
This is a tool to help developers quantify changes to the ring builder. It takes a scenario (JSON file) describing the builder's basic parameters (part_power, replicas, etc.) and a number of "rounds", where each round is a set of operations to perform on the builder. For each round, the operations are applied, and then the builder is rebalanced until it reaches a steady state. The idea is that a developer observes the ring builder behaving suboptimally, writes a scenario to reproduce the behavior, modifies the ring builder to fix it, and references the scenario with the commit so that others can see that things have improved. I decided to write this after writing my fourth or fifth hacky one-off script to reproduce some bad behavior in the ring builder. Change-Id: I114242748368f142304aab90a6d99c1337bced4c
228 lines
8.7 KiB
Python
228 lines
8.7 KiB
Python
#! /usr/bin/env python
|
|
# Copyright (c) 2015 Samuel Merritt <sam@swiftstack.com>
|
|
#
|
|
# 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 json
|
|
import mock
|
|
import unittest
|
|
from StringIO import StringIO
|
|
|
|
from swift.cli.ring_builder_analyzer import parse_scenario, run_scenario
|
|
|
|
|
|
class TestRunScenario(unittest.TestCase):
|
|
def test_it_runs(self):
|
|
scenario = {
|
|
'replicas': 3, 'part_power': 8, 'random_seed': 123, 'overload': 0,
|
|
'rounds': [[['add', 'r1z2-3.4.5.6:7/sda8', 100],
|
|
['add', 'z2-3.4.5.6:7/sda9', 200]],
|
|
[['set_weight', 0, 150]],
|
|
[['remove', 1]]]}
|
|
parsed = parse_scenario(json.dumps(scenario))
|
|
|
|
fake_stdout = StringIO()
|
|
with mock.patch('sys.stdout', fake_stdout):
|
|
run_scenario(parsed)
|
|
|
|
# Just test that it produced some output as it ran; the fact that
|
|
# this doesn't crash and produces output that resembles something
|
|
# useful is good enough.
|
|
self.assertTrue('Rebalance' in fake_stdout.getvalue())
|
|
|
|
|
|
class TestParseScenario(unittest.TestCase):
|
|
def test_good(self):
|
|
scenario = {
|
|
'replicas': 3, 'part_power': 8, 'random_seed': 123, 'overload': 0,
|
|
'rounds': [[['add', 'r1z2-3.4.5.6:7/sda8', 100],
|
|
['add', 'z2-3.4.5.6:7/sda9', 200]],
|
|
[['set_weight', 0, 150]],
|
|
[['remove', 1]]]}
|
|
parsed = parse_scenario(json.dumps(scenario))
|
|
|
|
self.assertEqual(parsed['replicas'], 3)
|
|
self.assertEqual(parsed['part_power'], 8)
|
|
self.assertEqual(parsed['random_seed'], 123)
|
|
self.assertEqual(parsed['overload'], 0)
|
|
self.assertEqual(parsed['rounds'], [
|
|
[['add', {'device': 'sda8',
|
|
'ip': '3.4.5.6',
|
|
'meta': '',
|
|
'port': 7,
|
|
'region': 1,
|
|
'replication_ip': None,
|
|
'replication_port': None,
|
|
'weight': 100.0,
|
|
'zone': 2}],
|
|
['add', {'device': u'sda9',
|
|
'ip': u'3.4.5.6',
|
|
'meta': '',
|
|
'port': 7,
|
|
'region': 1,
|
|
'replication_ip': None,
|
|
'replication_port': None,
|
|
'weight': 200.0,
|
|
'zone': 2}]],
|
|
[['set_weight', 0, 150.0]],
|
|
[['remove', 1]]])
|
|
|
|
# The rest of this test class is just a catalog of the myriad ways that
|
|
# the input can be malformed.
|
|
def test_invalid_json(self):
|
|
self.assertRaises(ValueError, parse_scenario, "{")
|
|
|
|
def test_json_not_object(self):
|
|
self.assertRaises(ValueError, parse_scenario, "[]")
|
|
self.assertRaises(ValueError, parse_scenario, "\"stuff\"")
|
|
|
|
def test_bad_replicas(self):
|
|
working_scenario = {
|
|
'replicas': 3, 'part_power': 8, 'random_seed': 123, 'overload': 0,
|
|
'rounds': [[['add', 'r1z2-3.4.5.6:7/sda8', 100]]]}
|
|
|
|
busted = dict(working_scenario)
|
|
del busted['replicas']
|
|
self.assertRaises(ValueError, parse_scenario, json.dumps(busted))
|
|
|
|
busted = dict(working_scenario, replicas='blahblah')
|
|
self.assertRaises(ValueError, parse_scenario, json.dumps(busted))
|
|
|
|
busted = dict(working_scenario, replicas=-1)
|
|
self.assertRaises(ValueError, parse_scenario, json.dumps(busted))
|
|
|
|
def test_bad_part_power(self):
|
|
working_scenario = {
|
|
'replicas': 3, 'part_power': 8, 'random_seed': 123, 'overload': 0,
|
|
'rounds': [[['add', 'r1z2-3.4.5.6:7/sda8', 100]]]}
|
|
|
|
busted = dict(working_scenario)
|
|
del busted['part_power']
|
|
self.assertRaises(ValueError, parse_scenario, json.dumps(busted))
|
|
|
|
busted = dict(working_scenario, part_power='blahblah')
|
|
self.assertRaises(ValueError, parse_scenario, json.dumps(busted))
|
|
|
|
busted = dict(working_scenario, part_power=0)
|
|
self.assertRaises(ValueError, parse_scenario, json.dumps(busted))
|
|
|
|
busted = dict(working_scenario, part_power=33)
|
|
self.assertRaises(ValueError, parse_scenario, json.dumps(busted))
|
|
|
|
def test_bad_random_seed(self):
|
|
working_scenario = {
|
|
'replicas': 3, 'part_power': 8, 'random_seed': 123, 'overload': 0,
|
|
'rounds': [[['add', 'r1z2-3.4.5.6:7/sda8', 100]]]}
|
|
|
|
busted = dict(working_scenario)
|
|
del busted['random_seed']
|
|
self.assertRaises(ValueError, parse_scenario, json.dumps(busted))
|
|
|
|
busted = dict(working_scenario, random_seed='blahblah')
|
|
self.assertRaises(ValueError, parse_scenario, json.dumps(busted))
|
|
|
|
def test_bad_overload(self):
|
|
working_scenario = {
|
|
'replicas': 3, 'part_power': 8, 'random_seed': 123, 'overload': 0,
|
|
'rounds': [[['add', 'r1z2-3.4.5.6:7/sda8', 100]]]}
|
|
|
|
busted = dict(working_scenario)
|
|
del busted['overload']
|
|
self.assertRaises(ValueError, parse_scenario, json.dumps(busted))
|
|
|
|
busted = dict(working_scenario, overload='blahblah')
|
|
self.assertRaises(ValueError, parse_scenario, json.dumps(busted))
|
|
|
|
busted = dict(working_scenario, overload=-0.01)
|
|
self.assertRaises(ValueError, parse_scenario, json.dumps(busted))
|
|
|
|
def test_bad_rounds(self):
|
|
base = {
|
|
'replicas': 3, 'part_power': 8, 'random_seed': 123, 'overload': 0}
|
|
|
|
self.assertRaises(ValueError, parse_scenario, json.dumps(base))
|
|
|
|
busted = dict(base, rounds={})
|
|
self.assertRaises(ValueError, parse_scenario, json.dumps(busted))
|
|
|
|
busted = dict(base, rounds=[{}])
|
|
self.assertRaises(ValueError, parse_scenario, json.dumps(busted))
|
|
|
|
busted = dict(base, rounds=[[['bork']]])
|
|
self.assertRaises(ValueError, parse_scenario, json.dumps(busted))
|
|
|
|
def test_bad_add(self):
|
|
base = {
|
|
'replicas': 3, 'part_power': 8, 'random_seed': 123, 'overload': 0}
|
|
|
|
# no dev
|
|
busted = dict(base, rounds=[[['add']]])
|
|
self.assertRaises(ValueError, parse_scenario, json.dumps(busted))
|
|
|
|
# no weight
|
|
busted = dict(base, rounds=[[['add', 'r1z2-1.2.3.4:6000/d7']]])
|
|
self.assertRaises(ValueError, parse_scenario, json.dumps(busted))
|
|
|
|
# too many fields
|
|
busted = dict(base, rounds=[[['add', 'r1z2-1.2.3.4:6000/d7', 1, 2]]])
|
|
self.assertRaises(ValueError, parse_scenario, json.dumps(busted))
|
|
|
|
# can't parse
|
|
busted = dict(base, rounds=[[['add', 'not a good value', 100]]])
|
|
self.assertRaises(ValueError, parse_scenario, json.dumps(busted))
|
|
|
|
# negative weight
|
|
busted = dict(base, rounds=[[['add', 'r1z2-1.2.3.4:6000/d7', -1]]])
|
|
self.assertRaises(ValueError, parse_scenario, json.dumps(busted))
|
|
|
|
def test_bad_remove(self):
|
|
base = {
|
|
'replicas': 3, 'part_power': 8, 'random_seed': 123, 'overload': 0}
|
|
|
|
# no dev
|
|
busted = dict(base, rounds=[[['remove']]])
|
|
self.assertRaises(ValueError, parse_scenario, json.dumps(busted))
|
|
|
|
# bad dev id
|
|
busted = dict(base, rounds=[[['remove', 'not an int']]])
|
|
self.assertRaises(ValueError, parse_scenario, json.dumps(busted))
|
|
|
|
# too many fields
|
|
busted = dict(base, rounds=[[['remove', 1, 2]]])
|
|
self.assertRaises(ValueError, parse_scenario, json.dumps(busted))
|
|
|
|
def test_bad_set_weight(self):
|
|
base = {
|
|
'replicas': 3, 'part_power': 8, 'random_seed': 123, 'overload': 0}
|
|
|
|
# no dev
|
|
busted = dict(base, rounds=[[['set_weight']]])
|
|
self.assertRaises(ValueError, parse_scenario, json.dumps(busted))
|
|
|
|
# no weight
|
|
busted = dict(base, rounds=[[['set_weight', 0]]])
|
|
self.assertRaises(ValueError, parse_scenario, json.dumps(busted))
|
|
|
|
# bad dev id
|
|
busted = dict(base, rounds=[[['set_weight', 'not an int', 90]]])
|
|
self.assertRaises(ValueError, parse_scenario, json.dumps(busted))
|
|
|
|
# negative weight
|
|
busted = dict(base, rounds=[[['set_weight', 1, -1]]])
|
|
self.assertRaises(ValueError, parse_scenario, json.dumps(busted))
|
|
|
|
# bogus weight
|
|
busted = dict(base, rounds=[[['set_weight', 1, 'bogus']]])
|
|
self.assertRaises(ValueError, parse_scenario, json.dumps(busted))
|