ceph/doc/changelog/v0.67.5.txt

   1 commit a60ac9194718083a4b6a225fc17cad6096c69bd1
   2 Author: Gary Lowell <gary.lowell@inktank.com>
   3 Date:   Thu Dec 19 20:53:10 2013 +0000
   4
   5     v0.67.5
   6
   7 commit 8745ba63533ba77b2ec9d40780cad822a577089c
   8 Author: Yehuda Sadeh <yehuda@inktank.com>
   9 Date:   Wed Dec 18 13:11:01 2013 -0800
  10
  11     rgw: fix use-after-free when releasing completion handle
  12
  13     Backport: emperor, dumpling
  14     Signed-off-by: Yehuda Sadeh <yehuda@inktank.com>
  15     Reviewed-by: Sage Weil <sage@inktank.com>
  16     (cherry picked from commit c8890ab2d46fe8e12200a0d2f9eab31c461fb871)
  17
  18 commit 7587ee52161d74a41bf89dc741e8ece9c964fb67
  19 Author: Yehuda Sadeh <yehuda@inktank.com>
  20 Date:   Wed Dec 18 13:10:21 2013 -0800
  21
  22     rgw: don't return data within the librados cb
  23
  24     Fixes: #7030
  25     The callback is running within a single Finisher thread, thus we
  26     shouldn't block there. Append read data to a list and flush it within
  27     the iterate context.
  28
  29     Reviewed-by: Sage Weil <sage@inktank.com>
  30     Signed-off-by: Yehuda Sadeh <yehuda@inktank.com>
  31     (cherry picked from commit d6a4f6adfaa75c3140d07d6df7be03586cc16183)
  32
  33 commit 9bd1ebe960aae69f6528a33809a58c13481f8df3
  34 Author: Christophe Courtaut <christophe.courtaut@gmail.com>
  35 Date:   Mon Jul 22 15:15:38 2013 +0200
  36
  37     rgw: Allow wildcard in supported keystone roles.
  38
  39     http://tracker.ceph.com/issues/4365 fixes #4365
  40
  41     Signed-off-by: Christophe Courtaut <christophe.courtaut@gmail.com>
  42     (cherry picked from commit 60195581273aee259e8c83d0b471af859d928342)
  43
  44 commit e52b9b76844f935daf2dc060905f869248bba60b
  45 Author: Roald J. van Loon <roaldvanloon@gmail.com>
  46 Date:   Fri Aug 9 13:31:10 2013 +0200
  47
  48     Validate S3 tokens against Keystone
  49
  50     - Added config option to allow S3 to use Keystone auth
  51     - Implemented JSONDecoder for KeystoneToken
  52     - RGW_Auth_S3::authorize now uses rgw_store_user_info on keystone auth
  53     - Minor fix in get_canon_resource; dout is now after the assignment
  54
  55     Reviewed-by: Yehuda Sadeh<yehuda@inktank.com>
  56     Signed-off-by: Roald J. van Loon <roaldvanloon@gmail.com>
  57     (cherry picked from commit a200e184b15a03a4ca382e94caf01efb41cb9db7)
  58
  59     Conflicts:
  60         src/rgw/rgw_swift.h
  61
  62 commit 5df2f35ee5797ffe85e8398b1ecf75f2eefb5031
  63 Author: Yehuda Sadeh <yehuda@inktank.com>
  64 Date:   Mon Oct 21 18:02:32 2013 -0700
  65
  66     rgw: init src_bucket_name, src_object in the S3 handler
  67
  68     Be consistent and initialize these fields also in the S3 case.
  69
  70     Signed-off-by: Yehuda Sadeh <yehuda@inktank.com>
  71     (cherry picked from commit ccaab2abf9e49ce3e50c15df793a3a3ca6b29bb8)
  72
  73 commit 0edee8e9c9b140a280dc782b491c330c7f86edfc
  74 Author: Yehuda Sadeh <yehuda@inktank.com>
  75 Date:   Mon Oct 21 14:45:59 2013 -0700
  76
  77     rgw: get rid of req_state.bucket_name
  78
  79     No need for this field, as we already have req_state.bucket_name_str.
  80     This saves us some memory allocation / freeing through every request
  81     processing.
  82
  83     Signed-off-by: Yehuda Sadeh <yehuda@inktank.com>
  84     (cherry picked from commit db7eb771226beaa2372ef5860ce742cb3457cd89)
  85
  86     Conflicts:
  87         src/rgw/rgw_rest_s3.cc
  88
  89 commit 0eae05f31e4a03e923d51521ecd0078657eda9ba
  90 Author: Yehuda Sadeh <yehuda@inktank.com>
  91 Date:   Mon Oct 21 14:17:12 2013 -0700
  92
  93     rgw: turn swift COPY into PUT
  94
  95     Fixes: #6606
  96     The swift COPY operation is unique in a sense that it's a write
  97     operation that has its destination not set by the URI target, but by a
  98     different HTTP header. This is problematic as there are some hidden
  99     assumptions in the code that the specified bucket/object in the URI is
 100     the operation target. E.g., certain initialization functions, quota,
 101     etc. Instead of creating a specialized code everywhere for this case
 102     just turn it into a regular copy operation, that is, a PUT with
 103     a specified copy source.
 104
 105     Signed-off-by: Yehuda Sadeh <yehuda@inktank.com>
 106     (cherry picked from commit cbf8f9a51737e7d89fb0ec0832b50fd47c35b08d)
 107
 108 commit 455f28a78ee22b6a535831ff70a330fa1379567c
 109 Author: Samuel Just <sam.just@inktank.com>
 110 Date:   Tue Aug 13 20:32:14 2013 -0700
 111
 112     OSDMonitor: add 'osd perf' command to dump recent osd perf information
 113
 114     Signed-off-by: Samuel Just <sam.just@inktank.com>
 115     (cherry picked from commit 94c3f29a32cbf169d896015da6765febd3c724e0)
 116
 117 commit ec086df1beccda6d5a6c12c727c84c2555b0b2cd
 118 Author: Samuel Just <sam.just@inktank.com>
 119 Date:   Tue Aug 13 17:36:57 2013 -0700
 120
 121     ObjectStore: add ObjectStore::get_cur_stats and impl in FileStore
 122
 123     Signed-off-by: Samuel Just <sam.just@inktank.com>
 124     (cherry picked from commit ebde89d5602536b4bc651737e4280fdfb6634c32)
 125
 126 commit 5bf7e7be1fa39ec39726635581e98a0e3a1df20f
 127 Author: Samuel Just <sam.just@inktank.com>
 128 Date:   Tue Aug 13 17:21:39 2013 -0700
 129
 130     osd_types: add osd_perf_stat_t type
 131
 132     Signed-off-by: Samuel Just <sam.just@inktank.com>
 133     (cherry picked from commit 95f3353a6ab3a4dd2bf28eaca7122487942e913e)
 134
 135 commit b37c5d208778ee52292757e96265d59d75f19654
 136 Author: Sage Weil <sage@inktank.com>
 137 Date:   Tue Dec 3 11:22:56 2013 -0800
 138
 139     rgw: do not use UNPRIVILEGED_DAEMON option
 140
 141     This is sort of a backport for 4f403c26dc0048ad63e20d20369fa86bfb31c50e
 142     in that we get the same behavior change, but only for rgw.
 143
 144     Signed-off-by: Sage Weil <sage@inktank.com>
 145
 146 commit a447fb7d04fbad84f9ecb57726396bb6ca29d8f6
 147 Author: Sage Weil <sage@inktank.com>
 148 Date:   Tue Dec 3 21:51:26 2013 -0800
 149
 150     osd/OSDMonitor: accept 'osd pool set ...' value as string
 151
 152     Newer monitors take this as a CephString.  Accept that so that if we are
 153     mid-upgrade and get a forwarded message using the alternate schema from
 154     a future mon we will handle it properly.
 155
 156     Signed-off-by: Sage Weil <sage@inktank.com>
 157     Reviewed-by: Greg Farnum <greg@inktank.com>
 158
 159 commit 9875c8b1992c59cc0c40901a44573676cdff2669
 160 Author: Josh Durgin <josh.durgin@inktank.com>
 161 Date:   Mon Nov 25 13:43:43 2013 -0800
 162
 163     init, upstart: prevent daemons being started by both
 164
 165     There can be only one init system starting a daemon. If there is a
 166     host entry in ceph.conf for a daemon, sysvinit would try to start it
 167     even if the daemon's directory did not include a sysvinit file. This
 168     preserves backwards compatibility with older installs using sysvinit,
 169     but if an upstart file is present in the daemon's directory, upstart
 170     will try to start them, regardless of host entries in ceph.conf.
 171
 172     If there's an upstart file in a daemon's directory and a host entry
 173     for that daemon in ceph.conf, both sysvinit and upstart would attempt
 174     to manage it.
 175
 176     Fix this by only starting daemons if the marker file for the other
 177     init system is not present. This maintains backwards compatibility
 178     with older installs using neither sysvinit or upstart marker files,
 179     and does not break any valid configurations. The only configuration
 180     that would break is one with both sysvinit and upstart files present
 181     for the same daemon.
 182
 183     Backport: emperor, dumpling
 184     Reported-by: Tim Spriggs <tims@uahirise.org>
 185     Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
 186     (cherry picked from commit 5e34beb61b3f5a1ed4afd8ee2fe976de40f95ace)
 187
 188 commit bd08d332f5d23b6097a7e8344a6183896b2303a9
 189 Author: Samuel Just <sam.just@inktank.com>
 190 Date:   Sun Nov 3 21:02:36 2013 -0800
 191
 192     OSD: allow project_pg_history to handle a missing map
 193
 194     If we get a peering message for an old map we don't have, we
 195     can throwit out: the sending OSD will learn about the newer
 196     maps and update itself accordingly, and we don't have the
 197     information to know if the message is valid. This situation
 198     can only happen if the sender was down for a long enough time
 199     to create a map gap and its PGs have not yet advanced from
 200     their boot-up maps to the current ones, so we can rely on it
 201
 202     Fixes: #6712
 203     Signed-off-by: Samuel Just <sam.just@inktank.com>
 204     Reviewed-by: Greg Farnum <greg@inktank.com>
 205     (cherry picked from commit cd0d612e1abdf5c87082eeeccd4ca09dd14fd737)
 206
 207 commit 8cd33e3a8ebf7c2aa796ec9f92d6b554c39ff705
 208 Author: Yehuda Sadeh <yehuda@inktank.com>
 209 Date:   Mon Aug 19 16:56:27 2013 -0700
 210
 211     rgw: bucket meta remove don't overwrite entry point first
 212
 213     Fixes: #6056
 214
 215     When removing a bucket metadata entry we first unlink the bucket
 216     and then we remove the bucket entrypoint object. Originally
 217     when unlinking the bucket we first overwrote the bucket entrypoint
 218     entry marking it as 'unlinked'. However, this is not really needed
 219     as we're just about to remove it. The original version triggered
 220     a bug, as we needed to propagate the new header version first (which
 221     we didn't do, so the subsequent bucket removal failed).
 222
 223     Reviewed-by: Greg Farnum <greg@inktank.com>
 224     Signed-off-by: Yehuda Sadeh <yehuda@inktank.com>
 225     (cherry picked from commit 0373d749cea7d9b532069ba8ebca2f005b2c9f59)
 226
 227 commit a24f1326adcdda1d387e0f4e3c18dda5ec4aa055
 228 Author: Yehuda Sadeh <yehuda@inktank.com>
 229 Date:   Fri Nov 22 07:04:01 2013 -0800
 230
 231     rgw: lower some debug message
 232
 233     Fixes: #6084
 234     Backport: dumpling, emperor
 235
 236     Reported-by: Ron Allred <rallred@itrefined.com>
 237     Signed-off-by: Yehuda Sadeh <yehuda@inktank.com>
 238     (cherry picked from commit b35fc1bc2ec8c8376ec173eb1c3e538e02c1694e)
 239
 240 commit 9e2d0d311437c1ffdd43fc8c961c60aacfea339c
 241 Author: Josh Durgin <josh.durgin@inktank.com>
 242 Date:   Mon Nov 18 14:39:12 2013 -0800
 243
 244     osd: fix bench block size
 245
 246     The command was declared to take 'size' in dumpling, but was trying to
 247     read 'bsize' instead, so it always used the default of 4MiB. Change
 248     the bench command to read 'size', so it matches what existing clients
 249     are sending.
 250
 251     Fixes: #6795
 252     Backport: emperor, dumpling
 253     Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
 254     (cherry picked from commit 40a76ef0d09f8ecbea13712410d9d34f25b91935)
 255
 256     Conflicts:
 257         src/osd/OSD.cc
 258
 259 commit 35c91cac9a2b414ba3d52ea06cf6f71a0a1d4ba3
 260 Author: David Zafman <david.zafman@inktank.com>
 261 Date:   Wed Sep 25 09:19:16 2013 -0700
 262
 263     os, osd, tools: Add backportable compatibility checking for sharded objects
 264
 265     OSD
 266       New CEPH_OSD_FEATURE_INCOMPAT_SHARDS
 267     FileStore
 268       NEW CEPH_FS_FEATURE_INCOMPAT_SHARDS
 269       Add FSSuperblock with feature CompatSet in it
 270       Store sharded_objects state using CompatSet
 271       Add set_allow_sharded_objects() and get_allow_sharded_objects() to FileStore/ObjectStore
 272       Add read_superblock()/write_superblock() internal filestore functions
 273     ceph_filestore_dump
 274       Add OSDsuperblock to export format
 275       Use CompatSet from OSD code itself in filestore-dump tool
 276       Always check compatibility of OSD features with on-disk features
 277       On import verify compatibility of on-disk features with export data
 278       Bump super_ver due to export format change
 279
 280     Backport: dumpling, cuttlefish
 281
 282     Signed-off-by: David Zafman <david.zafman@inktank.com>
 283     Reviewed-by: Samuel Just <sam.just@inktank.com>
 284     (cherry picked from commit c6b83180f9f769de27ca7890f5f8ec507ee743ca)
 285
 286     Conflicts:
 287
 288         src/os/FileStore.cc
 289         src/os/FileStore.h
 290         src/osd/OSD.cc
 291
 292     Excluded from cherry-pick:
 293       Didn't add set_allow_sharded_objects() and get_allow_sharded_objects() to FileStore/ObjectStore
 294       Didn't add code to check for incomplete transition to sharded objects in ceph-filestore-dump
 295
 296 commit 372f62717c56d9ab883ae2942e13d6d8d37c2925
 297 Author: Ray Lv <xiangyulv@gmail.com>
 298 Date:   Tue Oct 29 19:34:51 2013 +0800
 299
 300     rgw: Wrap hex_to_num table into class HexTable
 301
 302     Reviewed-by: Yehuda Sadeh <yehuda@inktank.com>
 303     Signed-off-by: Ray Lv <xiangyulv@gmail.com>
 304     (cherry picked from commit 588ed60a8ec1d8db5e096fd8f7b762f2afcf7dd3)
 305
 306 commit 6917b02530103b8c86ed75592da33144b0dea168
 307 Merge: 7ef210a f1fa811
 308 Author: Yehuda Sadeh <yehuda@inktank.com>
 309 Date:   Tue Nov 5 12:27:42 2013 -0800
 310
 311     Merge pull request #805 from ceph/wip-rgw-replica-log-dumpling
 312
 313     backport a bunch of rgw stuff to dumpling
 314
 315     Reviewed-by: Yehuda Sadeh <yehuda@inktank.com>
 316
 317 commit f1fa8116d441924d44c99624829f3daa090c821c
 318 Author: Josh Durgin <josh.durgin@inktank.com>
 319 Date:   Fri Nov 1 16:12:52 2013 -0700
 320
 321     rgw: don't turn 404 into 400 for the replicalog api
 322
 323     404 is not actually a problem to clients like radosgw-agent, but 400
 324     implies something about the request was incorrect.
 325
 326     Backport: dumpling
 327     Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
 328     (cherry picked from commit 306ec7142d699c26ce874b11dc02ccdb3cf296c7)
 329
 330 commit 739e7861f6d5fde5c4a2b59170377b6147fecfa3
 331 Author: Yehuda Sadeh <yehuda@inktank.com>
 332 Date:   Tue Sep 3 13:27:21 2013 -0700
 333
 334     rgw: when failing read from client, return correct error
 335
 336     Fixes: #6214
 337     When getting a failed read from client when putting an object
 338     we returned the wrong value (always 0), which in the chunked-
 339     upload case ended up in assuming that the write was done
 340     successfully.
 341
 342     Signed-off-by: Yehuda Sadeh <yehuda@inktank.com>
 343     Reviewed-by: Sage Weil <sage@inktank.com>
 344     (cherry picked from commit 0499948aad64c6c66668b43963403fb0cb1a2737)
 345
 346 commit 465087e64ab2bd85d7f5e183565e3619b3aa76fc
 347 Author: Yehuda Sadeh <yehuda@inktank.com>
 348 Date:   Mon Sep 30 14:15:19 2013 -0700
 349
 350     rgw: fix leak in RGWMetadataManager::remove()
 351
 352     Backport: dumpling
 353     Fixes: #6445
 354     handler->get() returns a newly allocated obj, need to delete it when
 355     done.
 356
 357     Signed-off-by: Yehuda Sadeh <yehuda@inktank.com>
 358     (cherry picked from commit 8dd7ea3fadc06b5ebb330af41d686d3da155fb5d)
 359
 360 commit a4dc2694db6a7e45109f111dcdfd077da67b6030
 361 Author: Yehuda Sadeh <yehuda@inktank.com>
 362 Date:   Mon Aug 26 11:16:08 2013 -0700
 363
 364     rgw: quiet down warning message
 365
 366     Fixes: #6123
 367     We don't want to know about failing to read region map info
 368     if it's not found, only if failed on some other error. In
 369     any case it's just a warning.
 370
 371     Signed-off-by: Yehuda Sadeh <yehuda@inktank.com>
 372     Reviewed-by: Sage Weil <sage@inktank.com>
 373     (cherry picked from commit 055e31359ac977943f014270cc56a1b7786abee3)
 374
 375 commit 7f57d9dda74a1feb2da85be3c77d61c6f81ec1b2
 376 Author: Yehuda Sadeh <yehuda@inktank.com>
 377 Date:   Fri Sep 13 14:43:54 2013 -0700
 378
 379     rgw: try to create log pool if doesn't exist
 380
 381     When using replica log, if the log pool doesn't exist all operations are
 382     going to fail. Try to create it if doesn't exist.
 383
 384     Reviewed-by: Josh Durgin <josh.durgin@inktank.com>
 385     Signed-off-by: Yehuda Sadeh <yehuda@inktank.com>
 386     (cherry picked from commit 4216eac0f59af60f60d4ce909b9ace87a7b64ccc)
 387
 388 commit 4fb233cf702957e782083d760877afa48d4f6965
 389 Author: Yehuda Sadeh <yehuda@inktank.com>
 390 Date:   Fri Oct 11 16:56:46 2013 -0700
 391
 392     formatter: dump_bool dumps unquoted strings
 393
 394     Signed-off-by: Yehuda Sadeh <yehuda@inktank.com>
 395     (cherry picked from commit ad409f8a6d230e9b1199226a333bb54159c2c910)
 396
 397 commit de7ede925acc47a5aee84b9fe82b9cf0368aa1b2
 398 Author: Yehuda Sadeh <yehuda@inktank.com>
 399 Date:   Fri Sep 27 13:15:35 2013 -0700
 400
 401     Formatter: add dump_bool()
 402
 403     Signed-off-by: Yehuda Sadeh <yehuda@inktank.com>
 404     (cherry picked from commit 28949d5d43beba7cef37cb2f83e3399d978061a6)
 405
 406 commit c5991f26febd86878aeb78baab33c071660fdee6
 407 Author: Josh Durgin <josh.durgin@inktank.com>
 408 Date:   Thu Oct 24 08:42:48 2013 -0700
 409
 410     rgw: escape bucket and object names in StreamReadRequests
 411
 412     This fixes copy operations for objects that contain unsafe characters,
 413     like a newline, which would return a 403 otherwise, since the GET to
 414     the source rgw would be unable to verify the signature on a partially
 415     valid bucket name.
 416
 417     Fixes: #6604
 418     Backport: dumpling
 419     Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
 420     (cherry picked from commit ec45b3b88c485140781b23d2c4f582f2cc26ea43)
 421
 422 commit 672679d26110dd93393bc18a681ae41175e4a299
 423 Author: Josh Durgin <josh.durgin@inktank.com>
 424 Date:   Thu Oct 24 08:37:25 2013 -0700
 425
 426     rgw: move url escaping to a common place
 427
 428     This is useful outside of the s3 interface. Rename url_escape()
 429     url_encode() for consistency with the exsting common url_decode()
 430     function. This is in preparation for the next commit, which needs
 431     to escape url-unsafe characters in another place.
 432
 433     Backport: dumpling
 434     Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
 435     (cherry picked from commit dd308cd481b368f90a64220847b91fc233d92a59)
 436
 437 commit 91997fca6ec6c06beecd63c3e2ae521319fca7c2
 438 Author: Josh Durgin <josh.durgin@inktank.com>
 439 Date:   Thu Oct 24 08:34:24 2013 -0700
 440
 441     rgw: update metadata log list to match data log list
 442
 443     Send the last marker whether the log is truncated in the same format
 444     as data log list, so clients don't have more needless complexity
 445     handling the difference.  Keep bucket index logs the same, since they
 446     contain the marker already, and are not used in exactly the same way
 447     metadata and data logs are.
 448
 449     Backport: dumpling
 450     Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
 451     (cherry picked from commit e0e8fb1b2b4a308b2a9317e10c6fd53ad48dbfaf)
 452
 453 commit 8916648521e2362a54eed0f5ebe078a269e4d703
 454 Author: Josh Durgin <josh.durgin@inktank.com>
 455 Date:   Thu Oct 24 08:26:19 2013 -0700
 456
 457     rgw: include marker and truncated flag in data log list api
 458
 459     Consumers of this api need to know their position in the log. It's
 460     readily available when fetching the log, so return it.  Without the
 461     marker in this call, a client could not easily or efficiently figure
 462     out its position in the log, since it would require getting the global
 463     last marker in the log, and then reading all the log entries.
 464
 465     This would be slow for large logs, and would be subject to races that
 466     would cause potentially very expensive duplicate work.
 467
 468     Returning this atomically while fetching the log entries simplifies
 469     all of this.
 470
 471     Fixes: #6615
 472     Backport: dumpling
 473     Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
 474     (cherry picked from commit c275912509255f8bb4c854e181318b45ab0f8564)
 475
 476 commit a49c064b7a1b4083818b6f950d7e0aade71c6afe
 477 Author: Josh Durgin <josh.durgin@inktank.com>
 478 Date:   Thu Oct 24 08:18:19 2013 -0700
 479
 480     cls_log: always return final marker from log_list
 481
 482     There's no reason to restrict returning the marker to the case where
 483     less than the whole log is returned, since there's already a truncated
 484     flag to tell the client what happened.
 485
 486     Giving the client the last marker makes it easy to consume when the
 487     log entries do not contain their own marker. If the last marker is not
 488     returned, the client cannot get the last marker without racing with
 489     updates to the log.
 490
 491     Backport: dumpling
 492     Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
 493     (cherry picked from commit e74776f4176470122485a79a4c07e9c12c9fc036)
 494
 495 commit 90101aea6818c2f43bf31243d912ec9143f68882
 496 Author: Josh Durgin <josh.durgin@inktank.com>
 497 Date:   Thu Oct 10 15:50:39 2013 -0700
 498
 499     rgw: skip read_policy checks for system_users
 500
 501     A system user should still be able to examine suspended buckets, and
 502     get -ENOENT instead of -EACCESS for a deleted object.
 503
 504     Fixes: #6616
 505     Backport: dumpling
 506     Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
 507     (cherry picked from commit ea816c1c2fd47eab647d6fab96c9ca4bfeecd5bb)
 508
 509 commit 7ef210adbd4645bd45e5387c9b03a122a7ca82df
 510 Author: Sage Weil <sage@inktank.com>
 511 Date:   Tue Oct 29 08:39:27 2013 -0700
 512
 513     common/crc32c: fix #ifdef to be x86_64 specific
 514
 515     Signed-off-by: Sage Weil <sage@inktank.com>
 516
 517 commit 29f4eebf63bcba9ea86b9a2e4d093b782b5fbf1e
 518 Author: Loic Dachary <loic@dachary.org>
 519 Date:   Tue Sep 24 08:05:38 2013 +0200
 520
 521     rgw: fix rgw test to reflect usage change
 522
 523     otherwise src/test/cli/radosgw-admin/help.t fails when running make
 524     check when run after a configure --with-radosgw
 525
 526     Signed-off-by: Loic Dachary <loic@dachary.org>
 527     (cherry picked from commit e50343e4423e20130035c860ba47a0edea876f7c)
 528
 529 commit 0106569a88a92aa94fbc5efc6a0a142ad0de325d
 530 Author: Josh Durgin <josh.durgin@inktank.com>
 531 Date:   Thu Oct 24 17:25:16 2013 -0700
 532
 533     rbd.py: increase parent name size limit
 534
 535     64 characters isn't all that long. 4096 ought to be enough for anyone.
 536
 537     Fixes: #6072
 538     Backport: dumpling, cuttlefish
 539     Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
 540     (cherry picked from commit 3c0042cde5a12de0f554a16b227ab437c6254ddd)
 541
 542 commit 352d8c966c323a008362130eb979ae5dfa47494f
 543 Author: Sage Weil <sage@inktank.com>
 544 Date:   Mon Sep 23 14:10:36 2013 -0700
 545
 546     common/config: include --cluster in default usage message
 547
 548     Clean it up a bit too.
 549
 550     Signed-off-by: Sage Weil <sage@inktank.com>
 551     (cherry picked from commit 157754b3a0191c5ff534a84adbeed88025615898)
 552
 553 commit 299ddd31b29e332dc5e76bc4f871e4769698665d
 554 Author: Yan, Zheng <zheng.z.yan@intel.com>
 555 Date:   Thu Oct 10 10:35:48 2013 +0800
 556
 557     mds: fix infinite loop of MDCache::populate_mydir().
 558
 559     make MDCache::populate_mydir() only fetch bare-bone stray dirs.
 560     After all stray dirs are populated, call MDCache::scan_stray_dir(),
 561     it fetches incomplete stray dirs.
 562
 563     Fixes: #4405
 564     Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com>
 565     Reviewed-by: Greg Farnum <greg@inktank.com>
 566     Reviewed-by: Sage Weil <sage@inktank.com>
 567     (cherry picked from commit 007f06ec174d4ee5cfb578c8b3f1c96b2bb0c238)
 568
 569     Conflicts:
 570
 571         src/mds/MDCache.h
 572
 573     Reviewed-by: Greg Farnum <greg@inktank.com>
 574
 575 commit 7d99940467019bda74da83f3f1ebc2da46bd6967
 576 Author: Yehuda Sadeh <yehuda@inktank.com>
 577 Date:   Tue Oct 15 10:20:48 2013 -0700
 578
 579     rgw: fix authenticated users acl group check
 580
 581     Fixes: #6553
 582     Backport: bobtail, cuttlefish, dumpling
 583     Authenticated users group acl bit was not working correctly. Check to
 584     test whether user is anonymous was wrong.
 585
 586     Signed-off-by: Yehuda Sadeh <yehuda@inktank.com>
 587     (cherry picked from commit bebbd6cb7b71697b34b8f27652cabdc40c97a33b)
 588
 589 commit e5736de834027a281d29afcf6743347a8db97d26
 590 Author: Yehuda Sadeh <yehuda@inktank.com>
 591 Date:   Tue Oct 15 10:55:07 2013 -0700
 592
 593     rgw: change default log level
 594
 595     Fixes: #6554
 596     Backport: cuttlefish, dumpling
 597     Default log level was just too high, bring it down a bit.
 598
 599     Signed-off-by: Yehuda Sadeh <yehuda@inktank.com>
 600     Reviewed-by: Sage Weil <sage@inktank.com>
 601     (cherry picked from commit 8d7dbf85472cfca9268d81ecf057ea078cf345b3)
 602
 603 commit e6e9ee7145bcfc1774af738af607a653fe8dfe46
 604 Author: Yehuda Sadeh <yehuda@inktank.com>
 605 Date:   Fri Oct 4 13:00:26 2013 -0700
 606
 607     rgw: swift update obj metadata also add generic attrs
 608
 609     Fixes: #6462
 610     We were missing the generic attributes when we updated the object
 611     metadata (operation that only exists in the swift api).
 612
 613     Reviewed-by: Josh Durgin <josh.durgin@inktank.com>
 614     Signed-off-by: Yehuda Sadeh <yehuda@inktank.com>
 615     (cherry picked from commit f2645e1c6d7383a0ace3b239f4304e353249c4bb)
 616
 617 commit cd1c3c9e00e90b19e83c1f11a48e516a7de93665
 618 Author: Yan, Zheng <zheng.z.yan@intel.com>
 619 Date:   Wed Oct 2 11:43:12 2013 +0800
 620
 621     mds: return -EAGAIN if standby replay falls behind
 622
 623     standby replay may fall behind and get -ENOENT when reading the
 624     journal. return -EAGAIN in this case, it makes the MDS respawn itself.
 625
 626     fixes: #5458
 627
 628     Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com>
 629     Reviewed-by: Sage Weil <sage@inktank.com>
 630     (cherry picked from commit d2cb2bf6bac83ac6db9df6cb876317d30e7493cc)
 631     Reviewed-by: Greg Farnum <greg@inktank.com>
 632
 633 commit bd073eeac28d8cba969e5746c5e6adcb95820fdf
 634 Author: Yan, Zheng <zheng.z.yan@intel.com>
 635 Date:   Thu Sep 12 09:50:51 2013 +0800
 636
 637     mon/MDSMonitor: don't reset incarnation when creating newfs
 638
 639     Fixes: #6279
 640
 641     Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com>
 642     (cherry picked from commit 40613b700b87b495d67793101ae05d6ba58b2a9a)
 643     Reviewed-by: Greg Farnum <greg@inktank.com>
 644
 645 commit 3a3724b782c53c92c45624ad02df9f0de77f1545
 646 Author: Yan, Zheng <zheng.z.yan@intel.com>
 647 Date:   Thu Sep 12 10:36:39 2013 +0800
 648
 649     osdc/ObjectCacher: finish contexts after dropping object reference
 650
 651     The context to finish can be class C_Client_PutInode, which may drop
 652     inode's last reference. So we should first drop object's reference,
 653     then finish contexts.
 654
 655     Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com>
 656     (cherry picked from commit b66ac77fa7aa3ff37804918c4308a348f239af09)
 657
 658 commit 102f63993cacf3b3404256e62a3096c5836f2282
 659 Author: Sandon Van Ness <sandon@inktank.com>
 660 Date:   Tue Oct 8 12:08:08 2013 -0700
 661
 662     Go back to $PWD in fsstress.sh if compiling from source.
 663
 664     Although fsstress was being called with a static path the directory
 665     it was writing to was in the current directory so doing a cd to the
 666     source directory that is made in /tmp and then removing it later
 667     caused it to be unable to write the files in a non-existent dir.
 668
 669     This change gets the current path first and cd's back into it after
 670     it is done compiling fsstress.
 671
 672     Issue #6479.
 673
 674     Signed-off-by: Sandon Van Ness <sandon@inktank.com>
 675     Reviewed-by: Alfredo Deza <alfredo.deza@inktank.com>
 676
 677 commit 62db186a19f902b0a23d14034552c83db8de7ece
 678 Author: Gary Lowell <gary.lowell@inktank.com>
 679 Date:   Tue Aug 27 09:53:12 2013 -0700
 680
 681     ceph.spec.in:  radosgw package doesn't require mod_fcgi
 682
 683     Fixes #5702
 684
 685     Signed-off-by: Gary Lowell  <gary.lowell@inktank.com>
 686
 687 commit 42fc9b49a7bbc242096946b381aafd435ebdd530
 688 Merge: ad85b8b 67ac3cf
 689 Author: Gary Lowell <gary.lowell@inktank.com>
 690 Date:   Fri Oct 4 19:51:32 2013 +0000
 691
 692     Merge branch 'dumpling' of jenkins:ceph/ceph into dumpling
 693
 694 commit 67ac3cffa1066030ad840937958bd534e8b26272
 695 Author: Sage Weil <sage@inktank.com>
 696 Date:   Thu Oct 3 21:47:26 2013 -0700
 697
 698     ceph_test_rados: do not let rollback race with snap delete
 699
 700     Note that the OSD behaves in a weird way when you rollback to a non-
 701     existent snap, so the test probably isn't the only party at fault here.
 702
 703     Fixes (test half of): #6254
 704     Backport: dumpling, cuttlefish
 705     Signed-off-by: Sage Weil <sage@inktank.com>
 706     (cherry picked from commit 55d279b98553ba4542219b126fc7159b20b18b1f)
 707
 708     Conflicts:
 709
 710         src/test/osd/RadosModel.h
 711         src/test/osd/TestRados.cc