Merge pull request #365 from jkleint/devel

Fix hang on large command output.
pull/364/merge
Michael DeHaan 13 years ago
commit 38ad3da63b

@ -164,7 +164,7 @@ class ParamikoConnection(object):
stdin = chan.makefile('wb', bufsize)
stdout = chan.makefile('rb', bufsize)
stderr = chan.makefile_stderr('rb', bufsize) # stderr goes to stdout when using a pty, so this will never output anything.
stderr = '' # stderr goes to stdout when using a pty, so this will never output anything.
return stdin, stdout, stderr
def put_file(self, in_path, out_path):

@ -10,9 +10,9 @@ import os
import shutil
import time
try:
import json
import json
except:
import simplejson as json
import simplejson as json
from nose.plugins.skip import SkipTest
@ -25,202 +25,206 @@ def get_binary(name):
class TestRunner(unittest.TestCase):
def setUp(self):
self.user = getpass.getuser()
self.runner = ansible.runner.Runner(
module_name='ping',
module_path='library/',
module_args='',
remote_user=self.user,
remote_pass=None,
host_list='test/ansible_hosts',
timeout=5,
forks=1,
background=0,
pattern='all',
)
self.cwd = os.getcwd()
self.test_dir = os.path.join(self.cwd, 'test')
self.stage_dir = self._prepare_stage_dir()
def _prepare_stage_dir(self):
stage_path = os.path.join(self.test_dir, 'test_data')
if os.path.exists(stage_path):
shutil.rmtree(stage_path, ignore_errors=False)
assert not os.path.exists(stage_path)
os.makedirs(stage_path)
assert os.path.exists(stage_path)
return stage_path
def _get_test_file(self, filename):
# get a file inside the test input directory
filename = os.path.join(self.test_dir, filename)
assert os.path.exists(filename)
return filename
def _get_stage_file(self, filename):
# get a file inside the test output directory
filename = os.path.join(self.stage_dir, filename)
return filename
def _run(self, module_name, module_args, background=0):
''' run a module and get the localhost results '''
self.runner.module_name = module_name
args = ' '.join(module_args)
print "DEBUG: using args=%s" % args
self.runner.module_args = args
self.runner.background = background
results = self.runner.run()
# when using nosetests this will only show up on failure
# which is pretty useful
print "RESULTS=%s" % results
assert "127.0.0.2" in results['contacted']
return results['contacted']['127.0.0.2']
def test_ping(self):
result = self._run('ping',[])
assert "ping" in result
def test_facter(self):
if not get_binary("facter"):
raise SkipTest
result = self._run('facter',[])
assert "hostname" in result
# temporarily disbabled since it occasionally hangs
# ohai's fault, setup module doesn't actually run this
# to get ohai's "facts" anyway
#
#def test_ohai(self):
# if not get_binary("facter"):
# raise SkipTest
# result = self._run('ohai',[])
# assert "hostname" in result
def test_copy(self):
# test copy module, change trigger, etc
pass
def test_copy(self):
input = self._get_test_file('sample.j2')
output = self._get_stage_file('sample.out')
assert not os.path.exists(output)
result = self._run('copy', [
"src=%s" % input,
"dest=%s" % output,
])
assert os.path.exists(output)
data_in = file(input).read()
data_out = file(output).read()
assert data_in == data_out
assert 'failed' not in result
assert result['changed'] == True
assert 'md5sum' in result
result = self._run('copy', [
"src=%s" % input,
"dest=%s" % output,
])
assert result['changed'] == False
def test_template(self):
input = self._get_test_file('sample.j2')
metadata = self._get_test_file('metadata.json')
output = self._get_stage_file('sample.out')
result = self._run('template', [
"src=%s" % input,
"dest=%s" % output,
"metadata=%s" % metadata
])
assert os.path.exists(output)
out = file(output).read()
assert out.find("duck") != -1
assert result['changed'] == True
assert 'md5sum' in result
assert 'failed' not in result
result = self._run('template', [
"src=%s" % input,
"dest=%s" % output,
"metadata=%s" % metadata
])
assert result['changed'] == False
def test_command(self):
# test command module, change trigger, etc
result = self._run('command', [ "/bin/echo", "hi" ])
assert "failed" not in result
assert "msg" not in result
assert result['rc'] == 0
assert result['stdout'] == 'hi'
assert result['stderr'] == ''
result = self._run('command', [ "/bin/false" ])
assert result['rc'] == 1
assert 'failed' not in result
result = self._run('command', [ "/usr/bin/this_does_not_exist", "splat" ])
assert 'msg' in result
assert 'failed' in result
assert 'rc' not in result
result = self._run('shell', [ "/bin/echo", "$HOME" ])
assert 'failed' not in result
assert result['rc'] == 0
def test_setup(self):
output = self._get_stage_file('output.json')
result = self._run('setup', [ "metadata=%s" % output, "a=2", "b=3", "c=4" ])
assert 'failed' not in result
assert 'md5sum' in result
assert result['changed'] == True
outds = json.loads(file(output).read())
assert outds['c'] == '4'
# not bothering to test change hooks here since ohai/facter results change
# almost every time so changed is always true, this just tests that
# rewriting the file is ok
result = self._run('setup', [ "metadata=%s" % output, "a=2", "b=3", "c=4" ])
print "RAW RESULT=%s" % result
assert 'md5sum' in result
def test_async(self):
# test async launch and job status
# of any particular module
result = self._run('command', [ get_binary("sleep"), "3" ], background=20)
assert 'ansible_job_id' in result
assert 'started' in result
jid = result['ansible_job_id']
# no real chance of this op taking a while, but whatever
time.sleep(5)
# CLI will abstract this (when polling), but this is how it works internally
result = self._run('async_status', [ "jid=%s" % jid ])
# TODO: would be nice to have tests for supervisory process
# killing job after X seconds
assert 'finished' in result
assert 'failed' not in result
assert 'rc' in result
assert 'stdout' in result
assert result['ansible_job_id'] == jid
def test_fetch(self):
input = self._get_test_file('sample.j2')
output = os.path.join(self.stage_dir, '127.0.0.2', input)
result = self._run('fetch', [ "src=%s" % input, "dest=%s" % self.stage_dir ])
assert os.path.exists(output)
assert open(input).read() == open(output).read()
def test_yum(self):
if not get_binary("yum"):
raise SkipTest
result = self._run('yum', [ "list=repos" ])
assert 'failed' not in result
def test_git(self):
# TODO: tests for the git module
pass
def test_service(self):
# TODO: tests for the service module
pass
def setUp(self):
self.user = getpass.getuser()
self.runner = ansible.runner.Runner(
module_name='ping',
module_path='library/',
module_args='',
remote_user=self.user,
remote_pass=None,
host_list='test/ansible_hosts',
timeout=5,
forks=1,
background=0,
pattern='all',
)
self.cwd = os.getcwd()
self.test_dir = os.path.join(self.cwd, 'test')
self.stage_dir = self._prepare_stage_dir()
def _prepare_stage_dir(self):
stage_path = os.path.join(self.test_dir, 'test_data')
if os.path.exists(stage_path):
shutil.rmtree(stage_path, ignore_errors=False)
assert not os.path.exists(stage_path)
os.makedirs(stage_path)
assert os.path.exists(stage_path)
return stage_path
def _get_test_file(self, filename):
# get a file inside the test input directory
filename = os.path.join(self.test_dir, filename)
assert os.path.exists(filename)
return filename
def _get_stage_file(self, filename):
# get a file inside the test output directory
filename = os.path.join(self.stage_dir, filename)
return filename
def _run(self, module_name, module_args, background=0):
''' run a module and get the localhost results '''
self.runner.module_name = module_name
args = ' '.join(module_args)
print "DEBUG: using args=%s" % args
self.runner.module_args = args
self.runner.background = background
results = self.runner.run()
# when using nosetests this will only show up on failure
# which is pretty useful
print "RESULTS=%s" % results
assert "127.0.0.2" in results['contacted']
return results['contacted']['127.0.0.2']
def test_ping(self):
result = self._run('ping', [])
assert "ping" in result
def test_facter(self):
if not get_binary("facter"):
raise SkipTest
result = self._run('facter', [])
assert "hostname" in result
# temporarily disbabled since it occasionally hangs
# ohai's fault, setup module doesn't actually run this
# to get ohai's "facts" anyway
#
#def test_ohai(self):
# if not get_binary("facter"):
# raise SkipTest
# result = self._run('ohai',[])
# assert "hostname" in result
def test_copy(self):
# test copy module, change trigger, etc
input_ = self._get_test_file('sample.j2')
output = self._get_stage_file('sample.out')
assert not os.path.exists(output)
result = self._run('copy', [
"src=%s" % input_,
"dest=%s" % output,
])
assert os.path.exists(output)
data_in = file(input_).read()
data_out = file(output).read()
assert data_in == data_out
assert 'failed' not in result
assert result['changed'] == True
assert 'md5sum' in result
result = self._run('copy', [
"src=%s" % input_,
"dest=%s" % output,
])
assert result['changed'] == False
def test_template(self):
input_ = self._get_test_file('sample.j2')
metadata = self._get_test_file('metadata.json')
output = self._get_stage_file('sample.out')
result = self._run('template', [
"src=%s" % input_,
"dest=%s" % output,
"metadata=%s" % metadata
])
assert os.path.exists(output)
out = file(output).read()
assert out.find("duck") != -1
assert result['changed'] == True
assert 'md5sum' in result
assert 'failed' not in result
result = self._run('template', [
"src=%s" % input_,
"dest=%s" % output,
"metadata=%s" % metadata
])
assert result['changed'] == False
def test_command(self):
# test command module, change trigger, etc
result = self._run('command', [ "/bin/echo", "hi" ])
assert "failed" not in result
assert "msg" not in result
assert result['rc'] == 0
assert result['stdout'] == 'hi'
assert result['stderr'] == ''
result = self._run('command', [ "/bin/false" ])
assert result['rc'] == 1
assert 'failed' not in result
result = self._run('command', [ "/usr/bin/this_does_not_exist", "splat" ])
assert 'msg' in result
assert 'failed' in result
assert 'rc' not in result
result = self._run('shell', [ "/bin/echo", "$HOME" ])
assert 'failed' not in result
assert result['rc'] == 0
def test_large_output(self):
# Ensure reading a large amount of output from a command doesn't hang.
result = self._run('command', [ "/bin/cat", "/usr/share/dict/words" ])
assert "failed" not in result
assert "msg" not in result
assert result['rc'] == 0
assert len(result['stdout']) > 100000
assert result['stderr'] == ''
def test_setup(self):
output = self._get_stage_file('output.json')
result = self._run('setup', [ "metadata=%s" % output, "a=2", "b=3", "c=4" ])
assert 'failed' not in result
assert 'md5sum' in result
assert result['changed'] == True
outds = json.loads(file(output).read())
assert outds['c'] == '4'
# not bothering to test change hooks here since ohai/facter results change
# almost every time so changed is always true, this just tests that
# rewriting the file is ok
result = self._run('setup', [ "metadata=%s" % output, "a=2", "b=3", "c=4" ])
print "RAW RESULT=%s" % result
assert 'md5sum' in result
def test_async(self):
# test async launch and job status
# of any particular module
result = self._run('command', [ get_binary("sleep"), "3" ], background=20)
assert 'ansible_job_id' in result
assert 'started' in result
jid = result['ansible_job_id']
# no real chance of this op taking a while, but whatever
time.sleep(5)
# CLI will abstract this (when polling), but this is how it works internally
result = self._run('async_status', [ "jid=%s" % jid ])
# TODO: would be nice to have tests for supervisory process
# killing job after X seconds
assert 'finished' in result
assert 'failed' not in result
assert 'rc' in result
assert 'stdout' in result
assert result['ansible_job_id'] == jid
def test_fetch(self):
input_ = self._get_test_file('sample.j2')
output = os.path.join(self.stage_dir, '127.0.0.2', input_)
result = self._run('fetch', [ "src=%s" % input_, "dest=%s" % self.stage_dir ])
assert os.path.exists(output)
assert open(input_).read() == open(output).read()
def test_yum(self):
if not get_binary("yum"):
raise SkipTest
result = self._run('yum', [ "list=repos" ])
assert 'failed' not in result
def test_git(self):
# TODO: tests for the git module
pass
def test_service(self):
# TODO: tests for the service module
pass

Loading…
Cancel
Save