-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathCHANGELOG
241 lines (215 loc) · 14.3 KB
/
CHANGELOG
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
Release Notes - Mesos - Version 0.13.0
--------------------------------------
* The primary feature in this release is "Slave Recovery". This enables a slave to reconnect
with old running executors after it restarts (e.g., after an upgrade).
* This release also includes a major refactor of the internal testing infrastructure.
* There are also several bug fixes and stability improvements (esp. around ZooKeeper).
** Bug
* [MESOS-77] - ExceptionTest.AbortOnFrameworkError sometimes hangs if Mesos built without optimizations
* [MESOS-201] - CppFramework test occasionally fails
* [MESOS-217] - LOST tasks are incorrectly reconciled between mesos and framework
* [MESOS-232] - Unit test CoordinatorTest.Elect triggers non-deterministic assertion failure in libprocess.
* [MESOS-276] - SIGSEV with current trunk and OpenJDK 7u3
* [MESOS-277] - Java test framework test is flaky
* [MESOS-289] - Zookeeper tests are flaky
* [MESOS-301] - Coordinator test is flaky
* [MESOS-318] - os::memory does not consider sysinfo.mem_unit
* [MESOS-321] - libprocess http::encode fails test
* [MESOS-344] - --disable-java still uses java headers during make check
* [MESOS-353] - ZooKeepet state test GetSetGet hung
* [MESOS-362] - Inconsistent slave maps in the master.
* [MESOS-365] - Slave should reject tasks before registering with the master.
* [MESOS-366] - Master check failure during load tests.
* [MESOS-368] - HTTP.Endpoints test is flaky.
* [MESOS-369] - Mesos tests spitting out error messages.
* [MESOS-379] - Zookeeper MasterDetectorExpireSlaveZKSessionNewMaster test is flaky
* [MESOS-385] - MasterTest.TaskRunning flaky on Jenkins.
* [MESOS-392] - FaultTolerance SchedulerExit test hangs
* [MESOS-393] - Forking at an unlucky time on OS X can cause the C++ library to deadlock.
* [MESOS-394] - Don't do ExecutorLauncher in forked process but exec first instead.
* [MESOS-395] - FaultToleranceTest.SchedulerFailoverFrameworkMessage test is flaky.
* [MESOS-399] - MonitorTest.WatchUnwatch failed.
* [MESOS-400] - Example Java framework test is flaky
* [MESOS-401] - SlaveRecoveryTest/0.RecoverTerminatedExecutor is flaky on OSX.
* [MESOS-402] - CoordinatorTest.TruncateNotLearnedFill test is flaky
* [MESOS-403] - CoordinatorTest.TruncateLearnedFill test is flaky
* [MESOS-405] - SlaveRecoveryTest/1.ReconnectExecutor crashes.
* [MESOS-406] - Google mock throws a segfault when invoked by TestFilter
* [MESOS-407] - Google test filter processing is incorrect for the empty string.
* [MESOS-408] - FaultToleranceTest.SlavePartitioned is flaky
* [MESOS-412] - MasterTest.ShutdownUnregisteredExecutor flaky
* [MESOS-423] - A slave asked to shutdown should not re-register with a new slave id
* [MESOS-424] - CgroupsIsolatorTest.BalloonFramework runs forever
* [MESOS-436] - FaultToleranceTest.SchedulerFailover test is flaky
* [MESOS-437] - ResourceOffersTest.ResourceOfferWithMultipleSlaves is flaky
* [MESOS-440] - Allow for headroom in the GC algorithm.
* [MESOS-441] - AllocatorZooKeeperTest/0.FrameworkReregistersFirst is flaky
* [MESOS-446] - Master should shutdown slaves that were deactivated
* [MESOS-447] - Master should send TASK_LOST updates for unknown tasks when slave reregisters
* [MESOS-453] - AllocatorZookeeper tests are using /tmp/mesos work directory
* [MESOS-454] - ResourceOffers tests are using /tmp/mesos working directory
* [MESOS-462] - Resource usage collection failure messages have '1' as the failure message.
* [MESOS-466] - Master should always send a re-registered message to framework when the scheduler driver sends a re-register
* [MESOS-467] - AllocatorTest.FrameworkExited is flaky
* [MESOS-469] - Scheduler driver should call disconnected on master failover
* [MESOS-474] - Mesos 0.10.0: make check fails on Ubuntu 12.04LTS
* [MESOS-476] - Upgrade libev to 4.15
* [MESOS-481] - Slave needs to only inform the master about non-terminal executors, for proper resource accounting
* [MESOS-482] - Status update manager should not cleanup the stream when there are pending updates, even though it received an ACK for a terminal update
* [MESOS-484] - Latest ZooKeeperState.cpp doesn't compile on Mountain Lion
* [MESOS-502] - Slave crashes when handling duplicate terminal updates
** Improvement
* [MESOS-46] - Refactor MasterTest to use fixture
* [MESOS-134] - Add Python documentation
* [MESOS-140] - Unrecognized command line args should fail the process
* [MESOS-242] - Add more tests to Dominant Share Allocator
* [MESOS-305] - Inform the frameworks / slaves about a master failover
* [MESOS-409] - Master detector code should stat nodes before attempting to create
* [MESOS-472] - Separate ResourceStatistics::cpu_time into ResourceStatistics::cpu_user_time and ResourceStatistics::cpu_system_time.
* [MESOS-493] - Expose version information in http endpoints
** New Feature
* [MESOS-110] - Mesos deploys should not restart tasks
* [MESOS-169] - Ability to for tests to catch in-flight messages that are dispatched
Release Notes - Mesos - Version 0.12.0
--------------------------------------
* This release includes bug fixes and stability improvements.
* The primary feature in this release is executor resource consumption monitoring. Slaves now monitor resource consumption of running executors and expose it over JSON and through the webui.
* This release also includes a new and improved Hadoop framework. The new port doesn't require patching Hadoop (i.e., it's a self contained Hadoop contrib) and lets you use existing schedulers (e.g., the fair scheduler or capacity scheduler)! The tradeoff, however, is that it doesn't take as much advantage of the fine-grained nature of a Mesos task (i.e., there is no longer a 1-1 mapping between a Mesos task and a map/reduce task).
** Sub-task
* [MESOS-214] - Report resources being used by executors
* [MESOS-419] - Old slave directories should not be garbage collected based on file modification time
** Bug
* [MESOS-107] - Scheduler library should not acknowledge a status update if the driver has been aborted.
* [MESOS-152] - Slave should forward status updates for unknown tasks
* [MESOS-285] - configure.macosx checks for version "10.7" but should check for 10.7 or greater
* [MESOS-307] - Web UI file download links are broken.
* [MESOS-317] - python mesos core bindings rejects framework messages with null bytes
* [MESOS-319] - Fix buggy read / write calls.
* [MESOS-325] - make clean is broken
* [MESOS-332] - Executor launcher fetches resources as slave user, instead of executor user.
* [MESOS-340] - Gperftools target always rebuilds.
* [MESOS-374] - HTTP GET requests to /statistics/snapshot.json crash the slave
* [MESOS-422] - Master leader election should be more robust to stale ephemeral nodes
* [MESOS-486] - TaskInfo should include a 'source' in order to enable getting resource monitoring statistics.
** Improvement
* [MESOS-293] - Make clean deletes checked in files.
** New Feature
* [MESOS-99] - display slave resource usage information in the slave webui
** Task
* [MESOS-274] - Unicode / Binary files over http endpoints.
* [MESOS-324] - Monitor executor resource usage.
* [MESOS-331] - Add --disable-perftools to configure.
Release Notes - Mesos - Version 0.11.0
--------------------------------------
** Brainstorming
* [MESOS-357] - participate in GSoC 2013
* [MESOS-358] - do not participate in the GSoC 2013
** Bug
* [MESOS-260] - Implement a duration abstraction
* [MESOS-261] - bootstrap fails when automake version >= 1.12
* [MESOS-263] - Complete the new webui (slave, framework, executor pages)
* [MESOS-264] - Make fails on the latest ubuntu
* [MESOS-270] - Log viewing broken on mesos-local runs
* [MESOS-364] - cgroup tests fail on Ubuntu
* [MESOS-386] - AllocatorTest/0.TaskFinished has incomplete expectations.
* [MESOS-388] - Latest update breaks building on OSX
* [MESOS-404] - FilesTest.BrowseTest is flaky
* [MESOS-413] - AllocatorTest/0.TaskFinished test has bad expectations.
** Improvement
* [MESOS-252] - Web UI Improvements
* [MESOS-253] - Enable -Wall -Werror on the build
* [MESOS-254] - Improve mesos slave's garbage collection
* [MESOS-259] - Expose slave attributes in slave endpoint & sortable tables
** Task
* [MESOS-275] - HTTP endpoint for file download.
* [MESOS-279] - Impose a limit on HTTP Response size
* [MESOS-389] - Add OSX slave to the Jenkins build.
* [MESOS-491] - Add mesos-0.11.0-incubating jar to maven central
Release Notes - Mesos - Version 0.10.0
--------------------------------------
** Sub-task
* [MESOS-89] - create utilities to collect information from the proc filesystem
* [MESOS-212] - Eliminate Bottle and Python based webui.
* [MESOS-222] - Rename SimpleAllocator to DominantShareAllocator
* [MESOS-223] - Libprocess-ify Allocator
* [MESOS-224] - Write allocator tests
** Bug
* [MESOS-17] - Hadoop executors killed while tasks in COMMIT_PENDING
* [MESOS-34] - Rendering JSON needs to escape strings properly.
* [MESOS-44] - Master Detector uses the wrong ACL when auth is not required
* [MESOS-48] - Remove the failover flag from executor driver
* [MESOS-54] - Mesos ZooKeeper authentication is broken
* [MESOS-83] - Filters should be removed when more resources are available than rejected offer had
* [MESOS-145] - mesos executor holds on to fd spawned by slave after slave death, preventing slave from restarting
* [MESOS-148] - Building of included Hadoop broken
* [MESOS-164] - Crash on Mac OS X due to dlopen not being thread-safe
* [MESOS-183] - Included MPI Framework Fails to Start
* [MESOS-187] - Mesos should not pass an invalid task to a slave
* [MESOS-190] - Slave seg fault when executor exited
* [MESOS-199] - Attempting to use killtree.sh after forked pid has died is fruitless.
* [MESOS-200] - New Linux 'proc' code assuming too new a kernel.
* [MESOS-209] - A race bug in ProcessManager::spawn in libprocess.
* [MESOS-211] - Fix Slave GC tests
* [MESOS-218] - Master throws exception on removeTask() if Framework is not connected
* [MESOS-220] - Slave throws exception in librocess when master is down
* [MESOS-229] - mesos zookeeper group code fails to connect when pre-existing children of the group path are read-only
* [MESOS-233] - port should not be of type short
* [MESOS-239] - Allocator doesn't handle framework failover correctly
* [MESOS-244] - Mesos slave process is not shutting down cleanly
* [MESOS-247] - Ranges comparison has a bug
* [MESOS-248] - Python quit unexpectedly while using mesos plugin
* [MESOS-251] - DRF allocator doesn't expire filter correctly
* [MESOS-257] - Master doesn't recover resources when executor exits
* [MESOS-262] - Slave should not charge the resources required for launching a executor against the executor
* [MESOS-266] - When the master removes a slave, a shutdown should be sent.
* [MESOS-268] - Slave should force kill executors when it is shutting down
* [MESOS-278] - Master Fails to Connect to Zookeeper
* [MESOS-284] - Short-term fix for fire-walling slave shutdown and lost slave messages from the 'wrong' master
* [MESOS-286] - AllocatorTest is flaky
* [MESOS-288] - Latest trunk does not finish make check
* [MESOS-290] - Jobtracker can't get TaskTrackerInfo when the JobTracker log file is deleted
* [MESOS-299] - Master detector doesn't notify about leading master after network disconnection
* [MESOS-302] - Scheduler driver shouldn't send an ACK if the driver is aborted while sending stats update
* [MESOS-303] - mesos slave crashes during framework termination
* [MESOS-306] - Mesos-master frequently crashes
* [MESOS-310] - cgroups isolation module should not block on fetching executors
* [MESOS-339] - Release script is expecting enter, not any key
* [MESOS-382] - FaultToleranceTest.FrameworkReliableRegistration test is flaky.
* [MESOS-383] - AllocatorTest/0.FrameworkExited test is broken
* [MESOS-464] - mesos 0.10.0 fails to build on ubutu 13.04
** Improvement
* [MESOS-8] - Maintain a history of executed frameworks/tasks and show it on the web UI
* [MESOS-57] - Submit mesos.jar to Maven Central
* [MESOS-142] - Explictly set and clean test work directories
* [MESOS-149] - Garbage collection on slaves
* [MESOS-171] - Make CommandInfo 'uri' field be repeated, possibly making a URI embedded message to describe whether or not we should 'chmod +x' the resulting resource.
* [MESOS-180] - Update the Hadoop patch to list protobuf-2.4.1 as a dependency so Maven pulls it down.
* [MESOS-193] - Create a single-page javascript interface to replace the existing webui
* [MESOS-194] - Make killtree more verbose.
* [MESOS-255] - Expose files through HTTP endpoints.
* [MESOS-256] - Introduce a cluster name into the Web UI
* [MESOS-272] - Create a 'fs' namespace and migrate as appropriate from our 'os' namespace.
** New Feature
* [MESOS-86] - Expose master url to the scheduler
* [MESOS-158] - Make ExecutorInfo more rich
* [MESOS-185] - Provide a master stat indicating number of outstanding resource offers
* [MESOS-207] - A new isolation module on Linux that uses Linux control groups (cgroups) directly.
* [MESOS-208] - Add whitelist option to master
** Question
* [MESOS-258] - mesos-master / mesos-slave => error: [Errno 32] Broken pipe
* [MESOS-298] - Executor fails to start
* [MESOS-311] - ClassNotFoundException when deploying hadoop on mesos
** Task
* [MESOS-69] - Migrate to Apache wiki and de-activate github wiki
* [MESOS-80] - Add a page or section to the wiki defining project coding standards
* [MESOS-133] - Make Mesos clean of most GCC warnings
* [MESOS-398] - Add mesos-0.10.0-incubating jar to maven central
Release Notes - Mesos - Version 0.9.0
-------------------------------------
** Dependency upgrade
* [MESOS-174] - Upgrade protobuf dependency to version 2.4.1
** Improvement
* [MESOS-3] - Ask executors to shutdown when a framework goes away
* [MESOS-167] - Make API names consistent
** New Feature
* [MESOS-146] - EC2 scripts should find the latest AMI from a known URL