zuul/zuul
Code Issues Proposed changes
Files
master
zuul/tests/unit/test_zk.py
James E. Blair 263d766738 Expose launcher locks thorugh the web api 
The launcher uses four tree caches: requests, nodes, artifacts,
uploads.  The web server also has those caches so that it can answer
queries about their contents.  It would be useful to be able to
show what component has locked a given artifact, upload, or request,
or node.  To do so, this change does the following:

* Updates the ZuulTreeCache so that the parsePath method can indicate
  whether the cache should fetch the data in the object.  This is so
  that we can fetch the contents of lock contender nodes, but still
  avoid any other unecessary fetches.
* Stores information about the component in the ZKContext that we create
  and use for locking ZKObjects.
* Sets the lock identifier (which is stored in the lock contender node)
  to the component's hostname (retrieved from the ZKContext).
* Updates the LockableZKObjectCache to set the name of the current
  holder (the winning contender) as a non-serialized attribute on
  the object.
* Exposes whether an object is locked and the current lock holder
  through the REST API.

The scheduler was relying on the zk_client duck-typing as a ZKContext
when it acquired some locks; but now that we need more information
from the ZKContext, this has been updated to create an extra context
when acquiring the lock.

Change-Id: I00328433fb6ecf5d965e8a4d5bf492a80b4ee1b4
2025-06-30 06:39:15 -07:00

3237 lines
112 KiB
Python
Raw Permalink Blame History

# Copyright 2019 Red Hat, Inc.
# Copyright 2022, 2024 Acme Gating, LLC
#
# Licensed under the Apache License, Version 2.0 (the "License"); you may
# not use this file except in compliance with the License. You may obtain
# a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
# WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
# License for the specific language governing permissions and limitations
# under the License.
import binascii
from collections import defaultdict
import json
import math
import queue
import random
import string
import threading
import time
import uuid
import zlib
from unittest import mock
import fixtures
import testtools
from zuul import model
from zuul.lib import yamlutil as yaml
from zuul.model import (
BuildRequest,
HoldRequest,
MergeRequest,
QuotaInformation,
)
from zuul.zk import ZooKeeperClient
from zuul.zk.blob_store import BlobStore
from zuul.zk.branch_cache import BranchCache, BranchFlag, BranchInfo
from zuul.zk.change_cache import (
AbstractChangeCache,
ChangeKey,
ConcurrentUpdateError,
)
from zuul.zk.config_cache import SystemConfigCache, UnparsedConfigCache
from zuul.zk.election import SessionAwareElection, RendezvousElection
from zuul.zk.exceptions import LockException
from zuul.zk.executor import ExecutorApi
from zuul.zk.job_request_queue import JobRequestEvent
from zuul.zk.merger import MergerApi
from zuul.zk.launcher import LauncherApi, LockableZKObjectCache
from zuul.zk.layout import LayoutStateStore, LayoutState
from zuul.zk.locks import locked
from zuul.zk.nodepool import ZooKeeperNodepool
from zuul.zk.sharding import (
RawShardIO,
BufferedShardReader,
BufferedShardWriter,
NODE_BYTE_SIZE_LIMIT,
)
from zuul.zk.components import (
BaseComponent,
ComponentRegistry,
ExecutorComponent,
LauncherComponent,
SchedulerComponent,
COMPONENT_REGISTRY
)
from tests.base import (
BaseTestCase,
HoldableExecutorApi,
HoldableMergerApi,
iterate_timeout,
model_version,
ZOOKEEPER_SESSION_TIMEOUT,
)
from zuul.zk.zkobject import (
LockableZKObject,
PolymorphicZKObjectMixin,
ShardedZKObject,
ZKContext,
ZKObject,
)
from zuul.zk.locks import tenant_write_lock
import kazoo.recipe.lock
from kazoo.exceptions import ZookeeperError, OperationTimeoutError, NoNodeError
class ZooKeeperBaseTestCase(BaseTestCase):
def setUp(self):
super().setUp()
self.setupZK()
self.zk_client = ZooKeeperClient(
self.zk_chroot_fixture.zk_hosts,
tls_cert=self.zk_chroot_fixture.zookeeper_cert,
tls_key=self.zk_chroot_fixture.zookeeper_key,
tls_ca=self.zk_chroot_fixture.zookeeper_ca,
timeout=ZOOKEEPER_SESSION_TIMEOUT,
)
self.addCleanup(self.zk_client.disconnect)
self.zk_client.connect()
self.setupModelPin()
self.component_registry = ComponentRegistry(self.zk_client)
# We don't have any other component to initialize the global
# registry in these tests, so we do it ourselves.
COMPONENT_REGISTRY.create(self.zk_client)
def getRequest(self, api, request_uuid, state=None):
for _ in iterate_timeout(30, "cache to update"):
req = api.getRequest(request_uuid)
if req:
if state is None:
return req
if req.state == state:
return req
class TestZookeeperClient(ZooKeeperBaseTestCase):
def test_ltime(self):
ltime = self.zk_client.getCurrentLtime()
self.assertGreaterEqual(ltime, 0)
self.assertIsInstance(ltime, int)
self.assertGreater(self.zk_client.getCurrentLtime(), ltime)
class TestNodepool(ZooKeeperBaseTestCase):
def setUp(self):
super().setUp()
self.zk_nodepool = ZooKeeperNodepool(self.zk_client)
def _createRequest(self):
req = HoldRequest()
req.count = 1
req.reason = 'some reason'
req.expiration = 1
return req
def test_hold_requests_api(self):
# Test no requests returns empty list
self.assertEqual([], self.zk_nodepool.getHoldRequests())
# Test get on non-existent request is None
self.assertIsNone(self.zk_nodepool.getHoldRequest('anything'))
# Test creating a new request
req1 = self._createRequest()
self.zk_nodepool.storeHoldRequest(req1)
self.assertIsNotNone(req1.id)
self.assertEqual(1, len(self.zk_nodepool.getHoldRequests()))
# Test getting the request
req2 = self.zk_nodepool.getHoldRequest(req1.id)
self.assertEqual(req1.toDict(), req2.toDict())
# Test updating the request
req2.reason = 'a new reason'
self.zk_nodepool.storeHoldRequest(req2)
req2 = self.zk_nodepool.getHoldRequest(req2.id)
self.assertNotEqual(req1.reason, req2.reason)
# Test lock operations
self.zk_nodepool.lockHoldRequest(req2, blocking=False)
with testtools.ExpectedException(
LockException, "Timeout trying to acquire lock .*"
):
self.zk_nodepool.lockHoldRequest(req2, blocking=True, timeout=2)
self.zk_nodepool.unlockHoldRequest(req2)
self.assertIsNone(req2.lock)
# Test deleting the request
self.zk_nodepool.deleteHoldRequest(req1, None)
self.assertEqual([], self.zk_nodepool.getHoldRequests())
class TestSharding(ZooKeeperBaseTestCase):
def test_reader(self):
shard_io = RawShardIO(self.zk_client.client, "/test/shards")
with testtools.ExpectedException(NoNodeError):
len(shard_io._shards)
with BufferedShardReader(
self.zk_client.client, "/test/shards"
) as shard_reader:
with testtools.ExpectedException(NoNodeError):
shard_reader.read()
shard_io.write(b"foobar")
self.assertEqual(len(shard_io._shards), 1)
self.assertEqual(shard_io.read(), b"foobar")
def test_writer(self):
shard_io = RawShardIO(self.zk_client.client, "/test/shards")
with testtools.ExpectedException(NoNodeError):
len(shard_io._shards)
with BufferedShardWriter(
self.zk_client.client, "/test/shards"
) as shard_writer:
shard_writer.write(b"foobar")
self.assertEqual(len(shard_io._shards), 1)
self.assertEqual(shard_io.read(), b"foobar")
def test_truncate(self):
shard_io = RawShardIO(self.zk_client.client, "/test/shards")
shard_io.write(b"foobar")
self.assertEqual(len(shard_io._shards), 1)
with BufferedShardWriter(
self.zk_client.client, "/test/shards"
) as shard_writer:
shard_writer.truncate(0)
with testtools.ExpectedException(NoNodeError):
len(shard_io._shards)
def test_shard_bytes_limit(self):
with BufferedShardWriter(
self.zk_client.client, "/test/shards"
) as shard_writer:
shard_writer.write(b"x" * (NODE_BYTE_SIZE_LIMIT + 1))
shard_writer.flush()
self.assertEqual(len(shard_writer.raw._shards), 2)
def test_json(self):
data = {"key": "value"}
with BufferedShardWriter(
self.zk_client.client, "/test/shards"
) as shard_io:
shard_io.write(json.dumps(data).encode("utf8"))
with BufferedShardReader(
self.zk_client.client, "/test/shards"
) as shard_io:
self.assertDictEqual(json.load(shard_io), data)
def test_no_makepath(self):
with testtools.ExpectedException(NoNodeError):
with BufferedShardWriter(
self.zk_client.client, "/test/shards",
makepath=False
) as shard_writer:
shard_writer.write(b"foobar")
def _test_write_old_read_new(self, shard_count):
# Write shards in the old format where each shard is
# compressed individually
data = b'{"key": "value"}'
data_shards = []
shard_len = math.ceil(len(data) / shard_count)
for start in range(0, len(data), shard_len):
data_shards.append(data[start:start + shard_len])
# Make sure we split them correctly
self.assertEqual(data, b''.join(data_shards))
self.assertEqual(shard_count, len(data_shards))
for shard in data_shards:
shard = zlib.compress(shard)
self.log.debug(f"{binascii.hexlify(shard)=}")
self.zk_client.client.create("/test/shards/", shard,
sequence=True, makepath=True)
# Read shards, expecting the new format
with BufferedShardReader(
self.zk_client.client, "/test/shards"
) as shard_io:
read_data = shard_io.read()
decompressed_data = zlib.decompress(read_data)
self.log.debug(f"{binascii.hexlify(read_data)=}")
self.log.debug(f"{decompressed_data=}")
self.assertEqual(decompressed_data, data)
def test_write_old_read_new_1(self):
self._test_write_old_read_new(1)
def test_write_old_read_new_2(self):
self._test_write_old_read_new(2)
def test_write_old_read_new_3(self):
self._test_write_old_read_new(3)
def _test_write_new_read_new(self, shard_count):
# Write shards in the new format
data = b'{"key": "value"}'
compressed_data = zlib.compress(data)
data_shards = []
shard_len = math.ceil(len(compressed_data) / shard_count)
for start in range(0, len(compressed_data), shard_len):
data_shards.append(compressed_data[start:start + shard_len])
# Make sure we split them correctly
self.assertEqual(compressed_data, b''.join(data_shards))
self.assertEqual(shard_count, len(data_shards))
for shard in data_shards:
self.log.debug(f"{binascii.hexlify(shard)=}")
self.zk_client.client.create("/test/shards/", shard,
sequence=True, makepath=True)
# Read shards, expecting the new format
with BufferedShardReader(
self.zk_client.client, "/test/shards"
) as shard_io:
read_data = shard_io.read()
decompressed_data = zlib.decompress(read_data)
self.log.debug(f"{binascii.hexlify(read_data)=}")
self.log.debug(f"{decompressed_data=}")
self.assertEqual(decompressed_data, data)
def test_write_new_read_new_1(self):
self._test_write_new_read_new(1)
def test_write_new_read_new_2(self):
self._test_write_new_read_new(2)
def test_write_new_read_new_3(self):
self._test_write_new_read_new(3)
def _test_write_old_read_old(self, shard_count):
# Test that the writer can write in the old format
data = b'{"key": "value is longer"}'
compressed_data = zlib.compress(data)
shard_len = math.ceil(len(compressed_data) / shard_count)
# We subtract 1024 from the size limit when writing the old
# format
size_limit = shard_len + 1024
with mock.patch("zuul.zk.sharding.NODE_BYTE_SIZE_LIMIT", size_limit):
with BufferedShardWriter(
self.zk_client.client, "/test/shards"
) as shard_writer:
shard_writer.write(compressed_data)
read_shards = []
for shard in sorted(
self.zk_client.client.get_children("/test/shards")):
self.log.debug(f"{shard=}")
read_data, _ = self.zk_client.client.get(f"/test/shards/{shard}")
self.log.debug(f"{binascii.hexlify(read_data)=}")
self.log.debug(f"{len(read_data)=}")
read_shards.append(zlib.decompress(read_data))
self.assertEqual(b"".join(read_shards), data)
self.assertEqual(shard_count, len(read_shards))
@model_version(30)
def test_write_old_read_old_1(self):
self._test_write_old_read_old(1)
@model_version(30)
def test_write_old_read_old_2(self):
self._test_write_old_read_old(2)
@model_version(30)
def test_write_old_read_old_3(self):
self._test_write_old_read_old(3)
class TestUnparsedConfigCache(ZooKeeperBaseTestCase):
def setUp(self):
super().setUp()
self.config_cache = UnparsedConfigCache(self.zk_client)
def test_files_cache(self):
master_files = self.config_cache.getFilesCache("project", "master")
with self.config_cache.readLock("project"):
self.assertEqual(len(master_files), 0)
with self.config_cache.writeLock("project"):
master_files["/path/to/file"] = "content"
with self.config_cache.readLock("project"):
self.assertEqual(master_files["/path/to/file"], "content")
self.assertEqual(len(master_files), 1)
with self.config_cache.writeLock("project"):
master_files.clear()
self.assertEqual(len(master_files), 0)
def test_valid_for(self):
tpc = model.TenantProjectConfig("project")
tpc.extra_config_files = {"foo.yaml", "bar.yaml"}
tpc.extra_config_dirs = {"foo.d/", "bar.d/"}
master_files = self.config_cache.getFilesCache("project", "master")
self.assertFalse(master_files.isValidFor(tpc, min_ltime=-1))
master_files.setValidFor(tpc.extra_config_files, tpc.extra_config_dirs,
ltime=1)
self.assertTrue(master_files.isValidFor(tpc, min_ltime=-1))
tpc.extra_config_files = set()
tpc.extra_config_dirs = set()
self.assertTrue(master_files.isValidFor(tpc, min_ltime=-1))
self.assertFalse(master_files.isValidFor(tpc, min_ltime=2))
tpc.extra_config_files = {"bar.yaml"}
tpc.extra_config_dirs = {"bar.d/"}
# Valid for subset
self.assertTrue(master_files.isValidFor(tpc, min_ltime=-1))
tpc.extra_config_files = {"foo.yaml", "bar.yaml"}
tpc.extra_config_dirs = {"foo.d/", "bar.d/", "other.d/"}
# Invalid for additional dirs
self.assertFalse(master_files.isValidFor(tpc, min_ltime=-1))
self.assertFalse(master_files.isValidFor(tpc, min_ltime=2))
tpc.extra_config_files = {"foo.yaml", "bar.yaml", "other.yaml"}
tpc.extra_config_dirs = {"foo.d/", "bar.d/"}
# Invalid for additional files
self.assertFalse(master_files.isValidFor(tpc, min_ltime=-1))
self.assertFalse(master_files.isValidFor(tpc, min_ltime=2))
def test_cache_ltime(self):
cache = self.config_cache.getFilesCache("project", "master")
self.assertEqual(cache.ltime, -1)
cache.setValidFor(set(), set(), ltime=1)
self.assertEqual(cache.ltime, 1)
def test_branch_cleanup(self):
master_files = self.config_cache.getFilesCache("project", "master")
release_files = self.config_cache.getFilesCache("project", "release")
master_files["/path/to/file"] = "content"
release_files["/path/to/file"] = "content"
self.config_cache.clearCache("project", "master")
self.assertEqual(len(master_files), 0)
self.assertEqual(len(release_files), 1)
def test_project_cleanup(self):
master_files = self.config_cache.getFilesCache("project", "master")
stable_files = self.config_cache.getFilesCache("project", "stable")
other_files = self.config_cache.getFilesCache("other", "master")
self.assertEqual(len(master_files), 0)
self.assertEqual(len(stable_files), 0)
master_files["/path/to/file"] = "content"
stable_files["/path/to/file"] = "content"
other_files["/path/to/file"] = "content"
self.assertEqual(len(master_files), 1)
self.assertEqual(len(stable_files), 1)
self.assertEqual(len(other_files), 1)
self.config_cache.clearCache("project")
self.assertEqual(len(master_files), 0)
self.assertEqual(len(stable_files), 0)
self.assertEqual(len(other_files), 1)
class TestComponentRegistry(ZooKeeperBaseTestCase):
def setUp(self):
super().setUp()
self.second_zk_client = ZooKeeperClient(
self.zk_chroot_fixture.zk_hosts,
tls_cert=self.zk_chroot_fixture.zookeeper_cert,
tls_key=self.zk_chroot_fixture.zookeeper_key,
tls_ca=self.zk_chroot_fixture.zookeeper_ca,
timeout=ZOOKEEPER_SESSION_TIMEOUT,
)
self.addCleanup(self.second_zk_client.disconnect)
self.second_zk_client.connect()
self.second_component_registry = ComponentRegistry(
self.second_zk_client)
def assertComponentAttr(self, component_name, attr_name,
attr_value, timeout=25):
for _ in iterate_timeout(
timeout,
f"{component_name} in cache has {attr_name} set to {attr_value}",
):
components = list(self.second_component_registry.all(
component_name))
if (
len(components) > 0 and
getattr(components[0], attr_name) == attr_value
):
break
def assertComponentState(self, component_name, state, timeout=25):
return self.assertComponentAttr(
component_name, "state", state, timeout
)
def assertComponentStopped(self, component_name, timeout=25):
for _ in iterate_timeout(
timeout, f"{component_name} in cache is stopped"
):
components = list(self.second_component_registry.all(
component_name))
if len(components) == 0:
break
def test_component_registry(self):
self.component_info = ExecutorComponent(self.zk_client, 'test')
self.component_info.register()
self.assertComponentState("executor", BaseComponent.STOPPED)
self.zk_client.client.stop()
self.assertComponentStopped("executor")
self.zk_client.client.start()
self.assertComponentState("executor", BaseComponent.STOPPED)
self.component_info.state = self.component_info.RUNNING
self.assertComponentState("executor", BaseComponent.RUNNING)
self.log.debug("DISCONNECT")
self.second_zk_client.client.stop()
self.second_zk_client.client.start()
self.log.debug("RECONNECT")
self.component_info.state = self.component_info.PAUSED
self.assertComponentState("executor", BaseComponent.PAUSED)
# Make sure the registry didn't create any read/write
# component objects that re-registered themselves.
components = list(self.second_component_registry.all('executor'))
self.assertEqual(len(components), 1)
self.component_info.state = self.component_info.RUNNING
self.assertComponentState("executor", BaseComponent.RUNNING)
class TestExecutorApi(ZooKeeperBaseTestCase):
def test_build_request(self):
# Test the lifecycle of a build request
request_queue = queue.Queue()
event_queue = queue.Queue()
# A callback closure for the request queue
def rq_put():
request_queue.put(None)
# and the event queue
def eq_put(br, e):
event_queue.put((br, e))
# Simulate the client side
client = ExecutorApi(self.zk_client)
self.addCleanup(client.stop)
# Simulate the server side
server = ExecutorApi(self.zk_client,
build_request_callback=rq_put,
build_event_callback=eq_put)
self.addCleanup(server.stop)
# Scheduler submits request
request = BuildRequest(
"A", None, None, "job", "job_uuid", "tenant", "pipeline", '1')
client.submit(request, {'job': 'test'})
request_queue.get(timeout=30)
# Executor receives request
reqs = list(server.next())
self.assertEqual(len(reqs), 1)
a = reqs[0]
self.assertEqual(a.uuid, 'A')
params = client.getParams(a)
self.assertEqual(params, {'job': 'test'})
client.clearParams(a)
params = client.getParams(a)
self.assertIsNone(params)
# Executor locks request
self.assertTrue(server.lock(a, blocking=False))
a.state = BuildRequest.RUNNING
server.update(a)
self.getRequest(client, a.uuid, BuildRequest.RUNNING)
# Executor should see no pending requests
reqs = list(server.next())
self.assertEqual(len(reqs), 0)
# Executor pauses build
a.state = BuildRequest.PAUSED
server.update(a)
self.getRequest(client, a.uuid, BuildRequest.PAUSED)
# Scheduler resumes build
self.assertTrue(event_queue.empty())
sched_a = self.getRequest(client, a.uuid)
client.requestResume(sched_a)
(build_request, event) = event_queue.get(timeout=30)
self.assertEqual(build_request, a)
self.assertEqual(event, JobRequestEvent.RESUMED)
# Executor resumes build
a.state = BuildRequest.RUNNING
server.update(a)
server.fulfillResume(a)
self.getRequest(client, a.uuid, BuildRequest.RUNNING)
# Scheduler cancels build
self.assertTrue(event_queue.empty())
sched_a = self.getRequest(client, a.uuid)
client.requestCancel(sched_a)
(build_request, event) = event_queue.get(timeout=30)
self.assertEqual(build_request, a)
self.assertEqual(event, JobRequestEvent.CANCELED)
# Executor aborts build
a.state = BuildRequest.COMPLETED
server.update(a)
server.fulfillCancel(a)
server.unlock(a)
self.getRequest(client, a.uuid, BuildRequest.COMPLETED)
# Scheduler removes build request on completion
client.remove(sched_a)
self.assertEqual(set(self.getZKPaths('/zuul/executor')),
set(['/zuul/executor/unzoned',
'/zuul/executor/unzoned/locks',
'/zuul/executor/unzoned/params',
'/zuul/executor/unzoned/requests',
'/zuul/executor/unzoned/result-data',
'/zuul/executor/unzoned/results',
'/zuul/executor/unzoned/waiters']))
def test_build_request_remove(self):
# Test the scheduler forcibly removing a request (perhaps the
# tenant is being deleted, so there will be no result queue).
request_queue = queue.Queue()
event_queue = queue.Queue()
def rq_put():
request_queue.put(None)
def eq_put(br, e):
event_queue.put((br, e))
# Simulate the client side
client = ExecutorApi(self.zk_client)
self.addCleanup(client.stop)
# Simulate the server side
server = ExecutorApi(self.zk_client,
build_request_callback=rq_put,
build_event_callback=eq_put)
self.addCleanup(server.stop)
# Scheduler submits request
request = BuildRequest(
"A", None, None, "job", "job_uuid", "tenant", "pipeline", '1')
client.submit(request, {})
request_queue.get(timeout=30)
# Executor receives request
reqs = list(server.next())
self.assertEqual(len(reqs), 1)
a = reqs[0]
self.assertEqual(a.uuid, 'A')
# Executor locks request
self.assertTrue(server.lock(a, blocking=False))
a.state = BuildRequest.RUNNING
server.update(a)
self.getRequest(client, a.uuid, MergeRequest.RUNNING)
# Executor should see no pending requests
reqs = list(server.next())
self.assertEqual(len(reqs), 0)
self.assertTrue(event_queue.empty())
# Scheduler rudely removes build request
sched_a = self.getRequest(client, a.uuid)
client.remove(sched_a)
# Make sure it shows up as deleted
(build_request, event) = event_queue.get(timeout=30)
self.assertEqual(build_request, a)
self.assertEqual(event, JobRequestEvent.DELETED)
# Executor should not write anything else since the request
# was deleted.
def test_build_request_hold(self):
# Test that we can hold a build request in "queue"
request_queue = queue.Queue()
event_queue = queue.Queue()
def rq_put():
request_queue.put(None)
def eq_put(br, e):
event_queue.put((br, e))
# Simulate the client side
client = HoldableExecutorApi(self.zk_client)
self.addCleanup(client.stop)
client.hold_in_queue = True
# Simulate the server side
server = ExecutorApi(self.zk_client,
build_request_callback=rq_put,
build_event_callback=eq_put)
self.addCleanup(server.stop)
# Scheduler submits request
request = BuildRequest(
"A", None, None, "job", "job_uuid", "tenant", "pipeline", '1')
client.submit(request, {})
request_queue.get(timeout=30)
# Executor receives nothing
reqs = list(server.next())
self.assertEqual(len(reqs), 0)
# Test releases hold
a = self.getRequest(client, request.uuid)
a.state = BuildRequest.REQUESTED
client.update(a)
# Executor receives request
request_queue.get(timeout=30)
reqs = list(server.next())
self.assertEqual(len(reqs), 1)
a = reqs[0]
self.assertEqual(a.uuid, 'A')
# The rest is redundant.
def test_nonexistent_lock(self):
request_queue = queue.Queue()
event_queue = queue.Queue()
def rq_put():
request_queue.put(None)
def eq_put(br, e):
event_queue.put((br, e))
# Simulate the client side
client = ExecutorApi(self.zk_client)
self.addCleanup(client.stop)
# Scheduler submits request
request = BuildRequest(
"A", None, None, "job", "job_uuid", "tenant", "pipeline", '1')
client.submit(request, {})
sched_a = self.getRequest(client, request.uuid)
# Simulate the server side
server = ExecutorApi(self.zk_client,
build_request_callback=rq_put,
build_event_callback=eq_put)
self.addCleanup(server.stop)
exec_a = self.getRequest(client, request.uuid)
client.remove(sched_a)
# Try to lock a request that was just removed
self.assertFalse(server.lock(exec_a))
def test_efficient_removal(self):
# Test that we don't try to lock a removed request
request_queue = queue.Queue()
event_queue = queue.Queue()
def rq_put():
request_queue.put(None)
def eq_put(br, e):
event_queue.put((br, e))
# Simulate the client side
client = ExecutorApi(self.zk_client)
self.addCleanup(client.stop)
# Scheduler submits two requests
request_a = BuildRequest(
"A", None, None, "job", "job_uuid", "tenant", "pipeline", '1')
client.submit(request_a, {})
request_b = BuildRequest(
"B", None, None, "job", "job_uuid", "tenant", "pipeline", '2')
client.submit(request_b, {})
sched_b = self.getRequest(client, request_b.uuid)
request_c = BuildRequest(
"C", None, None, "job", "job_uuid", "tenant", "pipeline", '3')
client.submit(request_c, {})
sched_c = self.getRequest(client, request_c.uuid)
# Simulate the server side
server = ExecutorApi(self.zk_client,
build_request_callback=rq_put,
build_event_callback=eq_put)
self.addCleanup(server.stop)
count = 0
for exec_request in server.next():
count += 1
if count == 1:
# Someone starts the second request and client removes
# the third request all while we're processing the first.
sched_b.state = sched_b.RUNNING
client.update(sched_b)
client.remove(sched_c)
for _ in iterate_timeout(30, "cache to be up-to-date"):
if (len(server.zone_queues[None].cache._cached_objects
) == 2):
break
# Make sure we only got the first request
self.assertEqual(count, 1)
def test_lost_build_requests(self):
# Test that lostBuildRequests() returns unlocked running build
# requests
executor_api = ExecutorApi(self.zk_client)
self.addCleanup(executor_api.stop)
br = BuildRequest(
"A", "zone", None, "job", "job_uuid", "tenant", "pipeline", '1')
executor_api.submit(br, {})
br = BuildRequest(
"B", None, None, "job", "job_uuid", "tenant", "pipeline", '1')
executor_api.submit(br, {})
b_uuid = br.uuid
br = BuildRequest(
"C", "zone", None, "job", "job_uuid", "tenant", "pipeline", '1')
executor_api.submit(br, {})
c_uuid = br.uuid
br = BuildRequest(
"D", "zone", None, "job", "job_uuid", "tenant", "pipeline", '1')
executor_api.submit(br, {})
d_uuid = br.uuid
br = BuildRequest(
"E", "zone", None, "job", "job_uuid", "tenant", "pipeline", '1')
executor_api.submit(br, {})
e_uuid = br.uuid
executor_api.waitForSync()
b = self.getRequest(executor_api, b_uuid)
c = self.getRequest(executor_api, c_uuid)
d = self.getRequest(executor_api, d_uuid)
e = self.getRequest(executor_api, e_uuid)
# Make sure the get() method used the correct zone keys
self.assertEqual(set(executor_api.zone_queues.keys()), {"zone", None})
b.state = BuildRequest.RUNNING
executor_api.update(b)
c.state = BuildRequest.RUNNING
executor_api.lock(c)
executor_api.update(c)
d.state = BuildRequest.COMPLETED
executor_api.update(d)
e.state = BuildRequest.PAUSED
executor_api.update(e)
# The lost_builds method should only return builds which are running or
# paused, but not locked by any executor, in this case build b and e.
lost_build_requests = list(executor_api.lostRequests())
self.assertEqual(2, len(lost_build_requests))
self.assertEqual(b.path, lost_build_requests[0].path)
def test_lost_build_request_params(self):
# Test cleaning up orphaned request parameters
executor_api = ExecutorApi(self.zk_client)
self.addCleanup(executor_api.stop)
br = BuildRequest(
"A", "zone", None, "job", "job_uuid", "tenant", "pipeline", '1')
executor_api.submit(br, {})
params_root = executor_api.zone_queues['zone'].PARAM_ROOT
self.assertEqual(len(executor_api._getAllRequestIds()), 1)
self.assertEqual(len(
self.zk_client.client.get_children(params_root)), 1)
# Delete the request but not the params
self.zk_client.client.delete(br.path)
self.assertEqual(len(executor_api._getAllRequestIds()), 0)
self.assertEqual(len(
self.zk_client.client.get_children(params_root)), 1)
# Clean up leaked params
executor_api.cleanup(0)
self.assertEqual(len(
self.zk_client.client.get_children(params_root)), 0)
def test_existing_build_request(self):
# Test that an executor sees an existing build request when
# coming online
# Test the lifecycle of a build request
request_queue = queue.Queue()
event_queue = queue.Queue()
# A callback closure for the request queue
def rq_put():
request_queue.put(None)
# and the event queue
def eq_put(br, e):
event_queue.put((br, e))
# Simulate the client side
client = ExecutorApi(self.zk_client)
self.addCleanup(client.stop)
client.submit(
BuildRequest(
"A", None, None, "job", "job_uuid",
"tenant", "pipeline", '1'), {})
# Simulate the server side
server = ExecutorApi(self.zk_client,
build_request_callback=rq_put,
build_event_callback=eq_put)
self.addCleanup(server.stop)
# Scheduler submits request
request_queue.get(timeout=30)
# Executor receives request
reqs = list(server.next())
self.assertEqual(len(reqs), 1)
a = reqs[0]
self.assertEqual(a.uuid, 'A')
def test_unlock_request(self):
# Test that locking and unlocking works
request_queue = queue.Queue()
event_queue = queue.Queue()
# A callback closure for the request queue
def rq_put():
request_queue.put(None)
# and the event queue
def eq_put(br, e):
event_queue.put((br, e))
# Simulate the client side
client = ExecutorApi(self.zk_client)
self.addCleanup(client.stop)
# Simulate the server side
server = ExecutorApi(self.zk_client,
build_request_callback=rq_put,
build_event_callback=eq_put)
self.addCleanup(server.stop)
# Scheduler submits request
request = BuildRequest(
"A", None, None, "job", "job_uuid", "tenant", "pipeline", '1')
client.submit(request, {'job': 'test'})
request_queue.get(timeout=30)
# Executor receives request
reqs = list(server.next())
self.assertEqual(len(reqs), 1)
a = reqs[0]
# Get a client copy of the request for later. Normally the
# client does not lock requests, but we will use this to
# simulate a second executor attempting to lock a request
# while our first executor operates. This ensures the lock
# contender counting works correctly.
client_a = self.getRequest(client, a.uuid)
# Executor locks request
self.assertTrue(server.lock(a, blocking=False))
# Someone else attempts to lock it
t = threading.Thread(target=client.lock, args=(client_a, True))
t.start()
# Wait for is_locked to be updated and both lock contenders to
# show in the cache:
for _ in iterate_timeout(30, "lock to propagate"):
r1 = self.getRequest(server, a.uuid)
r2 = self.getRequest(client, a.uuid)
if (r1.is_locked and
r2.is_locked and
r1.lock_contenders == 2 and
r2.lock_contenders == 2):
break
# Should see no pending requests
reqs = list(server.next())
self.assertEqual(len(reqs), 0)
reqs = list(client.next())
self.assertEqual(len(reqs), 0)
# Unlock request
server.unlock(a)
# Wait for client to get lock
t.join()
# Wait for lock_contenders to be updated:
for _ in iterate_timeout(30, "lock to propagate"):
r1 = self.getRequest(server, a.uuid)
r2 = self.getRequest(client, a.uuid)
if r1.lock_contenders == r2.lock_contenders == 1:
break
# Release client lock
client.unlock(client_a)
# Wait for is_locked to be updated:
for _ in iterate_timeout(30, "lock to propagate"):
r1 = self.getRequest(server, a.uuid)
r2 = self.getRequest(client, a.uuid)
if not r1.is_locked and not r2.is_locked:
break
# Should see pending requests
reqs = list(server.next())
self.assertEqual(len(reqs), 1)
reqs = list(client.next())
self.assertEqual(len(reqs), 1)
client.remove(a)
class TestMergerApi(ZooKeeperBaseTestCase):
def _assertEmptyRoots(self, client):
self.assertEqual(self.getZKPaths(client.REQUEST_ROOT), [])
self.assertEqual(self.getZKPaths(client.PARAM_ROOT), [])
self.assertEqual(self.getZKPaths(client.RESULT_ROOT), [])
self.assertEqual(self.getZKPaths(client.RESULT_DATA_ROOT), [])
self.assertEqual(self.getZKPaths(client.WAITER_ROOT), [])
self.assertEqual(self.getZKPaths(client.LOCK_ROOT), [])
def test_merge_request(self):
# Test the lifecycle of a merge request
request_queue = queue.Queue()
# A callback closure for the request queue
def rq_put():
request_queue.put(None)
# Simulate the client side
client = MergerApi(self.zk_client)
# Simulate the server side
server = MergerApi(self.zk_client,
merge_request_callback=rq_put)
# Scheduler submits request
payload = {'merge': 'test'}
request = MergeRequest(
uuid='A',
job_type=MergeRequest.MERGE,
build_set_uuid='AA',
tenant_name='tenant',
pipeline_name='check',
event_id='1',
)
client.submit(request, payload)
request_queue.get(timeout=30)
# Merger receives request
reqs = list(server.next())
self.assertEqual(len(reqs), 1)
a = reqs[0]
self.assertEqual(a.uuid, 'A')
params = client.getParams(a)
self.assertEqual(params, payload)
client.clearParams(a)
params = client.getParams(a)
self.assertIsNone(params)
# Merger locks request
self.assertTrue(server.lock(a, blocking=False))
a.state = MergeRequest.RUNNING
server.update(a)
self.getRequest(client, a.uuid, MergeRequest.RUNNING)
# Merger should see no pending requests
reqs = list(server.next())
self.assertEqual(len(reqs), 0)
# Merger removes and unlocks merge request on completion
server.remove(a)
server.unlock(a)
self._assertEmptyRoots(client)
def test_merge_request_hold(self):
# Test that we can hold a merge request in "queue"
request_queue = queue.Queue()
def rq_put():
request_queue.put(None)
# Simulate the client side
client = HoldableMergerApi(self.zk_client)
client.hold_in_queue = True
# Simulate the server side
server = MergerApi(self.zk_client,
merge_request_callback=rq_put)
# Scheduler submits request
payload = {'merge': 'test'}
client.submit(MergeRequest(
uuid='A',
job_type=MergeRequest.MERGE,
build_set_uuid='AA',
tenant_name='tenant',
pipeline_name='check',
event_id='1',
), payload)
request_queue.get(timeout=30)
# Merger receives nothing
reqs = list(server.next())
self.assertEqual(len(reqs), 0)
# Test releases hold
# We have to get a new merge_request object to update it.
a = self.getRequest(client, 'A')
self.assertEqual(a.uuid, 'A')
a.state = MergeRequest.REQUESTED
client.update(a)
# Merger receives request
request_queue.get(timeout=30)
reqs = list(server.next())
self.assertEqual(len(reqs), 1)
a = reqs[0]
self.assertEqual(a.uuid, 'A')
server.remove(a)
# The rest is redundant.
self._assertEmptyRoots(client)
def test_merge_request_result(self):
# Test the lifecycle of a merge request
request_queue = queue.Queue()
# A callback closure for the request queue
def rq_put():
request_queue.put(None)
# Simulate the client side
client = MergerApi(self.zk_client)
# Simulate the server side
server = MergerApi(self.zk_client,
merge_request_callback=rq_put)
# Scheduler submits request
payload = {'merge': 'test'}
future = client.submit(MergeRequest(
uuid='A',
job_type=MergeRequest.MERGE,
build_set_uuid='AA',
tenant_name='tenant',
pipeline_name='check',
event_id='1',
), payload, needs_result=True)
request_queue.get(timeout=30)
# Merger receives request
reqs = list(server.next())
self.assertEqual(len(reqs), 1)
a = reqs[0]
self.assertEqual(a.uuid, 'A')
# Merger locks request
self.assertTrue(server.lock(a, blocking=False))
a.state = MergeRequest.RUNNING
server.update(a)
self.getRequest(client, a.uuid, MergeRequest.RUNNING)
# Merger reports result
result_data = {'result': 'ok'}
server.reportResult(a, result_data)
self.assertEqual(set(self.getZKPaths(client.RESULT_ROOT)),
set(['/zuul/merger/results/A']))
self.assertEqual(set(self.getZKPaths(client.RESULT_DATA_ROOT)),
set(['/zuul/merger/result-data/A',
'/zuul/merger/result-data/A/0000000000']))
self.assertEqual(self.getZKPaths(client.WAITER_ROOT),
['/zuul/merger/waiters/A'])
# Merger removes and unlocks merge request on completion
server.remove(a)
server.unlock(a)
# Scheduler awaits result
self.assertTrue(future.wait())
self.assertEqual(future.data, result_data)
self._assertEmptyRoots(client)
def test_lost_merge_request_params(self):
# Test cleaning up orphaned request parameters
merger_api = MergerApi(self.zk_client)
# Scheduler submits request
payload = {'merge': 'test'}
merger_api.submit(MergeRequest(
uuid='A',
job_type=MergeRequest.MERGE,
build_set_uuid='AA',
tenant_name='tenant',
pipeline_name='check',
event_id='1',
), payload)
path_a = '/'.join([merger_api.REQUEST_ROOT, 'A'])
params_root = merger_api.PARAM_ROOT
self.assertEqual(len(merger_api._getAllRequestIds()), 1)
self.assertEqual(len(
self.zk_client.client.get_children(params_root)), 1)
# Delete the request but not the params
self.zk_client.client.delete(path_a)
self.assertEqual(len(merger_api._getAllRequestIds()), 0)
self.assertEqual(len(
self.zk_client.client.get_children(params_root)), 1)
# Clean up leaked params
merger_api.cleanup(0)
self.assertEqual(len(
self.zk_client.client.get_children(params_root)), 0)
self._assertEmptyRoots(merger_api)
def test_lost_merge_request_result(self):
# Test that we can clean up orphaned merge results
request_queue = queue.Queue()
# A callback closure for the request queue
def rq_put():
request_queue.put(None)
# Simulate the client side
client = MergerApi(self.zk_client)
# Simulate the server side
server = MergerApi(self.zk_client,
merge_request_callback=rq_put)
# Scheduler submits request
payload = {'merge': 'test'}
future = client.submit(MergeRequest(
uuid='A',
job_type=MergeRequest.MERGE,
build_set_uuid='AA',
tenant_name='tenant',
pipeline_name='check',
event_id='1',
), payload, needs_result=True)
request_queue.get(timeout=30)
# Merger receives request
reqs = list(server.next())
self.assertEqual(len(reqs), 1)
a = reqs[0]
self.assertEqual(a.uuid, 'A')
# Merger locks request
self.assertTrue(server.lock(a, blocking=False))
a.state = MergeRequest.RUNNING
server.update(a)
a = self.getRequest(client, a.uuid, MergeRequest.RUNNING)
# Merger reports result
result_data = {'result': 'ok'}
server.reportResult(a, result_data)
# Merger removes and unlocks merge request on completion
server.remove(a)
server.unlock(a)
self.assertEqual(set(self.getZKPaths(client.RESULT_ROOT)),
set(['/zuul/merger/results/A']))
self.assertEqual(set(self.getZKPaths(client.RESULT_DATA_ROOT)),
set(['/zuul/merger/result-data/A',
'/zuul/merger/result-data/A/0000000000']))
self.assertEqual(self.getZKPaths(client.WAITER_ROOT),
['/zuul/merger/waiters/A'])
# Scheduler "disconnects"
self.zk_client.client.delete(future._waiter_path)
# Find orphaned results
client.cleanup(age=0)
self._assertEmptyRoots(client)
def test_nonexistent_lock(self):
request_queue = queue.Queue()
def rq_put():
request_queue.put(None)
# Simulate the client side
client = MergerApi(self.zk_client)
# Scheduler submits request
payload = {'merge': 'test'}
client.submit(MergeRequest(
uuid='A',
job_type=MergeRequest.MERGE,
build_set_uuid='AA',
tenant_name='tenant',
pipeline_name='check',
event_id='1',
), payload)
client_a = self.getRequest(client, 'A')
# Simulate the server side
server = MergerApi(self.zk_client,
merge_request_callback=rq_put)
server_a = list(server.next())[0]
client.remove(client_a)
# Try to lock a request that was just removed
self.assertFalse(server.lock(server_a))
self._assertEmptyRoots(client)
def test_efficient_removal(self):
# Test that we don't try to lock a removed request
request_queue = queue.Queue()
event_queue = queue.Queue()
def rq_put():
request_queue.put(None)
def eq_put(br, e):
event_queue.put((br, e))
# Simulate the client side
client = MergerApi(self.zk_client)
# Scheduler submits three requests
payload = {'merge': 'test'}
client.submit(MergeRequest(
uuid='A',
job_type=MergeRequest.MERGE,
build_set_uuid='AA',
tenant_name='tenant',
pipeline_name='check',
event_id='1',
), payload)
client.submit(MergeRequest(
uuid='B',
job_type=MergeRequest.MERGE,
build_set_uuid='BB',
tenant_name='tenant',
pipeline_name='check',
event_id='2',
), payload)
client_b = self.getRequest(client, 'B')
client.submit(MergeRequest(
uuid='C',
job_type=MergeRequest.MERGE,
build_set_uuid='CC',
tenant_name='tenant',
pipeline_name='check',
event_id='2',
), payload)
client_c = self.getRequest(client, 'C')
# Simulate the server side
server = MergerApi(self.zk_client,
merge_request_callback=rq_put)
count = 0
for merge_request in server.next():
count += 1
if count == 1:
# Someone starts the second request and client removes
# the third request all while we're processing the first.
client_b.state = client_b.RUNNING
client.update(client_b)
client.remove(client_c)
# Make sure we only got the first request
self.assertEqual(count, 1)
def test_leaked_lock(self):
client = MergerApi(self.zk_client)
# Manually create a lock with no underlying request
self.zk_client.client.create(f"{client.LOCK_ROOT}/A", b'')
client.cleanup(0)
self._assertEmptyRoots(client)
def test_lost_merge_requests(self):
# Test that lostMergeRequests() returns unlocked running merge
# requests
merger_api = MergerApi(self.zk_client)
payload = {'merge': 'test'}
merger_api.submit(MergeRequest(
uuid='A',
job_type=MergeRequest.MERGE,
build_set_uuid='AA',
tenant_name='tenant',
pipeline_name='check',
event_id='1',
), payload)
merger_api.submit(MergeRequest(
uuid='B',
job_type=MergeRequest.MERGE,
build_set_uuid='BB',
tenant_name='tenant',
pipeline_name='check',
event_id='1',
), payload)
merger_api.submit(MergeRequest(
uuid='C',
job_type=MergeRequest.MERGE,
build_set_uuid='CC',
tenant_name='tenant',
pipeline_name='check',
event_id='1',
), payload)
merger_api.submit(MergeRequest(
uuid='D',
job_type=MergeRequest.MERGE,
build_set_uuid='DD',
tenant_name='tenant',
pipeline_name='check',
event_id='1',
), payload)
b = self.getRequest(merger_api, 'B')
c = self.getRequest(merger_api, 'C')
d = self.getRequest(merger_api, 'D')
b.state = MergeRequest.RUNNING
merger_api.update(b)
merger_api.lock(c)
c.state = MergeRequest.RUNNING
merger_api.update(c)
d.state = MergeRequest.COMPLETED
merger_api.update(d)
# The lost_merges method should only return merges which are running
# but not locked by any merger, in this case merge b
lost_merge_requests = list(merger_api.lostRequests())
self.assertEqual(1, len(lost_merge_requests))
self.assertEqual(b.path, lost_merge_requests[0].path)
# This test does not clean them up, so we can't assert empty roots
def test_existing_merge_request(self):
# Test that a merger sees an existing merge request when
# coming online
# Test the lifecycle of a merge request
request_queue = queue.Queue()
# A callback closure for the request queue
def rq_put():
request_queue.put(None)
# Simulate the client side
client = MergerApi(self.zk_client)
payload = {'merge': 'test'}
client.submit(MergeRequest(
uuid='A',
job_type=MergeRequest.MERGE,
build_set_uuid='AA',
tenant_name='tenant',
pipeline_name='check',
event_id='1',
), payload)
# Simulate the server side
server = MergerApi(self.zk_client,
merge_request_callback=rq_put)
# Scheduler submits request
request_queue.get(timeout=30)
# Merger receives request
reqs = list(server.next())
self.assertEqual(len(reqs), 1)
a = reqs[0]
self.assertEqual(a.uuid, 'A')
client.remove(a)
self._assertEmptyRoots(client)
class TestLocks(ZooKeeperBaseTestCase):
def test_locking_ctx(self):
lock = self.zk_client.client.Lock("/lock")
with locked(lock) as ctx_lock:
self.assertIs(lock, ctx_lock)
self.assertTrue(lock.is_acquired)
self.assertFalse(lock.is_acquired)
def test_already_locked_ctx(self):
lock = self.zk_client.client.Lock("/lock")
other_lock = self.zk_client.client.Lock("/lock")
other_lock.acquire()
with testtools.ExpectedException(
LockException, "Failed to acquire lock .*"
):
with locked(lock, blocking=False):
pass
self.assertFalse(lock.is_acquired)
def test_unlock_exception(self):
lock = self.zk_client.client.Lock("/lock")
with testtools.ExpectedException(RuntimeError):
with locked(lock):
self.assertTrue(lock.is_acquired)
raise RuntimeError
self.assertFalse(lock.is_acquired)
class TestLayoutStore(ZooKeeperBaseTestCase):
def test_layout_state(self):
store = LayoutStateStore(self.zk_client, lambda: None)
layout_uuid = uuid.uuid4().hex
branch_cache_min_ltimes = {
"gerrit": 123,
"github": 456,
}
state = LayoutState("tenant", "hostname", 0, layout_uuid,
branch_cache_min_ltimes, -1)
store["tenant"] = state
self.assertEqual(state, store["tenant"])
self.assertNotEqual(state.ltime, -1)
self.assertNotEqual(store["tenant"].ltime, -1)
self.assertEqual(store["tenant"].branch_cache_min_ltimes,
branch_cache_min_ltimes)
def test_ordering(self):
layout_uuid = uuid.uuid4().hex
state_one = LayoutState("tenant", "hostname", 1, layout_uuid,
{}, -1, ltime=1)
state_two = LayoutState("tenant", "hostname", 2, layout_uuid,
{}, -1, ltime=2)
self.assertGreater(state_two, state_one)
def test_cleanup(self):
store = LayoutStateStore(self.zk_client, lambda: None)
min_ltimes = defaultdict(lambda x: -1)
min_ltimes['foo'] = 1
state_one = LayoutState("tenant", "hostname", 1, uuid.uuid4().hex,
{}, -1, ltime=1)
state_two = LayoutState("tenant", "hostname", 2, uuid.uuid4().hex,
{}, -1, ltime=2)
store.setMinLtimes(state_one, min_ltimes)
store.setMinLtimes(state_two, min_ltimes)
store['tenant'] = state_one
# Run with the default delay of 5 minutes; nothing should be deleted.
store.cleanup()
self.assertEqual(store.get('tenant'), state_one)
self.assertIsNotNone(
self.zk_client.client.exists(
f'/zuul/layout-data/{state_one.uuid}'))
self.assertIsNotNone(
self.zk_client.client.exists(
f'/zuul/layout-data/{state_two.uuid}'))
# Run again with immediate deletion
store.cleanup(delay=0)
self.assertEqual(store.get('tenant'), state_one)
self.assertIsNotNone(
self.zk_client.client.exists(
f'/zuul/layout-data/{state_one.uuid}'))
self.assertIsNone(
self.zk_client.client.exists(
f'/zuul/layout-data/{state_two.uuid}'))
class TestSystemConfigCache(ZooKeeperBaseTestCase):
def setUp(self):
super().setUp()
self.config_cache = SystemConfigCache(self.zk_client, lambda: None)
def test_set_get(self):
uac = model.UnparsedAbideConfig()
uac.tenants = {"foo": "bar"}
uac.authz_rules = ["bar", "foo"]
attrs = model.SystemAttributes.fromDict({
"use_relative_priority": True,
"max_hold_expiration": 7200,
"default_hold_expiration": 3600,
"default_ansible_version": "X",
"web_root": "/web/root",
"websocket_url": "/web/socket",
"oidc_signing_key_rotation_interval": 3600,
"oidc_supported_signing_algorithms": ["RS256", "ES256"],
"oidc_default_signing_algorithm": "RS256",
})
self.config_cache.set(uac, attrs)
uac_cached, cached_attrs = self.config_cache.get()
self.assertEqual(uac.uuid, uac_cached.uuid)
self.assertEqual(uac.tenants, uac_cached.tenants)
self.assertEqual(uac.authz_rules, uac_cached.authz_rules)
self.assertEqual(attrs, cached_attrs)
def test_cache_empty(self):
with testtools.ExpectedException(RuntimeError):
self.config_cache.get()
def test_ltime(self):
uac = model.UnparsedAbideConfig()
attrs = model.SystemAttributes()
self.assertEqual(self.config_cache.ltime, -1)
self.config_cache.set(uac, attrs)
self.assertGreater(self.config_cache.ltime, -1)
self.assertEqual(uac.ltime, self.config_cache.ltime)
old_ltime = self.config_cache.ltime
self.config_cache.set(uac, attrs)
self.assertGreater(self.config_cache.ltime, old_ltime)
self.assertEqual(uac.ltime, self.config_cache.ltime)
cache_uac, _ = self.config_cache.get()
self.assertEqual(uac.ltime, cache_uac.ltime)
def test_valid(self):
uac = model.UnparsedAbideConfig()
attrs = model.SystemAttributes()
self.assertFalse(self.config_cache.is_valid)
self.config_cache.set(uac, attrs)
self.assertTrue(self.config_cache.is_valid)
class DummyChange:
def __init__(self, project, data=None):
self.uid = uuid.uuid4().hex
self.project = project
self.cache_stat = None
if data is not None:
self.deserialize(data)
@property
def cache_version(self):
return -1 if self.cache_stat is None else self.cache_stat.version
def serialize(self):
d = self.__dict__.copy()
d.pop('cache_stat')
return d
def deserialize(self, data):
self.__dict__.update(data)
def getRelatedChanges(self, sched, relevant):
return
class DummyChangeCache(AbstractChangeCache):
CHANGE_TYPE_MAP = {
"DummyChange": DummyChange,
}
class DummySource:
def getProject(self, project_name):
return project_name
def getChange(self, change_key):
return DummyChange('project')
class DummyConnections:
def getSource(self, name):
return DummySource()
class DummyScheduler:
def __init__(self):
self.connections = DummyConnections()
class DummyConnection:
def __init__(self):
self.connection_name = "DummyConnection"
self.source = DummySource()
self.sched = DummyScheduler()
class TestChangeCache(ZooKeeperBaseTestCase):
def setUp(self):
super().setUp()
self.cache = DummyChangeCache(self.zk_client, DummyConnection())
self.addCleanup(self.cache.stop)
def test_insert(self):
change_foo = DummyChange("project", {"foo": "bar"})
change_bar = DummyChange("project", {"bar": "foo"})
key_foo = ChangeKey('conn', 'project', 'change', 'foo', '1')
key_bar = ChangeKey('conn', 'project', 'change', 'bar', '1')
self.cache.set(key_foo, change_foo)
self.cache.set(key_bar, change_bar)
self.assertEqual(self.cache.get(key_foo), change_foo)
self.assertEqual(self.cache.get(key_bar), change_bar)
compressed_size, uncompressed_size = self.cache.estimateDataSize()
self.assertTrue(compressed_size != uncompressed_size != 0)
def test_update(self):
change = DummyChange("project", {"foo": "bar"})
key = ChangeKey('conn', 'project', 'change', 'foo', '1')
self.cache.set(key, change)
change.number = 123
self.cache.set(key, change, change.cache_version)
# The change instance must stay the same
updated_change = self.cache.get(key)
self.assertIs(change, updated_change)
self.assertEqual(change.number, 123)
def test_delete(self):
change = DummyChange("project", {"foo": "bar"})
key = ChangeKey('conn', 'project', 'change', 'foo', '1')
self.cache.set(key, change)
self.cache.delete(key)
self.assertIsNone(self.cache.get(key))
# Deleting an non-existent key should not raise an exception
invalid_key = ChangeKey('conn', 'project', 'change', 'invalid', '1')
self.cache.delete(invalid_key)
def test_concurrent_delete(self):
change = DummyChange("project", {"foo": "bar"})
key = ChangeKey('conn', 'project', 'change', 'foo', '1')
self.cache.set(key, change)
old_version = change.cache_version
# Simulate someone updating the cache after we decided to
# delete the entry
self.cache.set(key, change, old_version)
self.assertNotEqual(old_version, change.cache_version)
self.cache.delete(key, old_version)
# The change should still be in the cache
self.assertIsNotNone(self.cache.get(key))
def test_prune(self):
change1 = DummyChange("project", {"foo": "bar"})
change2 = DummyChange("project", {"foo": "baz"})
key1 = ChangeKey('conn', 'project', 'change', 'foo', '1')
key2 = ChangeKey('conn', 'project', 'change', 'foo', '2')
self.cache.set(key1, change1)
self.cache.set(key2, change2)
self.cache.prune([key1], max_age=0)
self.assertIsNotNone(self.cache.get(key1))
self.assertIsNone(self.cache.get(key2))
def test_concurrent_update(self):
change = DummyChange("project", {"foo": "bar"})
key = ChangeKey('conn', 'project', 'change', 'foo', '1')
self.cache.set(key, change)
# Attempt to update with the old change stat
with testtools.ExpectedException(ConcurrentUpdateError):
self.cache.set(key, change, change.cache_version - 1)
def test_change_update_retry(self):
change = DummyChange("project", {"foobar": 0})
key = ChangeKey('conn', 'project', 'change', 'foo', '1')
self.cache.set(key, change)
# Update the change so we have a new cache stat.
change.foobar = 1
self.cache.set(key, change, change.cache_version)
self.assertEqual(self.cache.get(key).foobar, 1)
def updater(c):
c.foobar += 1
# Change the cache stat so the change is considered outdated and we
# need to retry because of a concurrent update error.
change.cache_stat = model.CacheStat(change.cache_stat.key,
uuid.uuid4().hex,
change.cache_version - 1,
change.cache_stat.mzxid - 1,
0, 0, 0)
updated_change = self.cache.updateChangeWithRetry(
key, change, updater)
self.assertEqual(updated_change.foobar, 2)
def test_cache_sync(self):
other_cache = DummyChangeCache(self.zk_client, DummyConnection())
key = ChangeKey('conn', 'project', 'change', 'foo', '1')
change = DummyChange("project", {"foo": "bar"})
self.cache.set(key, change)
self.assertIsNotNone(other_cache.get(key))
change_other = other_cache.get(key)
change_other.number = 123
other_cache.set(key, change_other, change_other.cache_version)
for _ in iterate_timeout(10, "update to propagate"):
if getattr(change, "number", None) == 123:
break
other_cache.delete(key)
self.assertIsNone(self.cache.get(key))
def test_cache_sync_on_start(self):
key = ChangeKey('conn', 'project', 'change', 'foo', '1')
change = DummyChange("project", {"foo": "bar"})
self.cache.set(key, change)
change.number = 123
self.cache.set(key, change, change.cache_version)
other_cache = DummyChangeCache(self.zk_client, DummyConnection())
other_cache.cleanup()
other_cache.cleanup()
self.assertIsNotNone(other_cache.get(key))
def test_cleanup(self):
change = DummyChange("project", {"foo": "bar"})
key = ChangeKey('conn', 'project', 'change', 'foo', '1')
self.cache.set(key, change)
self.cache.cleanup()
self.assertEqual(len(self.cache._data_cleanup_candidates), 0)
self.assertEqual(
len(self.zk_client.client.get_children(self.cache.data_root)), 1)
change.number = 123
self.cache.set(key, change, change.cache_version)
self.cache.cleanup()
self.assertEqual(len(self.cache._data_cleanup_candidates), 1)
self.assertEqual(
len(self.zk_client.client.get_children(self.cache.data_root)), 2)
self.cache.cleanup()
self.assertEqual(len(self.cache._data_cleanup_candidates), 0)
self.assertEqual(
len(self.zk_client.client.get_children(self.cache.data_root)), 1)
def test_watch_cleanup(self):
change = DummyChange("project", {"foo": "bar"})
key = ChangeKey('conn', 'project', 'change', 'foo', '1')
self.cache.set(key, change)
for _ in iterate_timeout(10, "watch to be registered"):
if change.cache_stat.key._hash in self.cache._watched_keys:
break
self.cache.delete(key)
self.assertIsNone(self.cache.get(key))
for _ in iterate_timeout(10, "watch to be removed"):
if change.cache_stat.key._hash not in self.cache._watched_keys:
break
# No need for extra waiting here
class DummyZKObjectMixin:
_retry_interval = 0.1
def getPath(self):
return f'/zuul/pipeline/{self.name}'
def serialize(self, context):
d = {'name': self.name,
'foo': self.foo}
return json.dumps(d).encode('utf-8')
class DummyZKObject(DummyZKObjectMixin, ZKObject):
pass
class DummyNoMakepathZKObject(DummyZKObjectMixin, ZKObject):
makepath = False
class DummyShardedZKObject(DummyZKObjectMixin, ShardedZKObject):
pass
class TestZKObject(ZooKeeperBaseTestCase):
def _test_zk_object(self, zkobject_class):
stop_event = threading.Event()
self.zk_client.client.create('/zuul/pipeline', makepath=True)
# Create a new object
tenant_name = 'fake_tenant'
with tenant_write_lock(self.zk_client, tenant_name) as lock:
context = ZKContext(self.zk_client, lock, stop_event, self.log)
pipeline1 = zkobject_class.new(context,
name=tenant_name,
foo='bar')
self.assertEqual(pipeline1.foo, 'bar')
compressed_size, uncompressed_size = pipeline1.estimateDataSize()
self.assertTrue(compressed_size != uncompressed_size != 0)
# Load an object from ZK (that we don't already have)
with tenant_write_lock(self.zk_client, tenant_name) as lock:
context = ZKContext(self.zk_client, lock, stop_event, self.log)
pipeline2 = zkobject_class.fromZK(context,
'/zuul/pipeline/fake_tenant')
self.assertEqual(pipeline2.foo, 'bar')
compressed_size, uncompressed_size = pipeline2.estimateDataSize()
self.assertTrue(compressed_size != uncompressed_size != 0)
# Test that nested ZKObject sizes are summed up correctly
p1_compressed, p1_uncompressed = pipeline1.estimateDataSize()
p2_compressed, p2_uncompressed = pipeline2.estimateDataSize()
pipeline2._set(other=pipeline1)
compressed_size, uncompressed_size = pipeline2.estimateDataSize()
self.assertEqual(compressed_size, p1_compressed + p2_compressed)
self.assertEqual(uncompressed_size, p1_uncompressed + p2_uncompressed)
def get_ltime(obj):
zstat = self.zk_client.client.exists(obj.getPath())
return zstat.last_modified_transaction_id
# Update an object
with (tenant_write_lock(self.zk_client, tenant_name) as lock,
ZKContext(
self.zk_client, lock, stop_event, self.log) as context):
ltime1 = get_ltime(pipeline1)
pipeline1.updateAttributes(context, foo='qux')
self.assertEqual(pipeline1.foo, 'qux')
ltime2 = get_ltime(pipeline1)
self.assertNotEqual(ltime1, ltime2)
# This should not produce an unnecessary write
pipeline1.updateAttributes(context, foo='qux')
ltime3 = get_ltime(pipeline1)
self.assertEqual(ltime2, ltime3)
# Update an object using an active context
with (tenant_write_lock(self.zk_client, tenant_name) as lock,
ZKContext(
self.zk_client, lock, stop_event, self.log) as context):
ltime1 = get_ltime(pipeline1)
with pipeline1.activeContext(context):
pipeline1.foo = 'baz'
self.assertEqual(pipeline1.foo, 'baz')
ltime2 = get_ltime(pipeline1)
self.assertNotEqual(ltime1, ltime2)
# This should not produce an unnecessary write
with pipeline1.activeContext(context):
pipeline1.foo = 'baz'
ltime3 = get_ltime(pipeline1)
self.assertEqual(ltime2, ltime3)
# Update of object w/o active context should not work
with testtools.ExpectedException(Exception):
pipeline1.foo = 'nope'
self.assertEqual(pipeline1.foo, 'baz')
# Refresh an existing object
with (tenant_write_lock(self.zk_client, tenant_name) as lock,
ZKContext(
self.zk_client, lock, stop_event, self.log) as context):
pipeline2.refresh(context)
self.assertEqual(pipeline2.foo, 'baz')
# Delete an object
with (tenant_write_lock(self.zk_client, tenant_name) as lock,
ZKContext(self.zk_client, lock, stop_event, self.log) as context):
self.assertIsNotNone(self.zk_client.client.exists(
'/zuul/pipeline/fake_tenant'))
pipeline2.delete(context)
self.assertIsNone(self.zk_client.client.exists(
'/zuul/pipeline/fake_tenant'))
def _test_zk_object_exception(self, zkobject_class):
# Exercise the exception handling in the _save method
stop_event = threading.Event()
self.zk_client.client.create('/zuul/pipeline', makepath=True)
# Create a new object
tenant_name = 'fake_tenant'
class ZKFailsOnUpdate:
def delete(self, *args, **kw):
raise ZookeeperError()
def set(self, *args, **kw):
raise ZookeeperError()
class FailsOnce:
def __init__(self, real_client):
self.count = 0
self._real_client = real_client
def create(self, *args, **kw):
return self._real_client.create(*args, **kw)
def delete(self, *args, **kw):
self.count += 1
if self.count < 2:
raise OperationTimeoutError()
return self._real_client.delete(*args, **kw)
def set(self, *args, **kw):
self.count += 1
if self.count < 2:
raise OperationTimeoutError()
return self._real_client.set(*args, **kw)
def exists(self, *args, **kw):
return self._real_client.exists(*args, **kw)
def ensure_path(self, *args, **kw):
return self._real_client.ensure_path(*args, **kw)
# Fail an update
with (tenant_write_lock(self.zk_client, tenant_name) as lock,
ZKContext(
self.zk_client, lock, stop_event, self.log) as context):
pipeline1 = zkobject_class.new(context,
name=tenant_name,
foo='one')
self.assertEqual(pipeline1.foo, 'one')
# Simulate a fatal ZK exception
context.client = ZKFailsOnUpdate()
with testtools.ExpectedException(ZookeeperError):
pipeline1.updateAttributes(context, foo='two')
# We should still have the old attribute
self.assertEqual(pipeline1.foo, 'one')
# Any other error is retryable
context.client = FailsOnce(self.zk_client.client)
pipeline1.updateAttributes(context, foo='three')
# This time it should be updated
self.assertEqual(pipeline1.foo, 'three')
# Repeat test using an active context
context.client = ZKFailsOnUpdate()
with testtools.ExpectedException(ZookeeperError):
with pipeline1.activeContext(context):
pipeline1.foo = 'four'
self.assertEqual(pipeline1.foo, 'three')
context.client = FailsOnce(self.zk_client.client)
with pipeline1.activeContext(context):
pipeline1.foo = 'five'
self.assertEqual(pipeline1.foo, 'five')
def _test_zk_object_too_large(self, zkobject_class):
# This produces a consistent string that compresses to > 1MiB
rnd = random.Random()
rnd.seed(42)
size = 1024 * 1200
foo = ''.join(rnd.choice(string.printable) for x in range(size))
stop_event = threading.Event()
self.zk_client.client.create('/zuul/pipeline', makepath=True)
# Create a new object
tenant_name = 'fake_tenant'
with tenant_write_lock(self.zk_client, tenant_name) as lock:
context = ZKContext(self.zk_client, lock, stop_event, self.log)
with testtools.ExpectedException(Exception,
'ZK data size too large'):
pipeline1 = zkobject_class.new(context,
name=tenant_name,
foo=foo)
pipeline1 = zkobject_class.new(context,
name=tenant_name,
foo='foo')
with testtools.ExpectedException(Exception,
'ZK data size too large'):
pipeline1.updateAttributes(context, foo=foo)
# Refresh an existing object
pipeline1.refresh(context)
self.assertEqual(pipeline1.foo, 'foo')
def test_zk_object(self):
self._test_zk_object(DummyZKObject)
def test_zk_object_no_makepath(self):
with testtools.ExpectedException(NoNodeError):
with ZKContext(self.zk_client, None, None, self.log) as context:
DummyNoMakepathZKObject.new(context, name="tenant", foo="bar")
def test_sharded_zk_object(self):
self._test_zk_object(DummyShardedZKObject)
def test_zk_object_exception(self):
self._test_zk_object_exception(DummyZKObject)
def test_sharded_zk_object_exception(self):
self._test_zk_object_exception(DummyShardedZKObject)
def test_zk_object_too_large(self):
# This only makes sense for a regular zkobject
self._test_zk_object_too_large(DummyZKObject)
class TestBranchCache(ZooKeeperBaseTestCase):
def test_branch_cache_protected_then_all(self):
conn = DummyConnection()
cache = BranchCache(self.zk_client, conn, self.component_registry)
protected_flags = BranchFlag.PROTECTED
all_flags = BranchFlag.PRESENT
test_data = {
'project1': {
'all': [
BranchInfo('protected1', present=True),
BranchInfo('protected2', present=True),
BranchInfo('unprotected1', present=True),
BranchInfo('unprotected2', present=True),
],
'protected': [
BranchInfo('protected1', protected=True),
BranchInfo('protected2', protected=True),
],
},
}
# Test a protected-only query followed by all
cache.setProjectBranches('project1', protected_flags,
test_data['project1']['protected'])
self.assertEqual(
sorted([bi.name for bi in
cache.getProjectBranches('project1', protected_flags)
if bi.protected is True]),
[bi.name for bi in test_data['project1']['protected']]
)
self.assertRaises(
LookupError,
lambda: cache.getProjectBranches('project1', all_flags),
)
cache.setProjectBranches('project1', all_flags,
test_data['project1']['all'])
self.assertEqual(
sorted([bi.name for bi in
cache.getProjectBranches('project1', protected_flags)
if bi.protected is True]),
[bi.name for bi in test_data['project1']['protected']]
)
self.assertEqual(
sorted([bi.name for bi in
cache.getProjectBranches('project1', all_flags)]),
[bi.name for bi in test_data['project1']['all']]
)
# There's a lot of exception catching in the branch cache,
# so exercise a serialize/deserialize cycle.
ctx = ZKContext(self.zk_client, None, None, self.log)
data = cache.cache.serialize(ctx)
cache.cache.deserialize(data, ctx)
def test_branch_cache_all_then_protected(self):
conn = DummyConnection()
cache = BranchCache(self.zk_client, conn, self.component_registry)
protected_flags = BranchFlag.PROTECTED
all_flags = BranchFlag.PRESENT
test_data = {
'project1': {
'all': [
BranchInfo('protected1', present=True),
BranchInfo('protected2', present=True),
BranchInfo('unprotected1', present=True),
BranchInfo('unprotected2', present=True),
],
'protected': [
BranchInfo('protected1', protected=True),
BranchInfo('protected2', protected=True),
],
},
}
self.assertRaises(
LookupError,
lambda: cache.getProjectBranches('project1', protected_flags)
)
self.assertRaises(
LookupError,
lambda: cache.getProjectBranches('project1', all_flags)
)
# Test the other order; all followed by protected-only
cache.setProjectBranches('project1', all_flags,
test_data['project1']['all'])
self.assertRaises(
LookupError,
lambda: cache.getProjectBranches('project1', protected_flags)
)
self.assertEqual(
sorted([bi.name for bi in
cache.getProjectBranches('project1', all_flags)]),
[bi.name for bi in test_data['project1']['all']]
)
cache.setProjectBranches('project1', protected_flags,
test_data['project1']['protected'])
self.assertEqual(
sorted([bi.name for bi in
cache.getProjectBranches('project1', protected_flags)
if bi.protected is True]),
[bi.name for bi in test_data['project1']['protected']]
)
self.assertEqual(
sorted([bi.name for bi in
cache.getProjectBranches('project1', all_flags)]),
[bi.name for bi in test_data['project1']['all']]
)
# There's a lot of exception catching in the branch cache,
# so exercise a serialize/deserialize cycle.
ctx = ZKContext(self.zk_client, None, None, self.log)
data = cache.cache.serialize(ctx)
cache.cache.deserialize(data, ctx)
def test_branch_cache_change_protected(self):
conn = DummyConnection()
cache = BranchCache(self.zk_client, conn, self.component_registry)
protected_flags = BranchFlag.PROTECTED
all_flags = BranchFlag.PRESENT
data1 = {
'project1': {
'all': [
BranchInfo('newbranch', present=True),
BranchInfo('protected', present=True),
],
'protected': [
BranchInfo('protected', protected=True),
],
},
}
data2 = {
'project1': {
'all': [
BranchInfo('newbranch', present=True),
BranchInfo('protected', present=True),
],
'protected': [
BranchInfo('newbranch', present=True, protected=True),
BranchInfo('protected', protected=True),
],
},
}
# Create a new unprotected branch
cache.setProjectBranches('project1', all_flags,
data1['project1']['all'])
cache.setProjectBranches('project1', protected_flags,
data1['project1']['protected'])
self.assertEqual(
sorted([bi.name for bi in
cache.getProjectBranches('project1', protected_flags)
if bi.protected is True]),
[bi.name for bi in data1['project1']['protected']]
)
self.assertEqual(
sorted([bi.name for bi in
cache.getProjectBranches('project1', all_flags)]),
[bi.name for bi in data1['project1']['all']]
)
# Change it to protected
cache.setProtected('project1', 'newbranch', True)
self.assertEqual(
sorted([bi.name for bi in
cache.getProjectBranches('project1', protected_flags)
if bi.protected is True]),
[bi.name for bi in data2['project1']['protected']]
)
self.assertEqual(
sorted([bi.name for bi in
cache.getProjectBranches('project1', all_flags)]),
[bi.name for bi in data2['project1']['all']]
)
# Change it back
cache.setProtected('project1', 'newbranch', False)
self.assertEqual(
sorted([bi.name for bi in
cache.getProjectBranches('project1', protected_flags)
if bi.protected is True]),
[bi.name for bi in data1['project1']['protected']]
)
self.assertEqual(
sorted([bi.name for bi in
cache.getProjectBranches('project1', all_flags)]),
[bi.name for bi in data1['project1']['all']]
)
def test_branch_cache_lookup_error(self):
# Test that a missing branch cache entry results in a LookupError
conn = DummyConnection()
cache = BranchCache(self.zk_client, conn, self.component_registry)
self.assertRaises(
LookupError,
lambda: cache.getProjectBranches('project1', True)
)
self.assertIsNone(
cache.getProjectBranches('project1', True, default=None)
)
class TestConfigurationErrorList(ZooKeeperBaseTestCase):
def test_config_error_list(self):
stop_event = threading.Event()
self.zk_client.client.create('/zuul/pipeline', makepath=True)
source_context = model.SourceContext(
'cname', 'project', 'connection', 'branch', 'test')
m1 = yaml.Mark('name', 0, 0, 0, '', 0)
m2 = yaml.Mark('name', 1, 0, 0, '', 0)
start_mark = model.ZuulMark(m1, m2, 'hello')
# Create a new object
with (tenant_write_lock(self.zk_client, 'test') as lock,
ZKContext(
self.zk_client, lock, stop_event, self.log) as context):
pipeline = DummyZKObject.new(context, name="test", foo="bar")
e1 = model.ConfigurationError(
source_context, start_mark, "Test error1")
e2 = model.ConfigurationError(
source_context, start_mark, "Test error2")
with pipeline.activeContext(context):
path = '/zuul/pipeline/test/errors'
el1 = model.ConfigurationErrorList.new(
context, errors=[e1, e2], _path=path)
el2 = model.ConfigurationErrorList.fromZK(
context, path, _path=path)
self.assertEqual(el1.errors, el2.errors)
self.assertFalse(el1 is el2)
self.assertEqual(el1.errors[0], el2.errors[0])
self.assertEqual(el1.errors[0], e1)
self.assertNotEqual(e1, e2)
self.assertEqual([e1, e2], [e1, e2])
class TestBlobStore(ZooKeeperBaseTestCase):
def _assertEmptyBlobStore(self, bs, path):
with testtools.ExpectedException(KeyError):
bs.get(path)
# Make sure all keys have been cleaned up
keys = self.zk_client.client.get_children(bs.lock_root)
self.assertEqual([], keys)
def test_blob_store(self):
stop_event = threading.Event()
self.zk_client.client.create('/zuul/pipeline', makepath=True)
# Create a new object
tenant_name = 'fake_tenant'
start_ltime = self.zk_client.getCurrentLtime()
with (tenant_write_lock(self.zk_client, tenant_name) as lock,
ZKContext(
self.zk_client, lock, stop_event, self.log) as context):
bs = BlobStore(context)
with testtools.ExpectedException(KeyError):
bs.get('nope')
key = bs.put(b'something')
self.assertEqual(bs.get(key), b'something')
self.assertEqual([x for x in bs], [key])
self.assertEqual(len(bs), 1)
self.assertTrue(key in bs)
self.assertFalse('nope' in bs)
self.assertTrue(bs._checkKey(key))
self.assertFalse(bs._checkKey('nope'))
cur_ltime = self.zk_client.getCurrentLtime()
self.assertEqual(bs.getKeysLastUsedBefore(cur_ltime), {key})
self.assertEqual(bs.getKeysLastUsedBefore(start_ltime), set())
# Test deletion
bs.delete(key, cur_ltime)
self._assertEmptyBlobStore(bs, key)
# Put the blob back and test cleanup
key = bs.put(b'something')
live_blobs = set()
cur_ltime = self.zk_client.getCurrentLtime()
bs.cleanup(cur_ltime, live_blobs)
self._assertEmptyBlobStore(bs, key)
# Test leaked lock dir cleanup
self.zk_client.client.create(bs._getLockPath(key))
bs.lock_grace_period = 0
cur_ltime = self.zk_client.getCurrentLtime()
bs.cleanup(cur_ltime, live_blobs)
self._assertEmptyBlobStore(bs, key)
def test_blob_store_lock_cleanup_race(self):
# The blob store lock cleanup can have a race condition as follows:
# [1] Put a blob
# [1] Delete the blob but fail to delete the lock path for the blob
# [2] Start to put the blob it a second time
# [2] Ensure the lock path exists
# [1] Delete the leaked lock path
# [2] Fail to create the lock because the lock path does not exist
# To address this, we will retry the lock if it fails with
# NoNodeError. This test verifies that behavior.
stop_event = threading.Event()
self.zk_client.client.create('/zuul/pipeline', makepath=True)
# Create a new object
tenant_name = 'fake_tenant'
created_event = threading.Event()
deleted_event = threading.Event()
orig_ensure_path = kazoo.recipe.lock.Lock._ensure_path
def _create_blob(bs):
bs.put(b'something')
def _ensure_path(*args, **kw):
orig_ensure_path(*args, **kw)
created_event.set()
deleted_event.wait()
with (tenant_write_lock(self.zk_client, tenant_name) as lock,
ZKContext(
self.zk_client, lock, stop_event, self.log) as context):
bs = BlobStore(context)
# Get the key
key = bs.put(b'something')
cur_ltime = self.zk_client.getCurrentLtime()
bs.delete(key, cur_ltime)
# Recreate the lock dir
self.zk_client.client.create(bs._getLockPath(key))
# Block the lock method so we can delete from under it
self.useFixture(fixtures.MonkeyPatch(
'kazoo.recipe.lock.Lock._ensure_path',
_ensure_path))
# Start recreating the blob
thread = threading.Thread(target=_create_blob, args=(bs,))
thread.start()
created_event.wait()
# Run the cleanup
live_blobs = set()
bs.lock_grace_period = 0
cur_ltime = self.zk_client.getCurrentLtime()
bs.cleanup(cur_ltime, live_blobs)
self._assertEmptyBlobStore(bs, key)
# Finish recreating the blob
deleted_event.set()
thread.join()
# Ensure the blob exists
self.assertEqual(bs.get(key), b'something')
class TestPipelineInit(ZooKeeperBaseTestCase):
# Test the initialize-on-refresh code paths of various pipeline objects
def test_pipeline_state_new_object(self):
# Test the initialize-on-refresh code path with no existing object
tenant = model.Tenant('tenant')
pipeline = model.Pipeline('gate')
layout = model.Layout(tenant)
tenant.layout = layout
manager = mock.Mock()
manager.pipeline = pipeline
manager.tenant = tenant
manager.state = model.PipelineState.create(
manager, None)
context = ZKContext(self.zk_client, None, None, self.log)
manager.state.refresh(context)
self.assertTrue(self.zk_client.client.exists(manager.state.getPath()))
self.assertEqual(manager.state.layout_uuid, layout.uuid)
def test_pipeline_state_existing_object(self):
# Test the initialize-on-refresh code path with a pre-existing object
tenant = model.Tenant('tenant')
pipeline = model.Pipeline('gate')
layout = model.Layout(tenant)
tenant.layout = layout
manager = mock.Mock()
manager.pipeline = pipeline
manager.tenant = tenant
manager.state = model.PipelineState.create(
manager, None)
manager.change_list = model.PipelineChangeList.create(
manager)
context = ZKContext(self.zk_client, None, None, self.log)
# We refresh the change list here purely for the side effect
# of creating the pipeline state object with no data (the list
# is a subpath of the state object).
manager.change_list.refresh(context)
manager.state.refresh(context)
self.assertTrue(
self.zk_client.client.exists(manager.change_list.getPath()))
self.assertTrue(self.zk_client.client.exists(manager.state.getPath()))
self.assertEqual(manager.state.layout_uuid, layout.uuid)
def test_pipeline_change_list_new_object(self):
# Test the initialize-on-refresh code path with no existing object
tenant = model.Tenant('tenant')
pipeline = model.Pipeline('gate')
layout = model.Layout(tenant)
tenant.layout = layout
manager = mock.Mock()
manager.pipeline = pipeline
manager.tenant = tenant
manager.state = model.PipelineState.create(
manager, None)
manager.change_list = model.PipelineChangeList.create(
manager)
context = ZKContext(self.zk_client, None, None, self.log)
manager.change_list.refresh(context)
self.assertTrue(
self.zk_client.client.exists(manager.change_list.getPath()))
manager.state.refresh(context)
self.assertEqual(manager.state.layout_uuid, layout.uuid)
def test_pipeline_change_list_new_object_without_lock(self):
# Test the initialize-on-refresh code path if we don't have
# the lock. This should fail.
tenant = model.Tenant('tenant')
pipeline = model.Pipeline('gate')
layout = model.Layout(tenant)
tenant.layout = layout
manager = mock.Mock()
manager.pipeline = pipeline
manager.tenant = tenant
manager.state = model.PipelineState.create(
manager, None)
manager.change_list = model.PipelineChangeList.create(
manager)
context = ZKContext(self.zk_client, None, None, self.log)
with testtools.ExpectedException(NoNodeError):
manager.change_list.refresh(context, allow_init=False)
self.assertIsNone(
self.zk_client.client.exists(manager.change_list.getPath()))
manager.state.refresh(context)
self.assertEqual(manager.state.layout_uuid, layout.uuid)
class TestPolymorphicZKObjectMixin(ZooKeeperBaseTestCase):
def test_create(self):
class Parent(PolymorphicZKObjectMixin, ZKObject):
def __init__(self):
super().__init__()
self._set(
_path=None,
uid=uuid.uuid4().hex,
state="test"
)
def serialize(self, context):
return json.dumps({
"uid": self.uid,
"state": self.state,
}).encode("utf8")
def getPath(self):
return f"/test/{self.uid}"
class ChildA(Parent, subclass_id="child-A"):
pass
class ChildB(Parent, subclass_id="chid-B"):
pass
context = ZKContext(self.zk_client, None, None, self.log)
child_a = ChildA.new(context)
child_a_from_zk = Parent.fromZK(context, child_a.getPath())
self.assertIsInstance(child_a_from_zk, ChildA)
child_b = ChildB.new(context)
child_b_from_zk = Parent.fromZK(context, child_b.getPath())
self.assertIsInstance(child_b_from_zk, ChildB)
child_b_from_zk.updateAttributes(context, state="update-1")
child_b.refresh(context)
self.assertEqual(child_b.state, "update-1")
def test_missing_base(self):
with testtools.ExpectedException(TypeError):
class Child(PolymorphicZKObjectMixin, subclass_id="child"):
pass
def test_wrong_nesting(self):
class Parent(PolymorphicZKObjectMixin):
pass
with testtools.ExpectedException(TypeError):
class NewParent(Parent):
pass
def test_duplicate_subclass_id(self):
class Parent(PolymorphicZKObjectMixin):
pass
class ChildA(Parent, subclass_id="child-a"):
pass
with testtools.ExpectedException(ValueError):
class ChildAWrong(Parent, subclass_id="child-a"):
pass
def test_independent_hierarchies(self):
class ParentA(PolymorphicZKObjectMixin):
pass
class ChildA(ParentA, subclass_id="child-A"):
pass
class ParentB(PolymorphicZKObjectMixin):
pass
class ChildB(ParentB, subclass_id="child-B"):
pass
self.assertIn(ChildA._subclass_id, ParentA._subclasses)
self.assertIs(ParentA._subclasses[ChildA._subclass_id], ChildA)
self.assertNotIn(ChildB._subclass_id, ParentA._subclasses)
self.assertIn(ChildB._subclass_id, ParentB._subclasses)
self.assertIs(ParentB._subclasses[ChildB._subclass_id], ChildB)
self.assertNotIn(ChildA._subclass_id, ParentB._subclasses)
class TestPolymorphicZKObjectMixinSharded(ZooKeeperBaseTestCase):
def test_create(self):
class Parent(PolymorphicZKObjectMixin, ShardedZKObject):
def __init__(self):
super().__init__()
self._set(
_path=None,
uid=uuid.uuid4().hex,
state="test"
)
def serialize(self, context):
return json.dumps({
"uid": self.uid,
"state": self.state,
}).encode("utf8")
def getPath(self):
return f"/test/{self.uid}"
class ChildA(Parent, subclass_id="child-A"):
pass
class ChildB(Parent, subclass_id="chid-B"):
pass
context = ZKContext(self.zk_client, None, None, self.log)
child_a = ChildA.new(context)
child_a_from_zk = Parent.fromZK(context, child_a.getPath())
self.assertIsInstance(child_a_from_zk, ChildA)
child_b = ChildB.new(context)
child_b_from_zk = Parent.fromZK(context, child_b.getPath())
self.assertIsInstance(child_b_from_zk, ChildB)
child_b_from_zk.updateAttributes(context, state="update-1")
child_b.refresh(context)
self.assertEqual(child_b.state, "update-1")
def test_missing_base(self):
with testtools.ExpectedException(TypeError):
class Child(PolymorphicZKObjectMixin, subclass_id="child"):
pass
def test_wrong_nesting(self):
class Parent(PolymorphicZKObjectMixin):
pass
with testtools.ExpectedException(TypeError):
class NewParent(Parent):
pass
def test_duplicate_subclass_id(self):
class Parent(PolymorphicZKObjectMixin):
pass
class ChildA(Parent, subclass_id="child-a"):
pass
with testtools.ExpectedException(ValueError):
class ChildAWrong(Parent, subclass_id="child-a"):
pass
def test_independent_hierarchies(self):
class ParentA(PolymorphicZKObjectMixin):
pass
class ChildA(ParentA, subclass_id="child-A"):
pass
class ParentB(PolymorphicZKObjectMixin):
pass
class ChildB(ParentB, subclass_id="child-B"):
pass
self.assertIn(ChildA._subclass_id, ParentA._subclasses)
self.assertIs(ParentA._subclasses[ChildA._subclass_id], ChildA)
self.assertNotIn(ChildB._subclass_id, ParentA._subclasses)
self.assertIn(ChildB._subclass_id, ParentB._subclasses)
self.assertIs(ParentB._subclasses[ChildB._subclass_id], ChildB)
self.assertNotIn(ChildA._subclass_id, ParentB._subclasses)
class DummyAProviderNode(model.ProviderNode, subclass_id="dummy-A-node"):
pass
class DummyBProviderNode(model.ProviderNode, subclass_id="dummy-B-node"):
pass
class TestLauncherApi(ZooKeeperBaseTestCase):
def setUp(self):
super().setUp()
self.component_info = LauncherComponent(self.zk_client, "test")
self.component_info.state = self.component_info.RUNNING
self.component_info.register()
self.api = LauncherApi(
self.zk_client, self.component_registry, self.component_info,
lambda: None)
self.addCleanup(self.api.stop)
def test_launcher(self):
labels = ["foo", "bar"]
context = ZKContext(self.zk_client, None, None, self.log)
request = model.NodesetRequest.new(
context,
tenant_name="tenant",
pipeline_name="check",
buildset_uuid=uuid.uuid4().hex,
job_uuid=uuid.uuid4().hex,
job_name="foobar",
labels=labels,
priority=100,
request_time=time.time(),
zuul_event_id=None,
span_info=None,
)
# Wait for request to show up in the cache
for _ in iterate_timeout(10, "request to show up"):
request_list = self.api.getNodesetRequests()
if len(request_list):
break
self.assertEqual(len(request_list), 1)
for req in request_list:
request = self.api.getNodesetRequest(req.uuid)
self.assertIs(request, req)
self.assertIsNotNone(request)
self.assertEqual(labels, request.labels)
self.assertIsNotNone(request._zstat)
self.assertIsNotNone(request.getPath())
self.assertIsNotNone(request.acquireLock(context))
self.assertTrue(request.hasLock())
for _ in iterate_timeout(10, "request to be locked"):
if request.is_locked:
break
# Create provider nodes for the requested labels
for i, label in enumerate(request.labels):
# Alternate between the two dummy provider nodes classes
node_class = DummyAProviderNode if i % 2 else DummyBProviderNode
node = node_class.new(
context, request_id=request.uuid, uuid=uuid.uuid4().hex,
label=label)
# Wait for the nodes to show up in the cache
for _ in iterate_timeout(10, "nodes to show up"):
provider_nodes = self.api.getProviderNodes()
if len(provider_nodes) == 2:
break
# Accept and update the nodeset request
with request.activeContext(context):
for n in provider_nodes:
request.addProviderNode(n)
request.state = model.NodesetRequest.State.ACCEPTED
# "Fulfill" requested provider nodes
for node in self.api.getMatchingProviderNodes():
self.assertIsNotNone(node.acquireLock(context))
self.assertTrue(node.hasLock())
for _ in iterate_timeout(10, "node to be locked"):
if node.is_locked:
break
node.updateAttributes(
context,
state=model.ProviderNode.State.READY
)
node.releaseLock(context)
self.assertFalse(node.hasLock())
for _ in iterate_timeout(10, "node to be unlocked"):
if not node.is_locked:
break
# Wait for nodes to show up be ready and unlocked
for _ in iterate_timeout(10, "nodes to be ready"):
requested_nodes = [self.api.getProviderNode(ni)
for ni in request.nodes]
if len(requested_nodes) != 2:
continue
if all(n.state == model.ProviderNode.State.READY
and not n.is_locked for n in requested_nodes):
break
# Mark the request as fulfilled + unlock
request.updateAttributes(
context,
state=model.NodesetRequest.State.FULFILLED,
)
request.releaseLock(context)
self.assertFalse(request.hasLock())
for _ in iterate_timeout(10, "request to be unlocked"):
if not request.is_locked:
break
# Should be a no-op
self.api.cleanupNodes()
# Remove request and wait for it to be removed from the cache
request.delete(context)
for _ in iterate_timeout(10, "request to be removed"):
request_list = self.api.getNodesetRequests()
if not len(request_list):
break
not_request = self.api.getNodesetRequest(request.uuid)
self.assertIsNone(not_request)
# Make sure we still have the provider nodes
provider_nodes = self.api.getProviderNodes()
self.assertEqual(len(provider_nodes), 2)
# Mark nodes as used
for node in provider_nodes:
self.assertIsNotNone(node.acquireLock(context))
for _ in iterate_timeout(10, "wait for lock to show up"):
if node.is_locked:
break
node.updateAttributes(
context,
state=model.ProviderNode.State.USED,
)
node.releaseLock(context)
# Cleanup used nodes and wait for them to be removed from the cache
for _ in iterate_timeout(10, "nodes to be removed"):
self.api.cleanupNodes()
provider_nodes = self.api.getProviderNodes()
if not provider_nodes:
break
def test_nodeset_request_revision(self):
labels = ["foo", "bar"]
context = ZKContext(self.zk_client, None, None, self.log)
request = model.NodesetRequest.new(
context,
tenant_name="tenant",
pipeline_name="check",
buildset_uuid=uuid.uuid4().hex,
job_uuid=uuid.uuid4().hex,
job_name="foobar",
labels=labels,
priority=100,
request_time=time.time(),
zuul_event_id=None,
span_info=None,
_relative_priority=100,
)
self.assertEqual(100, request.relative_priority)
# Wait for request to show up in the cache
for _ in iterate_timeout(10, "request to show up"):
request_list = self.api.getNodesetRequests()
if len(request_list):
break
self.assertEqual(len(request_list), 1)
for req in request_list:
cached_request = self.api.getNodesetRequest(req.uuid)
self.assertEqual(100, request.relative_priority)
model.NodesetRequestRevision.new(
context, request=request, relative_priority=50)
for _ in iterate_timeout(10, "revision to be applied"):
if cached_request.relative_priority == 50:
break
# Relative priority should still be the initial value
self.assertEqual(100, request.relative_priority)
# Refresh request inc. revision
request.refresh(context)
self.assertEqual(50, request.relative_priority)
# Update revision
request.revise(context, relative_priority=10)
self.assertEqual(10, request.relative_priority)
def test_node_quota_cache(self):
context = ZKContext(self.zk_client, None, None, self.log)
class Dummy:
pass
provider = Dummy()
provider.canonical_name = 'provider'
quota = QuotaInformation(instances=1)
node_class = DummyAProviderNode
node1 = node_class.new(
context, request_id='1', uuid='1', quota=quota, label='foo',
provider='provider', state=node_class.State.BUILDING)
for _ in iterate_timeout(10, "cache to sync"):
used = self.api.nodes_cache.getQuota(provider)
if used.quota.get('instances') == 1:
break
node2 = node_class.new(
context, request_id='2', uuid='2', quota=quota, label='foo',
provider='provider', state=node_class.State.BUILDING)
for _ in iterate_timeout(10, "cache to sync"):
used = self.api.nodes_cache.getQuota(provider)
if used.quota.get('instances') == 2:
break
node1.delete(context)
for _ in iterate_timeout(10, "cache to sync"):
used = self.api.nodes_cache.getQuota(provider)
if used.quota.get('instances') == 1:
break
node2.delete(context)
for _ in iterate_timeout(10, "cache to sync"):
used = self.api.nodes_cache.getQuota(provider)
if used.quota.get('instances') == 0:
break
class DummyLockable(LockableZKObject):
ROOT = "/test/dummy"
DUMMY_PATH = "dummy"
LOCKS_PATH = "locks"
def __init__(self):
super().__init__()
self._set(
uuid=uuid.uuid4().hex,
is_locked=False,
lock_holder=None,
)
def serialize(self, context):
return json.dumps({
"uuid": self.uuid,
}).encode("utf8")
def getPath(self):
return f"{self.ROOT}/{self.DUMMY_PATH}/{self.uuid}"
def getLockPath(self):
return f"{self.ROOT}/{self.LOCKS_PATH}/{self.uuid}"
class TestLockableZKObjectCache(ZooKeeperBaseTestCase):
def test_is_locked_contenders(self):
cache = LockableZKObjectCache(
self.zk_client,
None,
root=DummyLockable.ROOT,
items_path=DummyLockable.DUMMY_PATH,
locks_path=DummyLockable.LOCKS_PATH,
zkobject_class=DummyLockable)
ctx = ZKContext(self.zk_client, None, None, self.log)
dummy = DummyLockable.new(ctx)
for _ in iterate_timeout(10, "cache to sync"):
if cache.getItems():
break
# Acquire lock for all items
for item in cache.getItems():
item.acquireLock(ctx)
for _ in iterate_timeout(10, "cache to sync"):
if all(d.is_locked for d in cache.getItems()):
break
# Create a dummy lock contender
contender_path = f"{dummy.getLockPath()}/deadbeef__lock__0000000001"
self.zk_client.client.create(contender_path)
# Make sure items are still considered locked
for _ in iterate_timeout(10, "cache to sync"):
if all(d.is_locked for d in cache.getItems()):
break
# Discard the pending lock and make sure items are still
# considered locked
self.zk_client.client.delete(contender_path)
for _ in iterate_timeout(10, "cache to sync"):
if all(d.is_locked for d in cache.getItems()):
break
# Release the lock and confirm items are no longer considered locked
for item in cache.getItems():
item.releaseLock(ctx)
for _ in iterate_timeout(10, "cache to sync"):
if not any(d.is_locked for d in cache.getItems()):
break
class TestRendezvousElection(ZooKeeperBaseTestCase):
def test_rendezvous_election(self):
# The strings foo and bar are not entirely arbitrary; they
# rendezvous hash with the path to produce a list of
# components in the order expected by the test.
c1 = SchedulerComponent(self.zk_client, "foo")
c1.register()
c2 = SchedulerComponent(self.zk_client, "bar")
c2.register()
e1 = RendezvousElection(
self.zk_client.client,
'/test/election/lock',
'scheduler',
c1,
)
e2 = RendezvousElection(
self.zk_client.client,
'/test/election/lock',
'scheduler',
c2,
)
self.assertEqual(0, len(e1._getScores()))
self.assertEqual(None, e1._getWinner())
self.assertFalse(e1.is_still_valid())
self.assertEqual(0, len(e2._getScores()))
self.assertEqual(None, e2._getWinner())
self.assertFalse(e2.is_still_valid())
# c2 will be the winner since it's the only one running
c2.state = c2.RUNNING
time.sleep(1)
self.assertEqual(1, len(e1._getScores()))
self.assertEqual(c2.hostname, e1._getWinner().hostname)
self.assertFalse(e1.is_still_valid())
self.assertEqual(1, len(e2._getScores()))
self.assertEqual(c2.hostname, e2._getWinner().hostname)
self.assertTrue(e2.is_still_valid())
# Once both are enabled, c1 is the winner
c1.state = c1.RUNNING
time.sleep(1)
self.assertEqual(2, len(e1._getScores()))
self.assertEqual(c1.hostname, e1._getWinner().hostname)
self.assertTrue(e1.is_still_valid())
self.assertEqual(2, len(e2._getScores()))
self.assertEqual(c1.hostname, e2._getWinner().hostname)
self.assertFalse(e2.is_still_valid())
event1 = threading.Event()
event2 = threading.Event()
def run1():
event1.set()
while e1.is_still_valid():
time.sleep(0.1)
def run2():
event2.set()
while e2.is_still_valid():
time.sleep(0.1)
t1 = threading.Thread(
target=e1.run,
args=(run1,),
)
t1.start()
t2 = threading.Thread(
target=e2.run,
args=(run2,),
)
t2.start()
# Wait for the thread to start
event1.wait()
# Second component should still be waiting
self.assertFalse(event2.is_set())
# Stop the thread
e1.cancel()
# Wait for the thread to stop
t1.join()
# The second election still should not have started yet since
# the first component is still running.
self.assertFalse(event2.is_set())
self.log.debug("Stop c1")
# Let the second election win
c1.state = c1.STOPPED
event2.wait()
self.log.debug("Stop c2")
# Stop the second election by stopping the second component
e2.running = False
c2.state = c2.STOPPED
t2.join()
def test_rendezvous_election_upgrade(self):
# Test that we can upgrade from a "regular" election to a
# rendezvous election.
# The numbers (e1, e2) mirror the test above, but "1" is the
# regular election and "2" is the rendezvous.
c2 = SchedulerComponent(self.zk_client, "bar")
c2.register()
e1 = SessionAwareElection(
self.zk_client.client,
'/test/election/lock',
)
e2 = RendezvousElection(
self.zk_client.client,
'/test/election/lock',
'scheduler',
c2,
)
self.assertEqual(0, len(e2._getScores()))
self.assertEqual(None, e2._getWinner())
self.assertFalse(e2.is_still_valid())
# c2 will be the winner since it's the only one running
c2.state = c2.RUNNING
time.sleep(1)
self.assertEqual(1, len(e2._getScores()))
self.assertEqual(c2.hostname, e2._getWinner().hostname)
self.assertTrue(e2.is_still_valid())
event1 = threading.Event()
event2 = threading.Event()
e1._test_stop = False
def run1():
event1.set()
while not e1._test_stop:
time.sleep(0.1)
def run2():
event2.set()
while e2.is_still_valid():
time.sleep(0.1)
t1 = threading.Thread(
target=e1.run,
args=(run1,),
)
t1.start()
# Wait for the thread to start
event1.wait()
t2 = threading.Thread(
target=e2.run,
args=(run2,),
)
t2.start()
time.sleep(1)
# Second component should still be waiting
self.assertFalse(event2.is_set())
# Stop the thread
self.log.debug("Stop c1")
e1._test_stop = True
# Wait for the thread to stop
t1.join()
# Wait for the second election to win
event2.wait()
self.log.debug("Stop c2")
# Stop the second election by stopping the second component
e2.running = False
c2.state = c2.STOPPED
t2.join()
View Git Blame Copy Permalink