def test_rpma_fio_bench(tool_mode, readwrite, mode, monkeypatch):
"""test all arguments variants of rpma_fio_bench.sh"""
def run_mock(args, env):
assert len(args) == 5
# XXX is it always correct to assume the tool is in the current working
# directory?
assert args[0] == './' + TOOL_RPMA_FIO_BENCH
assert args[1] == IP_DUMMY
assert args[2] == tool_mode
assert args[3] == readwrite
assert args[4] == mode
assert env['OUTPUT_FILE'] == output_file(ID_DUMMY, RESULT_DIR)
return ProcessMock()
monkeypatch.setattr(subprocess, 'run', run_mock)
oneseries = {
**ONESERIES_DUMMY, 'tool': TOOL_RPMA_FIO_BENCH,
'mode': mode,
'tool_mode': tool_mode,
'rw': readwrite,
'busy_wait_polling': True
}
benchmark = Benchmark(oneseries)
benchmark.run(CONFIG_DUMMY, RESULT_DIR)
assert benchmark.is_done()
def tearDown(self):
command = 'deletetable %s -f\n' % self.tablename
log.debug("Running Command %r", command)
code, out, err = cloudshell.run(self.username, self.password, command)
self.assertEqual(code, 0, "Could not delete table")
log.debug("Process finished")
Benchmark.tearDown(self)
def tearDown(self):
command = 'deletetable %s -f\n' % self.tablename
log.debug("Running Command %r", command)
code, out, err = cloudshell.run(self.username, self.password, command)
self.assertEqual(code, 0, "Could not delete table")
log.debug("Process finished")
Benchmark.tearDown(self)
def setUp(self):
random.jumpahead(int(time.time()))
num = random.randint(1, 100000)
self.input_table = self.input_table + "_" + str(num)
self.output_table = self.output_table + "_" + str(num)
#if (not os.getenv("HADOOP_CLASSPATH")):
# os.putenv("HADOOP_CLASSPATH", self.getjars(":"))
dir = os.path.dirname(os.path.realpath(__file__))
file = os.path.join( dir, 'splits' )
# code, out, err = cloudshell.run(self.username, self.password, 'table RowHashTestInput\n')
# if out.find('no such table') == -1:
# code, out, err = cloudshell.run(self.username, self.password, 'deletetable RowHashTestInput\n')
# self.sleep(15)
code, out, err = cloudshell.run(self.username, self.password, "createtable %s -sf %s\n" % (self.input_table, file))
#code, out, err = cloudshell.run('table RowHashTest\n')
#if out.find('no such table') == -1:
# code, out, err = cloudshell.run('user root\nsecret\ndeletetable RowHashTest\n')
# self.sleep(15)
code, out, err = cloudshell.run(self.username, self.password, "createtable %s -sf %s\n" % (self.output_table, file))
command = self.buildcommand('org.apache.accumulo.examples.simple.mapreduce.TeraSortIngest',
self.numrows(),
self.keysizemin(),
self.keysizemax(),
self.minvaluesize(),
self.maxvaluesize(),
self.input_table,
self.getInstance(),
self.getZookeepers(),
self.getUsername(),
self.getPassword(),
self.maxmaps)
handle = runner.start(command, stdin=subprocess.PIPE)
log.debug("Running: %r", command)
out, err = handle.communicate("")
Benchmark.setUp(self)
def setUp(self):
random.jumpahead(int(time.time()))
num = random.randint(1, 100000)
self.input_table = self.input_table + "_" + str(num)
self.output_table = self.output_table + "_" + str(num)
#if (not os.getenv("HADOOP_CLASSPATH")):
# os.putenv("HADOOP_CLASSPATH", self.getjars(":"))
dir = os.path.dirname(os.path.realpath(__file__))
file = os.path.join(dir, 'splits')
# code, out, err = cloudshell.run(self.username, self.password, 'table RowHashTestInput\n')
# if out.find('no such table') == -1:
# code, out, err = cloudshell.run(self.username, self.password, 'deletetable RowHashTestInput\n')
# self.sleep(15)
code, out, err = cloudshell.run(
self.username, self.password,
"createtable %s -sf %s\n" % (self.input_table, file))
#code, out, err = cloudshell.run('table RowHashTest\n')
#if out.find('no such table') == -1:
# code, out, err = cloudshell.run('user root\nsecret\ndeletetable RowHashTest\n')
# self.sleep(15)
code, out, err = cloudshell.run(
self.username, self.password,
"createtable %s -sf %s\n" % (self.output_table, file))
command = self.buildcommand(
'org.apache.accumulo.examples.simple.mapreduce.TeraSortIngest',
self.numrows(), self.keysizemin(), self.keysizemax(),
self.minvaluesize(), self.maxvaluesize(), self.input_table,
self.getInstance(), self.getZookeepers(), self.getUsername(),
self.getPassword(), self.maxmaps)
handle = runner.start(command, stdin=subprocess.PIPE)
log.debug("Running: %r", command)
out, err = handle.communicate("")
Benchmark.setUp(self)
def setUp(self):
dir = os.path.dirname(os.path.realpath(__file__))
file = os.path.join( dir, 'splits' )
code, out, err = cloudshell.run(self.username, self.password, 'table %s\n' % self.input_table)
if out.find('does not exist') == -1:
code, out, err = cloudshell.run(self.username, self.password, 'deletetable -f %s\n' % self.input_table)
self.sleep(15)
code, out, err = cloudshell.run(self.username, self.password, "createtable %s -sf %s\n" % (self.input_table, file))
code, out, err = cloudshell.run(self.username, self.password, 'table %s\n' % self.output_table)
if out.find('does not exist') == -1:
code, out, err = cloudshell.run(self.username, self.password, 'deletetable -f %s\n' %
self.output_table)
self.sleep(15)
code, out, err = cloudshell.run(self.username, self.password, "createtable %s -sf %s\n" % (self.output_table, file))
command = self.buildcommand('org.apache.accumulo.examples.simple.mapreduce.TeraSortIngest',
'--count', self.numrows(),
'-nk', self.keysizemin(),
'-xk', self.keysizemax(),
'-nv', self.minvaluesize(),
'-xv', self.maxvaluesize(),
'--table', self.input_table,
'-i', self.getInstance(),
'-z', self.getZookeepers(),
'-u', self.getUsername(),
'-p', self.getPassword(),
'--splits', self.maxmaps)
handle = runner.start(command, stdin=subprocess.PIPE)
log.debug("Running: %r", command)
out, err = handle.communicate("")
Benchmark.setUp(self)
def setUp(self):
# Need to generate a splits file for each speed
code, out, err = cloudshell.run(self.username, self.password, 'table %s\n' % self.tablename)
if out.find('does not exist') == -1:
log.debug('Deleting table %s' % self.tablename)
code, out, err = cloudshell.run(self.username, self.password, 'deletetable %s -f\n' % self.tablename)
self.assertEqual(code, 0, "Could not delete table")
Benchmark.setUp(self)
def setUp(self):
# Need to generate a splits file for each speed
code, out, err = cloudshell.run(self.username, self.password, 'table %s\n' % self.tablename)
if out.find('does not exist') == -1:
log.debug('Deleting table %s' % self.tablename)
code, out, err = cloudshell.run(self.username, self.password, 'deletetable %s -f\n' % self.tablename)
self.assertEqual(code, 0, "Could not delete table")
Benchmark.setUp(self)
def setUp(self):
code, out, err = cloudshell.run(self.username, self.password, 'table test_ingest\n')
if out.find('does not exist') == -1:
log.debug("Deleting table test_ingest")
code, out, err = cloudshell.run(self.username, self.password, 'deletetable test_ingest -f\n')
self.assertEquals(code, 0, "Could not delete the table 'test_ingest'")
code, out, err = cloudshell.run(self.username, self.password, 'createtable test_ingest\n')
self.assertEqual(code, 0, "Could not create the table 'test_ingest'")
Benchmark.setUp(self)
def setUp(self):
code, out, err = cloudshell.run(self.username, self.password,
'table %s\n' % self.tablename)
if out.find('does not exist') == -1:
log.debug('Deleting table %s' % self.tablename)
code, out, err = cloudshell.run(
self.username, self.password,
'deletetable -f %s\n' % self.tablename)
Benchmark.setUp(self)
def setUp(self):
code, out, err = cloudshell.run(self.username, self.password, 'table test_ingest\n')
if out.find('no such table') >= 0:
log.debug("Deleting table test_ingest")
code, out, err = cloudshell.run(self.username, self.password, 'deletetable test_ingest\n')
self.sleep(10)
code, out, err = cloudshell.run(self.username, self.password, 'createtable test_ingest\n')
self.assertEqual(code, 0)
Benchmark.setUp(self)
def tearDown(self):
code, out, err = cloudshell.run(self.username, self.password,
"deletetable %s\n" % self.input_table)
self.assertEqual(code, 0,
'Could not delete %s, %s' % (self.input_table, out))
code, out, err = cloudshell.run(self.username, self.password,
"deletetable %s\n" % self.output_table)
self.assertEqual(code, 0,
'Could not delete %s, %s' % (self.output_table, out))
Benchmark.tearDown(self)
def setUp(self):
for x in range(1, self.tables):
currentTable = 'test_ingest%d' % (x)
log.debug("Checking for table existence: %s" % currentTable)
code, out, err = cloudshell.run(self.username, self.password, 'table %s\n' % currentTable)
if out.find('does not exist') == -1:
command = 'deletetable -f %s\n' % (currentTable)
log.debug("Running Command %r", command)
code, out, err = cloudshell.run(self.username, self.password, command)
self.assertEqual(code, 0, 'Did not successfully delete table: %s' % currentTable)
Benchmark.setUp(self)
def setUp(self):
random.jumpahead(int(time.time()))
num = random.randint(1, 100000)
#self.tablename = self.tablename + "-" + str(num)
# Find which hadoop version
# code, out, err = cloudshell.run(self.username, self.password, 'table %s\n' % self.tablename)
#if out.find('no such table') == -1:
# log.debug('Deleting table %s' % self.tablename)
# code, out, err = cloudshell.run(self.username, self.password, 'deletetable %s\n' % self.tablename)
# self.sleep(10)
Benchmark.setUp(self)
def test_incomplete_benchmark(key, monkeypatch):
"""an incomplete benchmark definition"""
def run_mock(_args, **_):
assert False, "subprocess.run() should not be called"
monkeypatch.setattr(subprocess, 'run', run_mock)
oneseries = {**ONESERIES_DUMMY}
oneseries.pop(key, None)
benchmark = Benchmark(oneseries)
with pytest.raises(ValueError):
benchmark.run(CONFIG_DUMMY, RESULT_DIR)
assert not benchmark.is_done()
def setUp(self):
code, out, err = cloudshell.run(self.username, self.password,
'table test_ingest\n')
if out.find('no such table') >= 0:
log.debug("Deleting table test_ingest")
code, out, err = cloudshell.run(self.username, self.password,
'deletetable test_ingest\n')
self.sleep(10)
code, out, err = cloudshell.run(self.username, self.password,
'createtable test_ingest\n')
self.assertEqual(code, 0)
Benchmark.setUp(self)
def setUp(self):
random.jumpahead(int(time.time()))
num = random.randint(1, 100000)
self.tablename = self.tablename + "_" + str(num)
# Need to generate a splits file for each speed
#code, out, err = cloudshell.run(self.username, self.password, 'table %s\n' % self.tablename)
#if out.find('no such table') == -1:
# log.debug('Deleting table %s' % self.tablename)
# code, out, err = cloudshell.run('user %s\n%s\ndeletetable %s\n' % (self.user,
# self.password,
# self.tablename))
# self.sleep(5)
Benchmark.setUp(self)
def setUp(self):
code, out, err = cloudshell.run(self.username, self.password,
'table test_ingest\n')
if out.find('does not exist') == -1:
log.debug("Deleting table test_ingest")
code, out, err = cloudshell.run(self.username, self.password,
'deletetable test_ingest -f\n')
self.assertEquals(code, 0,
"Could not delete the table 'test_ingest'")
code, out, err = cloudshell.run(self.username, self.password,
'createtable test_ingest\n')
self.assertEqual(code, 0, "Could not create the table 'test_ingest'")
Benchmark.setUp(self)
def test_busy_wait_polling(config_busy_wait_polling, busy_wait_polling,
expected_busy_wait_polling, monkeypatch):
"""busy_wait_polling to BUSY_WAIT_POLLING mapping"""
def run_mock(_args, env):
assert env['BUSY_WAIT_POLLING'] == expected_busy_wait_polling
return ProcessMock()
monkeypatch.setattr(subprocess, 'run', run_mock)
oneseries = {**ONESERIES_DUMMY, 'busy_wait_polling': busy_wait_polling}
benchmark = Benchmark(oneseries)
config = {**CONFIG_BIG, 'BUSY_WAIT_POLLING': config_busy_wait_polling}
benchmark.run(config, RESULT_DIR)
assert benchmark.is_done()
def test_filetype_pmem_no_mem_path(config_remote_job_mem_path, monkeypatch):
"""filetype=pmem when no REMOTE_JOB_MEM_PATH provided"""
def run_mock(_args, **_):
assert False, "subprocess.run() should not be called"
monkeypatch.setattr(subprocess, 'run', run_mock)
oneseries = {**ONESERIES_DUMMY, 'filetype': 'pmem'}
benchmark = Benchmark(oneseries)
if config_remote_job_mem_path is None:
CONFIG_BIG.pop('REMOTE_JOB_MEM_PATH', None)
else:
CONFIG_BIG['REMOTE_JOB_MEM_PATH'] = config_remote_job_mem_path
with pytest.raises(ValueError):
benchmark.run(CONFIG_BIG, RESULT_DIR)
assert not benchmark.is_done()
def setUp(self):
for x in range(1, self.tables):
currentTable = 'test_ingest%d' % (x)
log.debug("Checking for table existence: %s" % currentTable)
code, out, err = cloudshell.run(self.username, self.password,
'table %s\n' % currentTable)
if out.find('does not exist') == -1:
command = 'deletetable -f %s\n' % (currentTable)
log.debug("Running Command %r", command)
code, out, err = cloudshell.run(self.username, self.password,
command)
self.assertEqual(
code, 0,
'Did not successfully delete table: %s' % currentTable)
Benchmark.setUp(self)
def test_gpspm_no_busy_wait_polling(readwrite, mode, monkeypatch):
"""filetype=pmem when no REMOTE_JOB_MEM_PATH provided"""
def run_mock(_args, **_):
assert False, "subprocess.run() should not be called"
monkeypatch.setattr(subprocess, 'run', run_mock)
oneseries = {
**ONESERIES_DUMMY, 'tool': TOOL_RPMA_FIO_BENCH,
'mode': mode,
'tool_mode': 'gpspm',
'rw': readwrite
}
benchmark = Benchmark(oneseries)
with pytest.raises(ValueError):
benchmark.run(CONFIG_BIG, RESULT_DIR)
assert not benchmark.is_done()
def test_ib_read(mode, monkeypatch):
"""test all arguments variants of ib_read.sh"""
def run_mock(args, env):
assert len(args) == 3
# XXX is it always correct to assume the tool is in the current working
# directory?
assert args[0] == './' + TOOL_IB_READ
assert args[1] == IP_DUMMY
assert args[2] == mode
assert env['OUTPUT_FILE'] == output_file(ID_DUMMY, RESULT_DIR)
return ProcessMock()
monkeypatch.setattr(subprocess, 'run', run_mock)
oneseries = {**ONESERIES_DUMMY, 'tool': TOOL_IB_READ, 'mode': mode}
benchmark = Benchmark(oneseries)
benchmark.run(CONFIG_DUMMY, RESULT_DIR)
assert benchmark.is_done()
def test_filetype(filetype, monkeypatch):
"""filetype to REMOTE_JOB_MEM_PATH mapping"""
# sanity check
assert CONFIG_BIG['REMOTE_JOB_MEM_PATH'] != 'malloc'
def run_mock(_args, env):
if filetype == 'malloc':
assert env['REMOTE_JOB_MEM_PATH'] == 'malloc'
else:
assert env['REMOTE_JOB_MEM_PATH'] == \
CONFIG_BIG['REMOTE_JOB_MEM_PATH']
return ProcessMock()
monkeypatch.setattr(subprocess, 'run', run_mock)
oneseries = {**ONESERIES_DUMMY, 'filetype': filetype}
benchmark = Benchmark(oneseries)
benchmark.run(CONFIG_BIG, RESULT_DIR)
assert benchmark.is_done()
def setUp(self):
Benchmark.setUp(self)
def tearDown(self):
Benchmark.tearDown(self)
def tearDown(self):
command = 'deletetable test_ingest -f\n'
log.debug("Running Command %r", command)
code, out, err = cloudshell.run(self.username, self.password, command)
self.assertEqual(code, 0, "Could not delete the table 'test_ingest'")
Benchmark.tearDown(self)
def tearDown(self):
code, out, err = cloudshell.run(self.username, self.password, "deletetable -f %s\n" % self.tablename)
self.assertEqual(code, 0, "Could not delete %s, %s" % (self.tablename, out))
Benchmark.tearDown(self)
def setUp(self):
code, out, err = cloudshell.run(self.username, self.password, 'table %s\n' % self.tablename)
if out.find('does not exist') == -1:
log.debug('Deleting table %s' % self.tablename)
code, out, err = cloudshell.run(self.username, self.password, 'deletetable -f %s\n' % self.tablename)
Benchmark.setUp(self)
def tearDown(self):
command = 'deletetable test_ingest -f\n'
log.debug("Running Command %r", command)
code, out, err = cloudshell.run(self.username, self.password, command)
self.assertEqual(code, 0, "Could not delete the table 'test_ingest'")
Benchmark.tearDown(self)
def fixture_benchmark_dummy():
"""create a very simple Benchmark instance"""
return Benchmark({**ONESERIES_DUMMY})
def tearDown(self):
code, out, err = cloudshell.run(self.username, self.password, "deletetable %s\n" % self.input_table)
self.assertEqual(code, 0, 'Could not delete %s, %s' % (self.input_table, out))
code, out, err = cloudshell.run(self.username, self.password, "deletetable %s\n" % self.output_table)
self.assertEqual(code, 0, 'Could not delete %s, %s' % (self.output_table, out))
Benchmark.tearDown(self)
def setUp(self):
code, out, err = cloudshell.run(self.username, self.password, "table %s\n" % self.tablename)
if out.find("does not exist") == -1:
log.debug("Deleting table %s" % self.tablename)
code, out, err = cloudshell.run(self.username, self.password, "deletetable -f %s\n" % self.tablename)
Benchmark.setUp(self)
