Skip to content

HTTPS clone URL

Subversion checkout URL

You can clone with HTTPS or Subversion.

Download ZIP
Newer
Older
100644 627 lines (528 sloc) 37.387 kb
945501f @jaredmorrow Update release notes for 1.0.3
jaredmorrow authored
1 * Riak 1.0.3 Release Notes
2 ** Bugs Fixed
3 -[[https://issues.basho.com/show_bug.cgi?id=1291][bz1291 - Incorrect match of erlang:decode_packet return value in mochiweb]]
4 -[[https://issues.basho.com/show_bug.cgi?id=1306][bz1306 - get_fsm crashes when riak_object value is not a binary]]
5 ** Other Additions
6 *** riak-admin top
2020caa @jaredmorrow Finalize release notes for 1.0.3 release
jaredmorrow authored
7 =riak-admin= can now run top against the server. Top provides
945501f @jaredmorrow Update release notes for 1.0.3
jaredmorrow authored
8 information about what the erlang processes inside of Riak are doing.
2020caa @jaredmorrow Finalize release notes for 1.0.3 release
jaredmorrow authored
9 Top reports process reductions (an indicator of CPU utilization),
945501f @jaredmorrow Update release notes for 1.0.3
jaredmorrow authored
10 memory used and message queue sizes.
11
12 Start with =riak-admin top= and exit with double-^C
13
14 *** busy_port / busy_dist_port
15 Riak will now log whenever distributed erlang is overloaded by logging
16 =busy_dist_port= messages on the console and to the logs. If the
17 message is being logged frequently, the buffer used for distributed
18 erlang can be increased by adding +zdbbl KBytes in the =vm.args= file
19 (in =/etc/riak= or =/opt/riak/etc= depending on platform).
20
21 *** Bitcask Support for O_SYNC
22 Bitcask has been reworked to perform I/O directly on the Erlang
23 scheduler threads; this is the result of extensive load testing in a
24 production environment. As a result of this change, it can now support
25 =O_SYNC= on Linux.
26
55f36ac @jaredmorrow Update RELEASE-NOTES for 1.0.3
jaredmorrow authored
27 *** New Claim Algorithm
28 The new claim algorithm introduced in the 1.0 release has been set as
2020caa @jaredmorrow Finalize release notes for 1.0.3 release
jaredmorrow authored
29 the default for 1.0.3.
30
31 The new claim algorithm significantly reduces the amount of ownerhip
32 shuffling for clusters with more than N+2 nodes in them. Changes to
33 the cluster membership code in 1.0 made the original claim algorithm
34 fall back to just reordering nodes in sequence in more cases than it
35 used to, causing massive handoff activity on larger clusters. Mixed
36 clusters including pre-1.0 nodes will still use the original algorithm
37 until all nodes are upgraded.
55f36ac @jaredmorrow Update RELEASE-NOTES for 1.0.3
jaredmorrow authored
38
945501f @jaredmorrow Update release notes for 1.0.3
jaredmorrow authored
39 *** Changes to Handoff Concurrency
40 The default value for =handoff_concurrency= was set to 1 to err on the
41 side of caution and not overload the cluster with handoff traffic.
42
43 ** Known Issues
44
f9f49d8 @jaredmorrow Update release notes for 1.0.2 changes
jaredmorrow authored
45 * Riak 1.0.2 Release Notes
46 ** Bugs Fixed
47 -[[https://issues.basho.com/show_bug.cgi?id=1227][bz1227 - badstate crash in handoff]]
48 -[[https://issues.basho.com/show_bug.cgi?id=1261][bz1261 - pipe mapreduce jobs fail in mixed 1.0.0/1.0.1 clusters]]
49 -[[https://issues.basho.com/show_bug.cgi?id=1263][bz1263 - bitcask-backed vnodes handle requests while dropping]]
50 -[[https://issues.basho.com/show_bug.cgi?id=1265][bz1265 - eleveldb not opened with configuration options from environment]]
51
52 ** Other Additions
53 *** New stats
54 Two new stats categories have been added to riak stats.
55
56 -=node_get_fsm_siblings= reports the statstics on the number of siblings retrieved with get over the last minute.
57 -=node_get_fsm_objsize= reports an estimate of the object sizes retrieved with get over the last minute.
58
59 For both categories mean, median, max, 95th & 99th percentiles are given.
60 The new stats are available using riak-admin status or from the /stats endpoint.
61 ** Known Issues
62 *** Pipe based MapReduce during rolling upgrades
63 [[https://issues.basho.com/show_bug.cgi?id=1261][BZ1261]]
64 A bug was found that caused pipe-based MapReduce to be
65 unusable during rolling upgrades from 1.0.0 to 1.0.1 (they work again
66 once the upgrade is complete). The problem has been fixed such that
67 MapReduce will stay available when upgrading from 1.0.1 to 1.0.2, and
68 also from 1.0.2 onward.
69
70 If you are still using Riak 1.0.0, pipe-based MapReduce will still be
71 unusable during the rolling upgrade to 1.0.2. We apologize for the
72 inconvenience. If your MapReduce queries work on the legacy
73 subsystem, you may want to switch back to it for the duration of your
74 upgrade to 1.0.2. To switch MapReduce subsystems, either edit the
75 'mapred_system' setting in your app.config and restart each node, or
76 connect to each nodes console and evaluate
77 'application:set_env(riak_kv, mapred_system, Choice).' where 'Choice'
78 is either 'legacy' or 'pipe'.
79
80 *** Large Ring Sizes
81
82 The new cluster protocol in 1.0 leads to a practical limit on the size
83 of the ring in a Riak cluster. While there is no hard coded limit, large
84 ring sizes lead to increasingly slower performance and potentially unstable
85 clusters. This limitation will be addressed in a future major release of
86 Riak. For the time being, it is recommended that users limit clusters to
87 no more than 256, 512, 1024 partitions depending on the performance of the
88 underlying hardware.
89
90 *** Ownership Handoff Stall
91
92 It has been observed that 1.0 clusters can end up in a state with ownership
93 handoff failing to complete. This state should not happen under normal
94 circumstances, but has been observed in cases where Riak nodes crashed due
95 to unrelated issues (eg. running out of disk space/memory) during cluster
96 membership changes. The behavior can be identified by looking at the output
97 of =riak-admin ring_status= and looking for transfers that are waiting on an
98 empty set. As an example:
99
100 #+BEGIN_SRC
101 ============================== Ownership Handoff ==============================
102 Owner: riak@host1
103 Next Owner: riak@host2
104
105 Index: 123456789123456789123456789123456789123456789123
106 Waiting on: []
107 Complete: [riak_kv_vnode,riak_pipe_vnode]
108 #+END_SRC
109
110 To fix this issue, copy/paste the following code sequence into an Erlang
111 shell for each =Owner= node in =riak-admin ring_status= for which
112 this case has been identified. The Erlang shell can be accessed with =riak attach=
113
114 #+BEGIN_SRC
115 fun() ->
116 Node = node(),
117 {_Claimant, _RingReady, _Down, _MarkedDown, Changes} =
118 riak_core_status:ring_status(),
119 Stuck =
120 [{Idx, Complete} || {{Owner, NextOwner}, Transfers} <- Changes,
121 {Idx, Waiting, Complete, Status} <- Transfers,
122 Owner =:= Node,
123 Waiting =:= [],
124 Status =:= awaiting],
125
126 RT = fun(Ring, _) ->
127 NewRing = lists:foldl(fun({Idx, Mods}, Ring1) ->
128 lists:foldl(fun(Mod, Ring2) ->
129 riak_core_ring:handoff_complete(Ring2, Idx, Mod)
130 end, Ring1, Mods)
131 end, Ring, Stuck),
132 {new_ring, NewRing}
133 end,
134
135 case Stuck of
136 [] ->
137 ok;
138 _ ->
139 riak_core_ring_manager:ring_trans(RT, []),
140 ok
141 end
142 end().
143 #+END_SRC
03947ed @kellymclaughlin Add mention in release notes about innostore and stack size on 32-bit sy...
kellymclaughlin authored
144
145 *** Innostore on 32-bit systems
146
147 Depending on the partition count and the default stack size of the
148 operating system, using Innostore as the backend on 32-bit systems can
149 cause problems. A symptom of this problem would be a message similar
150 to this on the console or in the Riak error log:
151
152 #+BEGIN_SRC
153 10:22:38.719 [error] Failed to start riak_kv_innostore_backend for index 1415829711164312202009819681693899175291684651008. Reason: {eagain,[{erlang,open_port,[{spawn,innostore_drv},[binary]]},{innostore,connect,0},{riak_kv_innostore_backend,start,2},{riak_kv_vnode,init,1},{riak_core_vnode,init,1},{gen_fsm,init_it,6},{proc_lib,init_p_do_apply,3}]}
154 10:22:38.871 [notice] "backend module failed to start."
155 #+END_SRC
156
157 The workaround for this problem is to reduce the stack size for the
4f7ee5d @jaredmorrow Fix grammar typo
jaredmorrow authored
158 user Riak is run as (=riak= for most distributions). For example
03947ed @kellymclaughlin Add mention in release notes about innostore and stack size on 32-bit sy...
kellymclaughlin authored
159 on Linux, the current stack size can be viewed using =ulimit -s= and
160 can be altered by adding entries to the =/etc/security/limits.conf=
161 file such as these:
162
163 #+BEGIN_SRC
164 riak soft stack 1024
165 riak hard stack 1024
166 #+END_SRC
167
7869692 @jaredmorrow Update release notes for 1.0.1 fixes
jaredmorrow authored
168 * Riak 1.0.1 Release Notes
169 ** Bugs Fixed
bb46ca8 @jaredmorrow Update release notes with latest bug fixes
jaredmorrow authored
170 -[[https://issues.basho.com/show_bug.cgi?id=15][bz0015 - Merge js-mapreduce.org and basic-mapreduce.txt]]
7869692 @jaredmorrow Update release notes for 1.0.1 fixes
jaredmorrow authored
171 -[[https://issues.basho.com/show_bug.cgi?id=1183][bz1183 - Handoff may cause riak_kv_w_reduce to over-reduce]]
172 -[[https://issues.basho.com/show_bug.cgi?id=1184][bz1184 - riak_kv_w_reduce should respect reduce_phase_only_1 during handoff]]
173 -[[https://issues.basho.com/show_bug.cgi?id=1213][bz1213 - HTTP and Javascript map phases can't use do_prereduce]]
174 -[[https://issues.basho.com/show_bug.cgi?id=1225][bz1225 - Exact phrase search fails on repeated words]]
175 -[[https://issues.basho.com/show_bug.cgi?id=1228][bz1228 - riak_core_ring_manager crashes with function clause and state of 'ok' ]]
176 -[[https://issues.basho.com/show_bug.cgi?id=1234][bz1234 - Mixed 0.14/1.0.0 cluster puts can fail for clusters with more than N nodes]]
177 -[[https://issues.basho.com/show_bug.cgi?id=1235][bz1235 - vnode crashes due to receiving handoff data in deleted state]]
178 -[[https://issues.basho.com/show_bug.cgi?id=1236][bz1236 - riak_pipe vnode crash in Java client test suite]]
179 -[[https://issues.basho.com/show_bug.cgi?id=1237][bz1237 - File descriptor leak in riak_kv_bitcask_backend fold]]
180 -[[https://issues.basho.com/show_bug.cgi?id=1238][bz1238 - Prereduce uses map phase arg instead of reduce phase arg]]
181 -[[https://issues.basho.com/show_bug.cgi?id=1242][bz1242 - Handoff stops when the ring is modified by fix-up modules]]
182 -[[https://issues.basho.com/show_bug.cgi?id=1243][bz1243 - JSONified riak_object drops multiple index entries]]
183 -[[https://issues.basho.com/show_bug.cgi?id=1244][bz1244 - riak_kv_wm_object makes call to riak_client:get/3 with invalid type for key]]
184 -[[https://issues.basho.com/show_bug.cgi?id=1249][bz1249 - Backup/restore fail if ring is fixed-up]]
185 -[[https://issues.basho.com/show_bug.cgi?id=1250][bz1250 - bitcask partition data isn't cleaned up after handoff]]
186
dffe517 @jaredmorrow Add major sections
jaredmorrow authored
187 * Riak 1.0.0 Release Notes
188
0c7c972 @jaredmorrow Initial attempt at Org mode release-notes
jaredmorrow authored
189 ** Major Features and Improvements for Riak
190 *** 2i
dffe517 @jaredmorrow Add major sections
jaredmorrow authored
191 Secondary Indexes (2I) makes it easier to find your data in
192 Riak. With 2i, an application can index an object under an arbitrary
193 number of field/value pairs specified at write time. Later, the
194 application can query the index, pulling back a list of keys that
195 match a given value or fall within a given range of values. This
196 simple mechanism enables applications to model alternate keys,
197 one-to-many relationships, and many-to-many relationships in
198 Riak. Indexing is atomic and real-time. Querying is integrated with
199 MapReduce. In this release, the query feature set is minimal: only
200 exact match and range queries are supported, and only on a single
201 field. In addition, the system must be configured to use the
ceb1eb8 @rzezeski Markup code fragments
rzezeski authored
202 =riak_kv_eleveldb_backend=. Future releases will bring support for more
dffe517 @jaredmorrow Add major sections
jaredmorrow authored
203 complex queries and additional backends, depending upon user feedback.
0c7c972 @jaredmorrow Initial attempt at Org mode release-notes
jaredmorrow authored
204
658c6f9 @jaredmorrow More monospace highlights to be consistent
jaredmorrow authored
205
0c7c972 @jaredmorrow Initial attempt at Org mode release-notes
jaredmorrow authored
206 *** Backend Refactoring
dffe517 @jaredmorrow Add major sections
jaredmorrow authored
207 There are 3 main areas of change with regard to the backend refactoring.
66228e4 @jaredmorrow Cleanup spacing
jaredmorrow authored
208 **** Changes to the suite of available backends
209 - Added: eleveldb backend
210 - Merged: ets and cache backend have been consolidated into a single
dffe517 @jaredmorrow Add major sections
jaredmorrow authored
211 memory backend.
66228e4 @jaredmorrow Cleanup spacing
jaredmorrow authored
212 - Removed: dets, filesystem (fs), and gb_trees backends
213 **** Standardized API
214 - All of the backend modules now share a common API.
215 - Vnode code has been greatly simplified.
216 **** Asynchronous folding
217 - The vnode no longer has to be blocked by fold operations.
218 - A node for a backend capable of asynchronous folds starts a worker
219 pool to handle fold requests.
220 - Asynchronous folds are done using read-only snapshots of the data.
221 - Bitcask, eleveldb, memory, and multi backends support asynchronous folding.
ceb1eb8 @rzezeski Markup code fragments
rzezeski authored
222 - Asynchronous folding may be disabled by adding ={async_folds, false}=
658c6f9 @jaredmorrow More monospace highlights to be consistent
jaredmorrow authored
223 in the =riak_kv= section of the =app.config= file.
224
0c7c972 @jaredmorrow Initial attempt at Org mode release-notes
jaredmorrow authored
225
226 *** Lager
dffe517 @jaredmorrow Add major sections
jaredmorrow authored
227 Lager is Riak’s new logging library. It aims to alleviate some of the
228 shortcomings in Erlang/OTP’s bundled logging library; specifically
229 more readable log messages, support for external logfile rotation,
230 more granular log levels and runtime changes.
66228e4 @jaredmorrow Cleanup spacing
jaredmorrow authored
231
658c6f9 @jaredmorrow More monospace highlights to be consistent
jaredmorrow authored
232
0c7c972 @jaredmorrow Initial attempt at Org mode release-notes
jaredmorrow authored
233 *** LevelDB
dffe517 @jaredmorrow Add major sections
jaredmorrow authored
234 Riak now has support for LevelDB, a fast key-value storage library
235 written at Google that provides an ordered mapping from string keys to
236 string values. This backend provides a viable alternative to the
237 innostore backend, with better performance. LevelDB has an append-only
238 design that provides high-performance writes. Unlike Bitcask, however,
239 LevelDB keeps keys in sorted order and uses a multi-level cache to
240 keep portions of the keyspace in memory. This design allows LevelDB to
241 effectively deal with very large keyspaces, but at the cost of
242 additional seeks on read.
66228e4 @jaredmorrow Cleanup spacing
jaredmorrow authored
243
658c6f9 @jaredmorrow More monospace highlights to be consistent
jaredmorrow authored
244
0c7c972 @jaredmorrow Initial attempt at Org mode release-notes
jaredmorrow authored
245 *** Pipe
dffe517 @jaredmorrow Add major sections
jaredmorrow authored
246 Riak Pipe is the new plumbing for Riak's MapReduce system. It improves
247 upon the legacy system in many ways, including built-in backpressure
248 to keep phases from overloading others downstream, and a design shaped
249 to make much better use of the parallel-processing resources available
250 in the cluster. Riak Pipe is also a general queue-based parallel
251 processing system, which new applications may find is a more natural
252 starting point than bare Riak Core. Riak Pipe is the default plumbing
253 for MapReduce queries on new clusters. To move old clusters onto Riak
ceb1eb8 @rzezeski Markup code fragments
rzezeski authored
254 Pipe MapReduce, add ={mapred_system, pipe}= to the =riak_kv= section
658c6f9 @jaredmorrow More monospace highlights to be consistent
jaredmorrow authored
255 of your node’s =app.config=. To force Riak to use the legacy system for
ceb1eb8 @rzezeski Markup code fragments
rzezeski authored
256 MapReduce, set ={mapred_system, legacy}= in the =riak_kv= section of
257 =app.config= instead.
66228e4 @jaredmorrow Cleanup spacing
jaredmorrow authored
258
658c6f9 @jaredmorrow More monospace highlights to be consistent
jaredmorrow authored
259
0c7c972 @jaredmorrow Initial attempt at Org mode release-notes
jaredmorrow authored
260 *** URL encoding changes over REST API
dffe517 @jaredmorrow Add major sections
jaredmorrow authored
261 Riak now decode URL encoded bucket/keys over the REST API, rather than
262 prior behavior which was to decode links but not decode
263 buckets/keys. The default cluster behavior is configurable in
658c6f9 @jaredmorrow More monospace highlights to be consistent
jaredmorrow authored
264 =riak_core= section of =app.config=: ={http_url_decoding, on}= provides the
dffe517 @jaredmorrow Add major sections
jaredmorrow authored
265 new behavior (decode everything), missing/anything else provides the
266 current behavior.
267
ceb1eb8 @rzezeski Markup code fragments
rzezeski authored
268 The module =riak_kv_encoding_migrate.erl= is also provided to help
dffe517 @jaredmorrow Add major sections
jaredmorrow authored
269 migrate existing encoded buckets/keys to their decoded equivalents
66228e4 @jaredmorrow Cleanup spacing
jaredmorrow authored
270
658c6f9 @jaredmorrow More monospace highlights to be consistent
jaredmorrow authored
271
0c7c972 @jaredmorrow Initial attempt at Org mode release-notes
jaredmorrow authored
272 *** Riak Clustering Overhaul
dffe517 @jaredmorrow Add major sections
jaredmorrow authored
273 The clustering subsystem of Riak has been overhauled for 1.0, addressing a
274 variety of bugs and limitations.
275 - When reassigning a partition to a new node, the partition data from
276 the existing owner is first transferred to the new owner before
277 officially changing the ownership in the ring. This fixes a bug
278 where 404s could appear while ownership was being changed.
279 - Adding/removing nodes to a cluster is now more robust, and it is no
ceb1eb8 @rzezeski Markup code fragments
rzezeski authored
280 longer necessary to check for ring convergence (=riak-admin ringready=)
dffe517 @jaredmorrow Add major sections
jaredmorrow authored
281 between adding/removing nodes. Adding multiple nodes all at once should
282 "just work".
283 - Handoff related to changing node owners can now occur while a cluster is
284 under load, therefore allowing a Riak cluster to scale up and down during
285 load / normal operation.
286 - Various other clustering bug/fixes. See the fixed bug list for details.
287 **** Notes
ceb1eb8 @rzezeski Markup code fragments
rzezeski authored
288 - =riak-admin join= has new semantics. The command is now a one-way
dffe517 @jaredmorrow Add major sections
jaredmorrow authored
289 operation that joins a single node to cluster. The node that the
290 command is executed under should be the desired joining node, and
291 the target of the command should be a member of the desired target
292 cluster. The new command requires the joining node to be a singleton
293 (1-node) cluster.
ceb1eb8 @rzezeski Markup code fragments
rzezeski authored
294 - =riak-admin leave= is now the only safe way to remove a node from a
dffe517 @jaredmorrow Add major sections
jaredmorrow authored
295 cluster. The leave command ensures that the exiting node will
296 handoff all its partitions before leaving the cluster. It should be
297 executed by the node intended to leave.
498236e @jtuple Update clustering section of relnotes
jtuple authored
298 - =riak-admin remove= no longer exists. Use =riak-admin leave= to safely
299 remove a node from the cluster, or =riak-admin force-remove= to remove
300 an unrecoverable node.
301 - =riak-admin force-remove= immediately removes a node from the cluster
302 without having it first handoff data. All data replicas are therefore
303 lost. This is designed for cases where a node is unrecoverable
dffe517 @jaredmorrow Add major sections
jaredmorrow authored
304 - The new cluster changes require all nodes to be up and reachable in
305 order for new members to be integrated into the cluster and for the
306 data to be rebalanced. During brief node outages, the new protocol
307 will wait until all nodes are eventually back online and continue
308 automatically. If it is known that a node will be offline for an
ceb1eb8 @rzezeski Markup code fragments
rzezeski authored
309 extended period, the new =riak-admin down= command can be used to
dffe517 @jaredmorrow Add major sections
jaredmorrow authored
310 mark a node as offline and the cluster will then resume integrating
311 nodes and performing ring rebalances. Nodes marked as down will
312 automatically rejoin and reintegrate into the cluster when they come
313 back online.
498236e @jtuple Update clustering section of relnotes
jtuple authored
314 - When performing a rolling upgrade, the cluster will auto-negotiate
315 the proper gossip protocol, using the legacy gossip protocol while
316 there is a mixed-verison cluster. During the upgrade, executing
317 =riak-admin ringready= and =riak-admin transfers= from a non-1.0
318 node will fail. However, executing those commands from a 1.0 node
319 will succeed and give the desired information.
66228e4 @jaredmorrow Cleanup spacing
jaredmorrow authored
320
658c6f9 @jaredmorrow More monospace highlights to be consistent
jaredmorrow authored
321
642b2a3 @jonmeredith Updated 1.0.0 release notes with delete/restore changes.
jonmeredith authored
322 *** Get/Put Improvements
dffe517 @jaredmorrow Add major sections
jaredmorrow authored
323 The way that Riak versions and updates objects has been
324 overhauled. ClientIds are no longer used when updating objects, the
325 server handles all versioning using a vector clock id per-vnode.
326
327 New clusters are configured with the new vclock behavior turned on.
328 If you are performing a rolling upgrade of an existing cluster, once
658c6f9 @jaredmorrow More monospace highlights to be consistent
jaredmorrow authored
329 all nodes have been upgraded the =app.config= needs to be updated to add
ceb1eb8 @rzezeski Markup code fragments
rzezeski authored
330 ={vnode_vclocks, true}=.
dffe517 @jaredmorrow Add major sections
jaredmorrow authored
331
332 Puts are now coordinated in the same way as on the original Dynamo
333 system. Requests must be handled by a node in the preference list
334 (primary or fallback) for that bucket/key. Nodes will automatically
335 forward to a valid node when necessary and increment the coord_redirs
336 stats counter. The put is initially written to the local vnode before
337 forwarding to the remote vnodes. This ensures that the updated vclock
338 for the riak object will replace the existing value or create siblings
339 in partitioning/failure scenario where the same client can see both
340 sides.
341
342 Error proofing for the failure scenarios has made it so that clients
ceb1eb8 @rzezeski Markup code fragments
rzezeski authored
343 no longer have to be well behaved. If =allow_mult= is set =true=, every
dffe517 @jaredmorrow Add major sections
jaredmorrow authored
344 time you create a new object and put over an existing one it will
345 create a sibling. Vector clocks should now be much smaller in size as
346 only a few vclock ids are now updated. This should resolve a number of
347 issues due to vclock pruning causing siblings.
348
349 Gets have been changed to return more information during failure.
350 Prior to 1.0 there were cases where Riak returned not found if not
351 enough valid responses were returned. The case of not enough responses
352 has been changed to an error instead reported as 503 over HTTP or as
ceb1eb8 @rzezeski Markup code fragments
rzezeski authored
353 ={error, {r_val_unsatistfied, R, NumResponses}}= for Erlang/PBC clients.
dffe517 @jaredmorrow Add major sections
jaredmorrow authored
354
355 New options have been added to the get requests for handling
356 notfound. Prior to 1.0 only successful reads were counted towards R
357 and there was some logic to try and fail early rather than wait until
358 the request timed out if not enough replies were received
359 (basic_quorum). This meant when a node went down and another node
360 didn’t response you would get a not found response that triggered a
361 read repair and then if you retrieved the object again it would be
362 present.
363
364 Now that other enhancements have been made (delete and asynchronous
365 improvements to the vnodes) we can change notfounds to be counted
366 towards R and disable the basic_quorum logic by setting bucket
ceb1eb8 @rzezeski Markup code fragments
rzezeski authored
367 properties to =[{notfound_ok, true}, {basic_quorum, false}]= and reduce
dffe517 @jaredmorrow Add major sections
jaredmorrow authored
368 the number of cases where notfound is returned on first request when
369 an object could be.
0c7c972 @jaredmorrow Initial attempt at Org mode release-notes
jaredmorrow authored
370
642b2a3 @jonmeredith Updated 1.0.0 release notes with delete/restore changes.
jonmeredith authored
371 *** Delete Changes
372
373 The changes to the vector clocks make it very important that the tombstones
374 written by deletes are removed from all vnodes. In 0.14.2 the tombstone
375 was removed as it was confirmed that all vnodes had the tombstone stored.
376 For 1.0.0 this has been changed to delay the removal by a short period of
377 time (default 3s) and is aborted if the object is updated. The behavior
378 is configurable by setting {delete_mode, Mode} in the riak_kv secion of
379 app.config and can be set to the following
380
381 keep - tombstones are kept forever
382
383 immediate - tombstones are removed without delay - 0.14.2 behavior.
384
18b0e82 @beerriot fix some release-notes typos
beerriot authored
385 NNNN - delay in milliseconds to check for changes before removing tombstone.
642b2a3 @jonmeredith Updated 1.0.0 release notes with delete/restore changes.
jonmeredith authored
386 The default is 3000 for 3s.
387
388 The riak_client, HTTP and PBC interfaces have been modified to return vclock
389 information for deleted objects. riak_client:get accepts a deletedvclock
390 option which changes a deleted object read from {error, notfound} to
391 {error, {deleted, VClock}}.
392
393 The HTTP interface returns X-Riak-Vclock with 404s now. The PBC interface
394 can request the vclock setting the deletedvclock option on get requests.
395
396 Clients that have not been updated to take advantage of the new information
397 may create siblings with tombstones if they create a new object over one
398 deleted recently enough the tombstone still exists.
399
400 *** Backup/Restore Changes
401
402 Restore has been changed to restore the exact objects that were backed up.
403 This means that if they have been updated since the backup, or deleted
404 recently enough that the tombstone has not been removed, then the backed
405 up object will not be restored. Waiting until the tombstones are removed
18b0e82 @beerriot fix some release-notes typos
beerriot authored
406 should enable the objects to be restored (however if delete_mode=keep
407 tombstones are never removed).
642b2a3 @jonmeredith Updated 1.0.0 release notes with delete/restore changes.
jonmeredith authored
408
18b0e82 @beerriot fix some release-notes typos
beerriot authored
409 In 0.14.2 restoring an object would have updated the vclock with a random
642b2a3 @jonmeredith Updated 1.0.0 release notes with delete/restore changes.
jonmeredith authored
410 client id and created a sibling, and if allow_mult=false the two resolved
411 by the last updated time.
66228e4 @jaredmorrow Cleanup spacing
jaredmorrow authored
412
e2d161a @rzezeski Add Search release notes
rzezeski authored
413 *** Search
414
415 **** Integration into Riak
416
417 Prior to the 1.0 release if you wanted a Riak cluster with search
418 capability you needed to install the Riak Search package. As of 1.0
419 this functionality is now included with the standard and enterprise
420 Riak packages. By default this functionality is turned off but
421 enabling it is a simple matter of changing the =enabled= flag to
422 =true= in the =riak_search= section of the =app.config= file.
423
424 **** Data Center Replication Support
425
426 Multi-datacenter replication that comes with Riak EDS now fully
427 supports Search. Now, not only will the standard KV data be
428 replicated but also any indexes created by Search. To be clear, this
429 includes all indexes no matter how they were created; whether by the
430 Search bucket hook, =search-cmd index=, or the Solr-like interface.
431
432 **** Removal of Java Support
433
434 Prior to 1.0 Riak Search provided the ability to interface with the
435 standard Lucene analyzers or even other customer analyzers written in
436 Java. While this certainly can be useful it added extra complexity to
437 both the code and the running system. After consulting with our
438 clients and community it was determined that removing Java support
439 makes the most sense at this point in time.
440
96c72de @rzezeski Add release notes about Search field listing
rzezeski authored
441 **** Add field listing support to Solr-like interface
442
443 [[https://github.com/basho/riak_search/pull/86][A patch submitted by Greg Pascale]] adds field listing support for
444 Search's Solr-like interface. This allows you to return only the
445 fields you want by specifying a list of comma-separated field names
446 for the query param =fl=. Furthermore, if you specify only the unique
447 field (which is =id= by default) then Search will perform an
448 optimization and not fetch any of the underlying objects. This is
449 very nice if you're only interested in the keys of the matching
450 objects as it potentially saves Search from doing a lot of unnecessary
451 work. However, note that if you specify something like
452 =fl=id&sort=other_field= that Search will return a =400 Bad Request=.
453 This is because the above optimization currently prevents Search from
454 access to the =other_field=.
455
e2d161a @rzezeski Add Search release notes
rzezeski authored
456 **** Miscellaneous
457
458 - [[https://github.com/basho/riak_search/pull/66][Fixed]] memory leak that could occur as the result of running
459 intersection queries.
460
461 - The Solr-like interface now allows to "presort" based on key (where
462 key is the matching "document" id, in the case of an indexed bucket
463 this is the object key) which may be useful if the key has a
464 meaningful order. For example, the timestamp of a tweet.
465
466 - Removed the search shell.
467
468 - Removed JavaScript extractor support.
469
470 - Ability to enabled KV indexing by setting the =search= bucket
471 property to =true=.
472
473 - [[https://github.com/basho/riak_search/pull/84][Streamlined]] custom extractor bucket property.
474
bc79803 @kellymclaughlin Add some release notes about a search fix to lucene_parser error handlin...
kellymclaughlin authored
475 - Fixed bug in =lucene_parser= to handle all errors returned from
476 calls to =lucene_scan:string=.
e2d161a @rzezeski Add Search release notes
rzezeski authored
477
9601884 @jaredmorrow Create 'known issues' section
jaredmorrow authored
478 ** Known Issues
479 *** Rolling Upgrade From Riak Search 14.2 to Riak 1.0.0
480
bcdb734 @jaredmorrow Change wrap to make github read properly
jaredmorrow authored
481 There are a couple of caveats for rolling upgrade from
482 Riak Search 0.14.2 to Riak 1.0.0.
9601884 @jaredmorrow Create 'known issues' section
jaredmorrow authored
483
484 First, there are some extra steps that need to be taken when
485 installing the new package. Instead of simply installing the new
486 package you must uninstall the old one, move the data dir, and then
487 install the new package.
488
489 Second, while in a mixed cluster state some queries will return
490 incorrect results. It's tough to say which queries will exhibit this
491 behavior because it depends on which node the data is stored and what
492 node is making the query. Essentially, if two nodes with different
493 versions need to coordinate on a query it will produce incorrect
494 results. Once all nodes have been upgrade to 1.0.0 all queries will
495 return the correct results.
496
c1ae7b5 @jonmeredith Added bz://1218 message about ring handler crash on riak_core:leave().
jonmeredith authored
497 *** Intermittent CRASH REPORT on node leave (bz://1218)
498
499 There is a harmless race condition that sometimes triggers a crash when a node leaves
500 the cluster. It can be ignored. It shows up on the console/logs as:
501
113be69 @jaredmorrow Last attempt to wrangle github's weird parsing
jaredmorrow authored
502 =(08:00:31.564 [notice] "node removal completed, exiting.")=
503
504 =(08:00:31.578 [error] CRASH REPORT Process riak_core_ring_manager with 0 neighbours crashed with reason: timeout_value)=
c1ae7b5 @jonmeredith Added bz://1218 message about ring handler crash on riak_core:leave().
jonmeredith authored
505
ad3d88c @massung Added pbc_connects_total issue to RELEASE-NOTES.org
massung authored
506 *** Node stats incorrectly report pbc_connects_total
507
508 The new code path for recording stats is not currently incrementing the
509 total number of protocol buffer connections made to the node, causing it
acace6f @jaredmorrow More formatting changes
jaredmorrow authored
510 to incorrectly report 0 in both =riak-admin status= and =GET /stats= .
9601884 @jaredmorrow Create 'known issues' section
jaredmorrow authored
511
2a6c8a1 @rustyio Add information about bz://1231 (multi-backend does not support 2i.)
rustyio authored
512 *** Secondary Indexes not supported under Multi Backend
513
514 Multi Backend does not correctly expose all capabilities of its
515 child backends. This prohibits using Secondary Indexes with Multi
516 Backend. Currently, Secondary Indexing is only supported for the
acace6f @jaredmorrow More formatting changes
jaredmorrow authored
517 ELevelDB backend (=riak_kv_eleveldb_backend=). Tracked as [[https://issues.basho.com/show_bug.cgi?id=1231][Bug 1231]].
2a6c8a1 @rustyio Add information about bz://1231 (multi-backend does not support 2i.)
rustyio authored
518
2d6e153 @beerriot add known issue of reduce-during-handoff to release notes
beerriot authored
519 *** MapReduce reduce phase may run more often than requested
520
521 If a reduce phase of a MapReduce query is handed off from one Riak
522 Pipe vnode to another it immediately and unconditionally reduces the
523 inputs it has accumulated. This may cause the reduce function to be
524 evaluated more often than requested by the batch size configuration
acace6f @jaredmorrow More formatting changes
jaredmorrow authored
525 options. Tracked as [[https://issues.basho.com/show_bug.cgi?id=1183][Bug 1183]] and [[https://issues.basho.com/show_bug.cgi?id=1184][Bug 1184]].
2d6e153 @beerriot add known issue of reduce-during-handoff to release notes
beerriot authored
526
043d299 @jtuple Add known gossip overload issue to release notes
jtuple authored
527 *** Potential Cluster/Gossip Overload
528
529 The new cluster protocol is designed to ensure that a Riak cluster
530 converges as quickly as possible. When running multiple Riak nodes on
531 a single-machine, the underlying gossip mechanism may become CPU-bound
532 for a period of time and cause cluster related commands to
533 timeout. This includes the following =riak-admin= commands: =join,
534 leave, remove, member_status, ring_status=. Incoming client requests
535 and other Riak operations will continue to function, although latency
536 may be impacted. The cluster will continue to handle gossip messages
537 and will eventually converge, resolving this issue. Note: This
538 behavior only occurs when adding/removing nodes from the cluster, and
539 will not occur when a cluster is stable. Also, this behavior has only
540 been observed when running multiple nodes on a single machine, and has
541 not been observed when running Riak on multiple servers or EC2
542 instances.
543
778f73f @jtuple Update RELEASE-NOTES.org
jtuple authored
544
dffe517 @jaredmorrow Add major sections
jaredmorrow authored
545 ** Bugs Fixed
864c948 @jaredmorrow Fix broken url on bug 105.
jaredmorrow authored
546 -[[https://issues.basho.com/show_bug.cgi?id=0105][bz0105 - Python client new_binary doesn't set the content_type well]]
dffe517 @jaredmorrow Add major sections
jaredmorrow authored
547 -[[https://issues.basho.com/show_bug.cgi?id=0123][bz0123 - default_bucket_props in app.config is not merged with the hardcoded defaults]]
548 -[[https://issues.basho.com/show_bug.cgi?id=0218][bz0218 - bin/riak-admin leave needs to remove abandoned ring files]]
549 -[[https://issues.basho.com/show_bug.cgi?id=0260][bz0260 - Expose tombstones as conflicts when allow_mult is true]]
550 -[[https://issues.basho.com/show_bug.cgi?id=0294][bz0294 - Possible race condition in nodetool]]
551 -[[https://issues.basho.com/show_bug.cgi?id=0325][bz0325 - Patch for mapred_builtins.js - reduceMin and reduceMax]]
552 -[[https://issues.basho.com/show_bug.cgi?id=0420][bz0420 - Links are incorrectly translated in riak_object:dejsonify_values/2]]
553 -[[https://issues.basho.com/show_bug.cgi?id=0426][bz0426 - bin/riak-admin leave has poor console output]]
554 -[[https://issues.basho.com/show_bug.cgi?id=0441][bz0441 - detect and report bad datafile entry]]
555 -[[https://issues.basho.com/show_bug.cgi?id=0461][bz0461 - Guard against non-string values of content-type in riak-erlang-client]]
323ec74 @jaredmorrow Update release notes
jaredmorrow authored
556 -[[https://issues.basho.com/show_bug.cgi?id=464][bz0464 - riak-admin status has garbage cpu_nprocs/avg1/5/15 on Solaris]]
94200e9 @jaredmorrow Update release notes for rc2
jaredmorrow authored
557 -[[https://issues.basho.com/show_bug.cgi?id=475][bz0475 - k/v FSMs should fail if no nodes are available - currently they time out]]
dffe517 @jaredmorrow Add major sections
jaredmorrow authored
558 -[[https://issues.basho.com/show_bug.cgi?id=0502][bz0502 - Minor merge_index code cleanup]]
559 -[[https://issues.basho.com/show_bug.cgi?id=0564][bz0564 - Planner's subprocesses run long after {timeout, range_loop} exception]]
560 -[[https://issues.basho.com/show_bug.cgi?id=0599][bz0599 - Consider adding erlang:memory/0 information to stats output]]
561 -[[https://issues.basho.com/show_bug.cgi?id=0605][bz0605 - riak_kv_wm_raw does not handle put_fsm timeout]]
562 -[[https://issues.basho.com/show_bug.cgi?id=0617][bz0617 - Riak URL decodes keys submitted in the Link header]]
563 -[[https://issues.basho.com/show_bug.cgi?id=0688][bz0688 - Ring does not settle when building large clusters ]]
564 -[[https://issues.basho.com/show_bug.cgi?id=0710][bz0710 - "riak ping" exits with status 0 when ping fails]]
565 -[[https://issues.basho.com/show_bug.cgi?id=0716][bz0716 - Handoff Sender crashes loudly when remote node dies]]
566 -[[https://issues.basho.com/show_bug.cgi?id=0808][bz0808 - The use of fold/3 function in do_list_keys/6 in riak_kv_vnode does not allow backends to take advantage of bucket aware optimizations]]
567 -[[https://issues.basho.com/show_bug.cgi?id=0823][bz0823 - Handoff processes crash irretrievably when receiving TCP garbage, resulting in node failure]]
75a23a6 @rzezeski Fix links
rzezeski authored
568 -[[https://issues.basho.com/show_bug.cgi?id=861][bz0861 - merge_index throws errors when data path contains a period]]
dffe517 @jaredmorrow Add major sections
jaredmorrow authored
569 -[[https://issues.basho.com/show_bug.cgi?id=0869][bz0869 - Any commands that change the ring should require the ringready command to return TRUE]]
570 -[[https://issues.basho.com/show_bug.cgi?id=0878][bz0878 - riak-admin leave, then stop node, then restart -> handoff transfers do not resume]]
571 -[[https://issues.basho.com/show_bug.cgi?id=0911][bz0911 - Fix #scope{} and #group{} operator preplanning]]
572 -[[https://issues.basho.com/show_bug.cgi?id=0931][bz0931 - Cluster should not use partition ownership to find list of nodes]]
573 -[[https://issues.basho.com/show_bug.cgi?id=0939][bz0939 - Fast map phase can overrun slower reduce phase]]
75a23a6 @rzezeski Fix links
rzezeski authored
574 -[[https://issues.basho.com/show_bug.cgi?id=948][bz0948 - Fix or remove commented out QC tests]]
dffe517 @jaredmorrow Add major sections
jaredmorrow authored
575 -[[https://issues.basho.com/show_bug.cgi?id=0953][bz0953 - Change Riak Search to use the Whitespace analyzer by default]]
576 -[[https://issues.basho.com/show_bug.cgi?id=0954][bz0954 - Wildcard queries are broken with Whitespace analyzer]]
75a23a6 @rzezeski Fix links
rzezeski authored
577 -[[https://issues.basho.com/show_bug.cgi?id=963][bz0963 - UTF8_test errors]]
dffe517 @jaredmorrow Add major sections
jaredmorrow authored
578 -[[https://issues.basho.com/show_bug.cgi?id=0967][bz0967 - Upgrade riak_search to compile on Erlang R14B01]]
75a23a6 @rzezeski Fix links
rzezeski authored
579 -[[https://issues.basho.com/show_bug.cgi?id=970][bz0970 - Deleting a non-indexed object from an indexed bucket throws an error]]
dffe517 @jaredmorrow Add major sections
jaredmorrow authored
580 -[[https://issues.basho.com/show_bug.cgi?id=0989][bz0989 - riak_kv_map_master crashes when counters are out of date]]
581 -[[https://issues.basho.com/show_bug.cgi?id=1003][bz1003 - REST API and PBC API have incompatible naming rules]]
582 -[[https://issues.basho.com/show_bug.cgi?id=1024][bz1024 - Valid objects return notfound during heavy partition transfer]]
583 -[[https://issues.basho.com/show_bug.cgi?id=1033][bz1033 - delete_resource doesn't handle case where object is no longer extant]]
584 -[[https://issues.basho.com/show_bug.cgi?id=1047][bz1047 - Javascript VM worker process is not restarted after crash]]
75a23a6 @rzezeski Fix links
rzezeski authored
585 -[[https://issues.basho.com/show_bug.cgi?id=1050][bz1050 - Add inline field support / filter support to the KV interface]]
dffe517 @jaredmorrow Add major sections
jaredmorrow authored
586 -[[https://issues.basho.com/show_bug.cgi?id=1052][bz1052 - riak_core_ring_handler:ensure_vnodes_started breaks on multiple vnode types]]
587 -[[https://issues.basho.com/show_bug.cgi?id=1055][bz1055 - riak_core_vnode_master keeps unnecessary "exclusions list" in its state]]
75a23a6 @rzezeski Fix links
rzezeski authored
588 -[[https://issues.basho.com/show_bug.cgi?id=1065][bz1065 - mi_buffer_converter processes sit idle with large heap]]
dffe517 @jaredmorrow Add major sections
jaredmorrow authored
589 -[[https://issues.basho.com/show_bug.cgi?id=1067][bz1067 - deprecate riak_kv_util:try_cast/fallback]]
590 -[[https://issues.basho.com/show_bug.cgi?id=1072][bz1072 - spiraltime crash (in riak_kv_stat)]]
591 -[[https://issues.basho.com/show_bug.cgi?id=1075][bz1075 - java.net.SocketException: Connection reset by peer from proto client (thundering (small) herd)]]
592 -[[https://issues.basho.com/show_bug.cgi?id=1077][bz1077 - nodetool needs to support Erlang SSL distribution]]
75a23a6 @rzezeski Fix links
rzezeski authored
593 -[[https://issues.basho.com/show_bug.cgi?id=1086][bz1086 - merge_index doesn't tolerate dashes in parent paths]]
dffe517 @jaredmorrow Add major sections
jaredmorrow authored
594 -[[https://issues.basho.com/show_bug.cgi?id=1097][bz1097 - Truncated data file then merge triggers error in bitcask_fileops:fold/3]]
75a23a6 @rzezeski Fix links
rzezeski authored
595 -[[https://issues.basho.com/show_bug.cgi?id=1103][bz1103 - RHEL/CentOS riaksearch init script uses 'riaksearch' as username but riaksearch install RPM creates 'riak' user]]
dffe517 @jaredmorrow Add major sections
jaredmorrow authored
596 -[[https://issues.basho.com/show_bug.cgi?id=1109][bz1109 - PB interface error when content-type is JSON and {not_found} in results]]
597 -[[https://issues.basho.com/show_bug.cgi?id=1110][bz1110 - Riak Search integration with MapReduce does not work as of Riak Search 0.14.2rc9]]
75a23a6 @rzezeski Fix links
rzezeski authored
598 -[[https://issues.basho.com/show_bug.cgi?id=1116][bz1116 - riak_search_sup never starts riak_kv_js_manager]]
dffe517 @jaredmorrow Add major sections
jaredmorrow authored
599 -[[https://issues.basho.com/show_bug.cgi?id=1125][bz1125 - HTTP Delete returns a 204 when the RW param cannot be satisfied, expected 500]]
600 -[[https://issues.basho.com/show_bug.cgi?id=1126][bz1126 - riak_kv_cache_backend doesn't stop]]
87279b8 @jaredmorrow Update release-notes bug list after az1130 fix and give credit in THANKS...
jaredmorrow authored
601 -[[https://issues.basho.com/show_bug.cgi?id=1130][bz1130 - Debian packages should depend on 'sudo']]
75a23a6 @rzezeski Fix links
rzezeski authored
602 -[[https://issues.basho.com/show_bug.cgi?id=1131][bz1131 - js_thread_stack isn't described in /etc/riaksearch/app.config]]
dffe517 @jaredmorrow Add major sections
jaredmorrow authored
603 -[[https://issues.basho.com/show_bug.cgi?id=1144][bz1144 - Riak Search custom JS extractor not initializing VM pool properly]]
604 -[[https://issues.basho.com/show_bug.cgi?id=1147][bz1147 - "Proxy Objects" are not cleaned]]
605 -[[https://issues.basho.com/show_bug.cgi?id=1149][bz1149 - Delete op should not use user-supplied timeout for tombstone harvest]]
606 -[[https://issues.basho.com/show_bug.cgi?id=1155][bz1155 - Regression in single negated term]]
607 -[[https://issues.basho.com/show_bug.cgi?id=1165][bz1165 - mi_buffer doesn’t check length when reading terms]]
608 -[[https://issues.basho.com/show_bug.cgi?id=1175][bz1175 - Riak_kv_pb_socket crashes when clientId is undefined]]
609 -[[https://issues.basho.com/show_bug.cgi?id=1176][bz1176 - Error on HTTP POST or PUT that specifies indexes with integer values > 255 and returnbody=true]]
610 -[[https://issues.basho.com/show_bug.cgi?id=1177][bz1177 - riak_kv_bitcask_backend.erl's use of symlinks breaks upgrade from 0.14.2]]
611 -[[https://issues.basho.com/show_bug.cgi?id=1178][bz1178 - ring mgr and bucket fixups not playing well on startup]]
2cec140 @jaredmorrow Update release notes to reflect changes for pre3
jaredmorrow authored
612 -[[https://issues.basho.com/show_bug.cgi?id=1186][bz1186 - riak_kv_w_reduce batch size should default to 20]]
8daa295 @jaredmorrow Update release notes for 1.0.0rc1
jaredmorrow authored
613 -[[https://issues.basho.com/show_bug.cgi?id=1188][bz1188 - Worker pools don't complete work on vnode shutdown]]
845a641 @jaredmorrow Update release notes for 1.0.0pre4
jaredmorrow authored
614 -[[https://issues.basho.com/show_bug.cgi?id=1191][bz1191 - Pipe-based mapred reverses inputs to reduce]]
8daa295 @jaredmorrow Update release notes for 1.0.0rc1
jaredmorrow authored
615 -[[https://issues.basho.com/show_bug.cgi?id=1195][bz1195 - Running "make rel" fails with riak-1.0.0pre3 source tarball]]
94200e9 @jaredmorrow Update release notes for rc2
jaredmorrow authored
616 -[[https://issues.basho.com/show_bug.cgi?id=1197][bz1197 - riak attach does not play well with scripting - stdin data may be lost]]
845a641 @jaredmorrow Update release notes for 1.0.0pre4
jaredmorrow authored
617 -[[https://issues.basho.com/show_bug.cgi?id=1200][bz1200 - Bitcask backend merges repeatedly, and misplaces files]]
8daa295 @jaredmorrow Update release notes for 1.0.0rc1
jaredmorrow authored
618 -[[https://issues.basho.com/show_bug.cgi?id=1202][bz1202 - Bucket listing fails when there are indexed objects]]
94200e9 @jaredmorrow Update release notes for rc2
jaredmorrow authored
619 -[[https://issues.basho.com/show_bug.cgi?id=1214][bz1214 - Handoff crash with async enabled+leveldb]]
620 -[[https://issues.basho.com/show_bug.cgi?id=1215][bz1215 - get FSM timeout causes new stats to crash]]
8daa295 @jaredmorrow Update release notes for 1.0.0rc1
jaredmorrow authored
621 -[[https://issues.basho.com/show_bug.cgi?id=1216][bz1216 - Not possible to control search hook order with bucket fixups]]
94200e9 @jaredmorrow Update release notes for rc2
jaredmorrow authored
622 -[[https://issues.basho.com/show_bug.cgi?id=1220][bz1220 - riak-admin ringready only shows 1.0 nodes in a mixed cluster]]
623 -[[https://issues.basho.com/show_bug.cgi?id=1224][bz1224 - platform_data_dir (/data) is not being created before accessed for some packages]]
d3515b7 @jaredmorrow Update release notes for 1.0.0
jaredmorrow authored
624 -[[https://issues.basho.com/show_bug.cgi?id=1226][bz1226 - Riak creates identical vtags for the same bucket/key with different values]]
498236e @jtuple Update clustering section of relnotes
jtuple authored
625 -[[https://issues.basho.com/show_bug.cgi?id=1229][bz1229 - "Downed" (riak-admin down) nodes don't rejoin cluster]]
d3515b7 @jaredmorrow Update release notes for 1.0.0
jaredmorrow authored
626
Something went wrong with that request. Please try again.