod_port', default="27999", help='Port the mongod will bind to (%default)') parser.add_option('--mongo', dest='shell_executable', default=os.path.join(mongo_repo, 'mongo'), help='Path to mongo, for .js test files (%default)') parser.add_option('--continue-on-failure', dest='continue_on_failure', action="store_true", default=False, help='If supplied, continue testing even after a test fails') parser.add_option('--from-file', dest='File', help="Run tests/suites named in FILE, one test per line, '-' means stdin") parser.add_option('--smoke-db-prefix', dest='smoke_db_prefix', default=smoke_db_prefix, help="Prefix to use for the mongods' dbpaths ('%default')") parser.add_option('--small-oplog', dest='small_oplog', default=False, action="store_true", help='Run tests with master/slave replication & use a small oplog') parser.add_option('--small-oplog-rs', dest='small_oplog_rs', default=False, action="store_true", help='Run tests with replica set replication & use a small oplog') parser.add_option('--storageEngine', dest='storage_engine', default=None, help='What storage engine to start mongod with') parser.add_option('--wiredTigerEngineConfig', dest='wiredtiger_engine_config_string', default=None, help='Wired Tiger configuration to pass through to mongod') parser.add_option('--wiredTigerCollectionConfig', dest='wiredtiger_collection_config_string', default=None, help='Wired Tiger collection configuration to pass through to mongod') parser.add_option('--wiredTigerIndexConfig', dest='wiredtiger_index_config_string', default=None, help='Wired Tiger index configuration to pass through to mongod') parser.add_option('--nojournal', dest='no_journal', default=False, action="store_true", help='Do not turn on journaling in tests') parser.add_option('--nopreallocj', dest='no_preallocj', default=False, action="store_true", help='Do not preallocate journal files in tests') parser.add_option('--auth', dest='auth', default=False, action="store_true", help='Run standalone mongods in tests with authentication enabled') parser.add_option('--authMechanism', dest='authMechanism', default='SCRAM-SHA-1', help='Use the given authentication mechanism, when --auth is used.') parser.add_option('--keyFile', dest='keyFile', default=None, help='Path to keyFile to use to run replSet and sharding tests with authentication enabled') parser.add_option('--ignore', dest='ignore_files', default=None, help='Pattern of files to ignore in tests') parser.add_option('--only-old-fails', dest='only_old_fails', default=False, action="store_true", help='Check the failfile and only run all tests that failed last time') parser.add_option('--reset-old-fails', dest='reset_old_fails', default=False, action="store_true", help='Clear the failfile. Do this if all tests pass') parser.add_option('--with-cleanbb', dest='with_cleanbb', action="store_true", default=False, help='Clear database files before first test') parser.add_option('--clean-every', dest='clean_every_n_tests', type='int', default=(1 if 'detect_leaks=1' in os.getenv("ASAN_OPTIONS", "") else 20), help='Clear database files every N tests [default %default]') parser.add_option('--dont-start-mongod', dest='start_mongod', default=True, action='store_false', help='Do not start mongod before commencing test running') parser.add_option('--use-ssl', dest='use_ssl', default=False, action='store_true', help='Run mongo shell and mongod instances with SSL encryption') parser.add_option('--set-parameters', dest='set_parameters', default="", help='Adds --setParameter to mongod for each passed in item in the csv list - ex. "param1=1,param2=foo" ') parser.add_option('--set-parameters-mongos', dest='set_parameters_mongos', default="", help='Adds --setParameter to mongos for each passed in item in the csv list - ex. "param1=1,param2=foo" ') parser.add_option('--temp-path', dest='temp_path', default=None, help='If present, passed as --tempPath to unittests and dbtests or TestData.tmpPath to mongo') # Buildlogger invocation from command line parser.add_option('--buildlogger-builder', dest='buildlogger_builder', default=os.getenv("BUILDLOGGER_BUILDER"), action="store", help='Set the "builder name" for buildlogger') parser.add_option('--buildlogger-buildnum', dest='buildlogger_buildnum', default=os.getenv("BUILDLOGGER_NUMBER"), action="store", help='Set the "build number" for buildlogger') parser.add_option('--buildlogger-url', dest='buildlogger_url', default=os.getenv('BUILDLOGGER_URL'), action="store", help='Set the url root for the buildlogger service') parser.add_option('--buildlogger-credentials', dest='buildlogger_credentials', default=os.getenv("BUILDLOGGER_CREDENTIALS"), action="store", help='Path to Python file containing buildlogger credentials') parser.add_option('--buildlogger-phase', dest='buildlogger_phase', default=os.getenv("BUILDLOGGER_PHASE"), action="store", help='Set the "phase" for buildlogger (e.g. "core", "auth") for display in the webapp (optional)') parser.add_option('--report-file', dest='report_file', default=None, action='store', help='Path to generate detailed json report containing all test details') parser.add_option('--use-write-commands', dest='use_write_commands', default=False, action='store_true', help='Deprecated(use --shell-write-mode): Sets the shell to use write commands by default') parser.add_option('--shell-write-mode', dest='shell_write_mode', default="commands", help='Sets the shell to use a specific write mode: commands/compatibility/legacy (default:legacy)') parser.add_option('--basisTechRootDirectory', dest='rlp_path', default=None, help='Basis Tech Rosette Linguistics Platform root directory') global tests (options, tests) = parser.parse_args() set_globals(options, tests) buildlogger_opts = (options.buildlogger_builder, options.buildlogger_buildnum, options.buildlogger_credentials) if all(buildlogger_opts): os.environ['MONGO_USE_BUILDLOGGER'] = 'true' os.environ['MONGO_BUILDER_NAME'] = options.buildlogger_builder os.environ['MONGO_BUILD_NUMBER'] = options.buildlogger_buildnum os.environ['BUILDLOGGER_CREDENTIALS'] = options.buildlogger_credentials if options.buildlogger_phase: os.environ['MONGO_PHASE'] = options.buildlogger_phase elif any(buildlogger_opts): # some but not all of the required options were sete raise Exception("you must set all of --buildlogger-builder, --buildlogger-buildnum, --buildlogger-credentials") if options.buildlogger_url: #optional; if None, defaults to const in buildlogger.py os.environ['BUILDLOGGER_URL'] = options.buildlogger_url if options.File: if options.File == '-': tests = sys.stdin.readlines() else: f = open(options.File) tests = f.readlines() tests = [t.rstrip('\n') for t in tests] if options.only_old_fails: run_old_fails() return elif options.reset_old_fails: clear_failfile() return # If we're in suite mode, tests is a list of names of sets of tests. if options.mode == 'suite': tests = expand_suites(tests) elif options.mode == 'files': tests = [(os.path.abspath(test), start_mongod) for test in tests] if options.ignore_files != None : ignore_patt = re.compile( options.ignore_files ) print "Ignoring files with pattern: ", ignore_patt def ignore_test( test ): if ignore_patt.search( test[0] ) != None: print "Ignoring test ", test[0] return False else: return True tests = filter( ignore_test, tests ) if not tests: print "warning: no tests specified" return if options.with_cleanbb: clean_dbroot(nokill=True) test_report["start"] = time.time() test_report["mongod_running_at_start"] = mongod().is_mongod_up(mongod_port) try: run_tests(tests) finally: add_to_failfile(fails, options) test_report["end"] = time.time() test_report["elapsed"] = test_report["end"] - test_report["start"] test_report["failures"] = len(losers.keys()) test_report["mongod_running_at_end"] = mongod().is_mongod_up(mongod_port) if report_file: f = open( report_file, "wb" ) f.write( json.dumps( test_report, indent=4, separators=(',', ': ')) ) f.close() report() if __name__ == "__main__": main()