1 commit f93610a4421cb670b08e974c6550ee715ac528ae (tag: refs/tags/v0.80.6, refs/remotes/gh/firefly)
2 Author: Jenkins <jenkins@inktank.com>
3 Date: Wed Oct 1 09:37:12 2014 -0700
7 commit 2ecf362e15301ee07b511b2a84585ad691543569
9 Author: Sage Weil <sage@redhat.com>
10 Date: Mon Sep 29 13:55:02 2014 -0700
12 Merge pull request #2603 from dachary/wip-9620-test-mon-thrash-firefly
14 qa/workunits/cephtool/test.sh: fix thrash (ultimate)
16 Reviewed-by: Sage Weil <sage@redhat.com>
18 commit 76341b0b7581194273ac787df2b940221825d083
19 Author: Loic Dachary <loic-201408@dachary.org>
20 Date: Mon Sep 29 13:47:06 2014 +0200
22 qa/workunits/cephtool/test.sh: fix thrash (ultimate)
24 Keep the osd trash test to ensure it is a valid command but make it a
25 noop by giving it a zero argument (meaning thrash 0 OSD maps).
27 Remove the loops that were added after the command in an attempt to wait
28 for the cluster to recover and not pollute the rest of the tests. Actual
29 testing of osd thrash would require a dedicated cluster because it the
30 side effects are random and it is unnecessarily difficult to ensure they
33 http://tracker.ceph.com/issues/9620 Fixes: #9620
35 Signed-off-by: Loic Dachary <loic-201408@dachary.org>
36 (cherry picked from commit beade63a17db2e6fc68d1f55332d602f8f7cb93a)
39 qa/workunits/cephtool/test.sh
41 commit 1fafd6bf2ef03672dfa27ec7a201a274927040b7
42 Merge: ebb5995 71005c1
43 Author: Josh Durgin <josh.durgin@inktank.com>
44 Date: Thu Sep 25 10:08:03 2014 -0700
46 Merge pull request #2576 from ceph/wip-9593
48 osd/ReplicatedPG: fix objecter locking in start_flush
50 Reviewed-by: Josh Durgin <josh.durgin@inktank.com>
52 commit 71005c1bdf38212ea2fde12de08c45598c815d37
53 Author: Sage Weil <sage@redhat.com>
54 Date: Thu Sep 25 09:59:29 2014 -0700
56 osd/ReplicatedPG: fix objecter locking in start_flush
58 Broken in backport fd96eb62ece27f5c660429584c2ff2e058bc6e94.
60 Signed-off-by: Sage Weil <sage@redhat.com>
62 commit ebb599545d70e4289e385f88b4f0594b4f23f3b7
63 Merge: 2abd7a3 2e60c2f
64 Author: Samuel Just <sam.just@inktank.com>
65 Date: Tue Sep 23 10:37:27 2014 -0700
67 Merge pull request #2548 from dachary/wip-9547-python-rados-truncate-firefly
69 python radio aio_read must not truncate on \000 (firefly)
71 Reviewed-by: Samuel Just <sam.just@inktank.com>
73 commit 2abd7a3fd018b52f45f9d6b7c9d847242aa97fe9
74 Merge: 2675b0d 80f2f9c
75 Author: Samuel Just <sam.just@inktank.com>
76 Date: Tue Sep 23 10:27:15 2014 -0700
78 Merge branch 'wip-sam-testing-firefly' into firefly
80 commit 80f2f9cf687e25d6f2c4cee34fef55215ac6b027
81 Merge: bb9e65b e6da732
82 Author: Samuel Just <sam.just@inktank.com>
83 Date: Sun Sep 21 10:04:30 2014 -0700
85 Merge remote-tracking branch 'origin/wip-9240' into wip-sam-testing-firefly
87 commit bb9e65bd9292e2ee03a260cc98aa6715008e6682
88 Author: Samuel Just <sam.just@inktank.com>
89 Date: Thu Aug 28 15:32:22 2014 -0700
91 PG: wait until we've build the missing set to discover_all_missing
95 Signed-off-by: Samuel Just <sam.just@inktank.com>
96 (cherry picked from commit 970d9830a3a6e8568337c660fb8b4c4a60a2b3bf)
101 commit fa645b21b587b6a1d9adbb9cedce1af3af1e5a62
102 Author: Samuel Just <sam.just@inktank.com>
103 Date: Tue Aug 26 16:53:02 2014 -0700
105 PG: mark_log_for_rewrite on resurrection
109 Signed-off-by: Samuel Just <sam.just@inktank.com>
110 (cherry picked from commit 8346e10755027e982f26bab4642334fd91cc31aa)
112 commit fd96eb62ece27f5c660429584c2ff2e058bc6e94
113 Author: Samuel Just <sam.just@inktank.com>
114 Date: Sun Sep 7 20:13:41 2014 -0700
116 ReplicatedPG:start_flush send a second delete
118 Suppose we start with the following in the cache pool:
120 30:[29,21,20,15,10,4]:[22(21), 15(15,10), 4(4)]+head
122 The object doesn't exist at 29 or 20.
124 First, we flush 4 leaving the backing pool with:
128 Then, we begin to flush 15 with a delete with snapc 4:[4] leaving the
133 Then, we finish flushing 15 with snapc 9:[4] with leaving the backing
138 Next, snaps 10 and 15 are removed causing clone 10 to be removed leaving
141 30:[29,21,20,4]:[22(21),4(4)]+head
143 We next begin to flush 22 by sending a delete with snapc 4(4) since
144 prev_snapc is 4 <---------- here is the bug
146 The backing pool ignores this request since 4 < 9 (ORDERSNAP) leaving it
151 Then, we complete flushing 22 with snapc 19:[4] leaving the backing pool
156 Then, we begin to flush head by deleting with snapc 22:[21,20,4] leaving
157 the backing pool with:
159 22[21,20,4]:[22(21,20), 4(4)]
161 Finally, we flush head leaving the backing pool with:
163 30:[29,21,20,4]:[22(21*,20*),4(4)]+head
165 When we go to flush clone 22, all we know is that 22 is dirty, has snaps
166 [21], and 4 is clean. As part of flushing 22, we need to do two things:
167 1) Ensure that the current head is cloned as cloneid 4 with snaps [4] by
168 sending a delete at snapc 4:[4].
169 2) Flush the data at snap sequence < 21 by sending a copyfrom with snapc
172 Unfortunately, it is possible that 1, 1&2, or 1 and part of the flush
173 process for some other now non-existent clone have already been
174 performed. Because of that, between 1) and 2), we need to send
175 a second delete ensuring that the object does not exist at 20.
179 Related: 66c7439ea0888777b5cfc08bcb0fbd7bfd8653c3
180 Signed-off-by: Samuel Just <sam.just@inktank.com>
181 (cherry picked from commit 4843fd510b33a71999cdf9c2cfa2b4c318fa80fd)
183 commit 5aa35ac262c6e0910f99086f46dd792775ab6790
184 Author: Samuel Just <sam.just@inktank.com>
185 Date: Mon Aug 11 12:59:16 2014 -0700
187 ReplicatedPG::start_flush: remove superfluous loop
189 Signed-off-by: Samuel Just <sam.just@inktank.com>
190 (cherry picked from commit 66c7439ea0888777b5cfc08bcb0fbd7bfd8653c3)
192 commit 14fb643bbda51cc31919197541973c332b0c37bc
193 Merge: c468507 3578b11
194 Author: Samuel Just <sam.just@inktank.com>
195 Date: Sun Sep 21 10:03:53 2014 -0700
197 Merge remote-tracking branch 'origin/wip-9339' into wip-sam-testing-firefly
199 commit 2e60c2f1ec8c2aaf56b73a64f55becd679dd2825
200 Author: Loic Dachary <loic-201408@dachary.org>
201 Date: Sat Sep 20 12:41:30 2014 +0200
203 test: check python rados aio_read with buffers containing null
205 http://tracker.ceph.com/issues/9547 Refs: #9547
207 Signed-off-by: Loic Dachary <loic-201408@dachary.org>
208 (cherry picked from commit 226c0c7ac6ee95ff2c1665d4e7164e2962c0346e)
210 commit 72a46dec211d5c08d1ca9f969013d939c4361ba9
211 Author: Mohammad Salehe <salehe+dev@gmail.com>
212 Date: Sun May 25 10:42:11 2014 +0430
214 pybind: Fix aio_read handling of string buffer
216 Read data may contain \0, and buf.value interprerts them as string terminator.
218 Signed-off-by: Mohammad Salehe <salehe+dev@gmail.com>
219 (cherry picked from commit 8bda44ff37fd04a0fc9498fbbc22f0daf515d721)
221 commit 2675b0d159ff0608bc6e607be3727ed23e7e2976
222 Merge: a67f915 d535fab
223 Author: Sage Weil <sage@redhat.com>
224 Date: Fri Sep 19 11:30:17 2014 -0700
226 Merge pull request #2535 from dachary/wip-9470-pidfile-firefly
228 daemons: write pid file even when told not to daemonize (firefly)
230 Reviewed-by: Sage Weil <sage@redhat.com>
232 commit d535fab6ac4a025541340ded0d3568f565acf1cd
233 Author: Alexandre Oliva <oliva@gnu.org>
234 Date: Wed Jul 30 23:08:43 2014 -0300
236 daemons: write pid file even when told not to daemonize
238 systemd wants to run daemons in foreground, but daemons wouldn't write
239 out the pid file with -f. Fixed.
241 Signed-off-by: Alexandre Oliva <oliva@gnu.org>
242 (cherry picked from commit bccb0eb64891f65fd475e96b6386494044cae8c1)
244 commit c4685075f583ff58cc05220a3044a10b8392033c (refs/remotes/gh/wip-log-crash-firefly)
245 Author: Samuel Just <sam.just@inktank.com>
246 Date: Mon Sep 15 15:44:11 2014 -0700
248 PGLog::claim_log_and_clear_rollback_info: fix rollback_info_trimmed_to
250 We have been setting it to the old head value. This is usually
251 harmless since the new head will virtually always be ahead of the
252 old head for claim_log_and_clear_rollback_info, but can cause trouble
257 Signed-off-by: Samuel Just <sam.just@inktank.com>
258 (cherry picked from commit 0769310ccd4e0dceebd8ea601e8eb5c0928e0603)
260 commit 112f6f021bd063e2647169ba42729a447e5848e1
261 Merge: a67f915 92cfd37 07f54f2
262 Author: Samuel Just <sam.just@inktank.com>
263 Date: Thu Sep 18 09:46:38 2014 -0700
265 Merge remote-tracking branches 'origin/wip-9497' and 'origin/wip-9482' into wip-log-crash-firefly
267 commit 07f54f28df7429a88dd9243800281ecf23f10036
268 Author: Samuel Just <sam.just@inktank.com>
269 Date: Mon Sep 15 16:53:21 2014 -0700
271 PG::find_best_info: let history.last_epoch_started provide a lower bound
273 If we find a info.history.last_epoch_started above any
274 info.last_epoch_started, we must be missing updates and
275 min_last_update_acceptable should provisionally be max().
279 Signed-off-by: Samuel Just <sam.just@inktank.com>
281 commit 92cfd370395385ca5537b5bc72220934c9f09026
282 Author: Samuel Just <sam.just@inktank.com>
283 Date: Tue Sep 16 20:36:51 2014 -0700
285 PG::choose_acting: let the pg go down if acting is smaller than min_size
287 Even if the backfill peer would bring us up to min_size, we can't go
288 active since build_prior will not consider the interval maybe_went_rw.
292 Signed-off-by: Samuel Just <sam.just@inktank.com>
294 commit a67f9152f128d693d160e5b08f2ac8b8bc83e8e0
295 Author: Josh Durgin <josh.durgin@inktank.com>
296 Date: Thu Jul 24 15:29:40 2014 -0700
298 librbd: fix crash using clone of flattened image
300 The crash occurs due to ImageCtx->parent->parent being uninitialized,
301 since the inital open_parent() -> open_image(parent) ->
302 ictx_refresh(parent) occurs before ImageCtx->parent->snap_id is set,
303 so refresh_parent() is not called to open an ImageCtx for the parent
304 of the parent. This leaves the ImageCtx->parent->parent NULL, but the
305 rest of ImageCtx->parent updated to point at the correct parent snapshot.
307 Setting the parent->snap_id earlier has some unintended side effects
308 currently, so for now just call refresh_parent() during
309 open_parent(). This is the easily backportable version of the
310 fix. Further patches can clean up this whole initialization process.
313 Backport: firefly, dumpling
314 Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
315 (cherry picked from commit 2545e80d274b23b6715f4d8b1f4c6b96182996fb)
317 commit 2422f9fd634c239a1159c99aa4a49cfb5c5d097f
318 Author: JuanJose 'JJ' Galvez <jgalvez@redhat.com>
319 Date: Sun Sep 14 20:38:20 2014 -0700
321 init-radosgw.sysv: Support systemd for starting the gateway
323 When using RHEL7 the radosgw daemon needs to start under systemd.
325 Check for systemd running on PID 1. If it is then start
326 the daemon using: systemd-run -r <cmd>. pidof returns null
327 as it is executed too quickly, adding one second of sleep and
328 script reports startup correctly.
330 Signed-off-by: JuanJose 'JJ' Galvez <jgalvez@redhat.com>
331 (cherry picked from commit ddd52e87b25a6861d3b758a40d8b3693a751dc4d)
333 commit 6070383bc2c433e36f7ab5dc1ec09a3e7439d523
334 Merge: 668cd13 7b25512
335 Author: Sage Weil <sage@redhat.com>
336 Date: Fri Sep 12 17:31:03 2014 -0700
338 Merge pull request #2479 from ceph/wip-9444
340 mds: fix root and mdsdir inodes' rsubdirs
342 Reviewed-by: Sage Weil <sage@redhat.com>
344 commit 7b25512d65c6337cbddc7d6c6f55747ba1591be5
345 Author: Yan, Zheng <zheng.z.yan@intel.com>
346 Date: Fri May 2 23:08:41 2014 +0800
348 mds: fix root and mdsdir inodes' rsubdirs
350 inode rstat accounts inode itself.
352 Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com>
353 (cherry picked from commit da17394941386dab88ddbfed4af2c8cb6b5eb72f)
355 commit 668cd1359c94698bbb4f7b841e2f30264d4af937
356 Author: Samuel Just <sam.just@inktank.com>
357 Date: Tue Sep 9 14:03:50 2014 -0700
359 FileStore: report l_os_j_lat as commit latency
361 l_os_commit_lat is actually the commit cycle latency.
365 Signed-off-by: Samuel Just <sam.just@inktank.com>
366 (cherry picked from commit d165238b4ee7e925e06ca22890c1e9dac101a7da)
368 commit 7686966677867d098141aa08f45364eaa97121dd
369 Author: Samuel Just <sam.just@inktank.com>
370 Date: Tue Sep 9 12:58:07 2014 -0700
372 Objecter::_recalc_linger_op: resend for any acting set change
376 Signed-off-by: Samuel Just <sam.just@inktank.com>
377 (cherry picked from commit 1349383ac416673cb6df2438729fd2182876a7d1)
384 commit 3abf95456783346e6686919b803c4819d5e82ab7
385 Author: Sage Weil <sage@redhat.com>
386 Date: Mon Sep 8 13:44:57 2014 -0700
388 osdc/Objecter: revoke rx_buffer on op_cancel
390 If we cancel a read, revoke the rx buffers to avoid a use-after-free and/or
391 other undefined badness by using user buffers that may no longer be
395 Backport: firefly, dumpling
396 Reported-by: Matthias Kiefer <matthias.kiefer@1und1.de>
397 Signed-off-by: Sage Weil <sage@redhat.com>
398 (cherry picked from commit 2305b2897acba38384358c33ca3bbfcae6f1c74e)
400 (adjusted for op->con instead of s->con)
402 commit bc4108443a17de47b7c582e50fb884396a287152
403 Author: Sage Weil <sage@redhat.com>
404 Date: Mon Sep 8 13:45:52 2014 -0700
406 ceph_test_rados_api_io: add read timeout test
408 Verify we don't receive data after a timeout.
410 Based on reproducer for #9362 written by
411 Matthias Kiefer <matthias.kiefer@1und1.de>.
413 Signed-off-by: Sage Weil <sage@redhat.com>
414 (cherry picked from commit f295c1fee4afb9447cdf46f05a44234274d23b6c)
416 commit 19e7606545f829e5b9ea0255049d9130c51e42fd
417 Author: Sage Weil <sage@redhat.com>
418 Date: Mon Sep 8 13:42:43 2014 -0700
420 ceph_test_rados_api_*: expose nspace
422 Signed-off-by: Sage Weil <sage@redhat.com>
423 (cherry picked from commit 977d289055d69ab8a7baaf7ef68c013019225833)
425 commit 881c267fa4f8257a9205ec897c631183ef5c6388
426 Author: Samuel Just <sam.just@inktank.com>
427 Date: Tue Sep 9 12:40:51 2014 -0700
429 Revert "PG: mark_log_for_rewrite on resurrection"
431 Actually, we don't want to backport this one without the fix
434 This reverts commit 7ddf0a252bb887553b29fd93e58d01cac38835e6.
436 commit 3578b1193d497373db14d040ff7f0fc9e06e5ac2
437 Author: Samuel Just <sam.just@inktank.com>
438 Date: Wed Sep 3 15:49:47 2014 -0700
440 ReplicatedPG: create max hitset size
442 Otherwise, hit_set_create could create an unbounded size hitset
447 Signed-off-by: Samuel Just <sam.just@inktank.com>
449 commit d73eb895cdc792ddef2626643e61c1521cf53434
450 Author: Samuel Just <sam.just@inktank.com>
451 Date: Wed Aug 27 16:21:41 2014 -0700
453 PG::can_discard_op: do discard old subopreplies
455 Otherwise, a sub_op_reply from a previous interval can stick around
456 until we either one day go active again and get rid of it or delete the
457 pg which is holding it on its waiting_for_active list. While it sticks
458 around futily waiting for the pg to once more go active, it will cause
459 harmless slow request warnings.
463 Signed-off-by: Samuel Just <sam.just@inktank.com>
464 (cherry picked from commit ae3d87348ca4e2dde809c9593b0d54ce0469f7a0)
466 commit 7ddf0a252bb887553b29fd93e58d01cac38835e6
467 Author: Samuel Just <sam.just@inktank.com>
468 Date: Tue Aug 26 16:53:02 2014 -0700
470 PG: mark_log_for_rewrite on resurrection
474 Signed-off-by: Samuel Just <sam.just@inktank.com>
475 (cherry picked from commit 8346e10755027e982f26bab4642334fd91cc31aa)
477 commit 49a08eb46ef277225fff0d9a0920c37427175f54
478 Author: Thorsten Glaser <tg@mirbsd.de>
479 Date: Mon Sep 8 12:49:50 2014 -0700
481 debian: only B-R yasm on amd64
483 Make yasm dependency amd64 only, it isn?t used elsewhere
484 but breaks x32 (which is mis-detected as amd64)
486 Signed-off-by: Sage Weil <sage@redhat.com>
487 (cherry picked from commit 9ab46dc5b49219aa6194861c393c938f23001c52)
489 commit f19162adc241e627ce20140360352ac96f809a6e
490 Author: Sage Weil <sage@redhat.com>
491 Date: Tue Aug 26 17:43:10 2014 -0700
493 osd: fix osd_tp shutdown
495 We need to clear the queue, not just drain the currently executing jobs.
499 Signed-off-by: Sage Weil <sage@redhat.com>
500 (cherry picked from commit c2f21c04207b9a2a65e514994a775632b36d6874)
506 commit 598cde42e7f57fd0b294f70c0a930f94fd361fb3
507 Author: Sage Weil <sage@redhat.com>
508 Date: Wed Aug 27 06:19:12 2014 -0700
510 osd/PG: fix crash from second backfill reservation rejection
512 If we get more than one reservation rejection we should ignore them; when
513 we got the first we already sent out cancellations. More importantly, we
517 Backport: firefly, dumpling
518 Signed-off-by: Sage Weil <sage@redhat.com>
519 (cherry picked from commit 2b13de16c522754e30a0a55fb9d072082dac455e)
521 commit 0a9d49e5b5867c58ca1f5c6b9a4c5106639a8d01
522 Author: Sage Weil <sage@redhat.com>
523 Date: Mon Sep 8 06:58:45 2014 -0700
525 mon/Paxos: don't spam log with is_readable at dout level 1
527 Backport: firefly, dumpling
528 Reported-by: Aanchal Agrawal <Aanchal.Agrawal@sandisk.com>
529 Signed-off-by: Sage Weil <sage@redhat.com>
530 (cherry picked from commit 62ca27d0b119b597ebad40dde64c4d86599e466d)
532 commit 1660503ffafd69fac3722aea9915008113906e05
533 Author: Alfredo Deza <alfredo.deza@inktank.com>
534 Date: Thu Sep 4 13:58:14 2014 -0400
536 doc: add note on soft JS dependency for navigating docs
538 Signed-off-by: Alfredo Deza <alfredo.deza@inktank.com>
539 (cherry picked from commit 657be818375bea2d8b5998ea1e5505eedc2f294d)
541 commit 2f075d442f58e9cfefdbc421b9b20f2034a56314
542 Author: Alfredo Deza <alfredo.deza@inktank.com>
543 Date: Wed Sep 3 21:21:45 2014 -0400
545 doc: fix missing bracket
547 Signed-off-by: Alfredo Deza <alfredo.deza@inktank.com>
548 (cherry picked from commit 69638dfaeb0dcd96dac4b5f5c00ed08042432487)
550 commit 4dacb593e830f42cd0394486333315f1709b502f
551 Author: Alfredo Deza <alfredo.deza@inktank.com>
552 Date: Wed Sep 3 20:47:54 2014 -0400
554 doc: attempt to get the ayni JS into all head tags
556 Signed-off-by: Alfredo Deza <alfredo.deza@inktank.com>
557 (cherry picked from commit 35663fa55ac1579a3b0c8b67028a3a8dfea87b48)
559 commit 275df4407b37923cb47510890a686ed370e2f39e
560 Author: Dmitry Smirnov <onlyjob@member.fsf.org>
561 Date: Sat Aug 23 22:41:30 2014 +1000
563 Fix FTBFS on alpha due to incorrect check on BLKGETSIZE
565 Ceph FTBFS on Alpha with:
568 libtool: compile: g++ -DHAVE_CONFIG_H -I. -D__CEPH__ -D_FILE_OFFSET_BITS=64 -D_REENTRANT -D_THREAD_SAFE -D__STDC_FORMAT_MACROS -D_GNU_SOURCE -DCEPH_LIBDIR=\"/usr/lib/alpha-linux-gnu\" -DCEPH_PKGLIBDIR=\"/usr/lib/alpha-linux-gnu/ceph\" -DGTEST_HAS_TR1_TUPLE=0 -D_FORTIFY_SOURCE=2 -I/usr/include/nss -I/usr/include/nspr -Wall -Wtype-limits -Wignored-qualifiers -Winit-self -Wpointer-arith -Werror=format-security -fno-strict-aliasing -fsigned-char -rdynamic -ftemplate-depth-1024 -Wnon-virtual-dtor -Wno-invalid-offsetof -Wstrict-null-sentinel -g -O2 -Wformat -Werror=format-security -c common/blkdev.cc -fPIC -DPIC -o common/.libs/blkdev.o
569 In file included from /usr/include/alpha-linux-gnu/asm/ioctls.h:4:0,
570 from /usr/include/alpha-linux-gnu/bits/ioctls.h:23,
571 from /usr/include/alpha-linux-gnu/sys/ioctl.h:26,
572 from common/blkdev.cc:3:
573 common/blkdev.cc:13:7: error: missing binary operator before token "int"
578 This error occurs because the value of BLKGETSIZE is tested in a
579 c-preprocessor conditional compilation test whereas the test should
582 From: Michael Cree <mcree@orcon.net.nz>
583 Bug-Debian: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=756892
584 Signed-off-by: Dmitry Smirnov <onlyjob@member.fsf.org>
585 (cherry picked from commit 6ad8e61a428cfc9fc60ccdb9bce812e1f49822ac)
586 Reviewed-by: Greg Farnum <greg@inktank.com>
588 commit fbbdc31a8cd4c6457e06570ddeb9b2021563498e
589 Merge: 46d9611 895bddc
590 Author: Sage Weil <sage@redhat.com>
591 Date: Fri Aug 29 17:31:29 2014 -0700
593 Merge pull request #2356 from dachary/wip-9273-mon-preload-erasure-code-firefly
595 erasure-code: preload the default plugins in the mon (firefly)
597 Reviewed-by: Sage Weil <sage@redhat.com>
599 commit 46d9611aedd3194cfd55a338e5f94011e6e70668
600 Author: Joao Eduardo Luis <joao.luis@inktank.com>
601 Date: Fri Aug 29 20:21:25 2014 +0100
603 osd: OSDMap: ordered blacklist on non-classic encode function
608 Signed-off-by: Joao Eduardo Luis <joao.luis@inktank.com>
609 Reviewed-by: Sage Weil <sage@redhat.com>
610 (cherry picked from commit 81102044f417bd99ca570d9234b1df5195e9a8c9)
612 commit 994a9e35e96984c0806261f7eb349d95c0fb4873
613 Author: Sage Weil <sage@redhat.com>
614 Date: Tue Aug 26 08:16:29 2014 -0700
616 osd/OSDMap: encode blacklist in deterministic order
618 When we use an unordered_map the encoding order is non-deterministic,
619 which is problematic for OSDMap. Construct an ordered map<> on encode
620 and use that. This lets us keep the hash table for lookups in the general
625 Signed-off-by: Sage Weil <sage@redhat.com>
626 (cherry picked from commit 4672e50922b75d642056020b9745a3a5844424d3)
628 commit 895bddc5d4dabd2f05c13c343eb0d7457d6f132f
629 Author: Loic Dachary <loic-201408@dachary.org>
630 Date: Fri Aug 29 18:13:08 2014 +0200
632 erasure-code: preload the default plugins in the mon
634 The commit 164f1a1959a863848319585fa752250c7b261381 preloads the
635 jerasure plugin in the OSD. They must also be preloaded in the mon for
638 http://tracker.ceph.com/issues/9273 Fixes: #9273
640 Signed-off-by: Loic Dachary <loic-201408@dachary.org>
642 commit db8d6e6e190de964ec3d5d1bbc87da2dcb3213f5
643 Author: John Spray <john.spray@redhat.com>
644 Date: Tue Aug 26 17:36:16 2014 +0100
646 mds: fix FP error in ROUND_UP_TO
648 Explicitly handle case where denominator is 0, instead of
649 passing into ROUND_UP_TO.
651 Regression from 9449520b121fc6ce0c64948386d4ff77f46f4f5f
653 Signed-off-by: John Spray <john.spray@redhat.com>
654 (cherry picked from commit bf3e4835dabc057982def1b5c9a6499c04ac5312)
656 commit 8e3120fcb379a00d370e4c04d34af35e596e2de9
657 Author: Sage Weil <sage@redhat.com>
658 Date: Thu Aug 21 11:14:39 2014 -0700
660 mon: generate cluster_fingerprint if null
662 This triggers after an upgrade of a legacy cluster that has no fingerprint.
664 Signed-off-by: Sage Weil <sage@redhat.com>
665 (cherry picked from commit b245d600163f6337af15aedd1fea68f4e2a668a8)
667 commit ebcdeb4cfe201dd0c630386226f9970650689ccf
668 Author: Sage Weil <sage@redhat.com>
669 Date: Wed Aug 20 08:59:46 2014 -0700
671 mon: add a cluster fingerprint
673 Generate it on cluster creations with the initial monmap. Include it in
674 the report. Provide no way for this uuid to be fed in to the cluster
675 (intentionally or not) so that it can be assumed to be a truly unique
676 identifier for the cluster.
678 Signed-off-by: Sage Weil <sage@redhat.com>
679 (cherry picked from commit 675b0042eff0ad5e1453838410210b1206c39004)
681 commit b3b029165d4ac0af3f2143fad205f4e8ab08a30f
682 Merge: 21e1faa 07e1ceb
683 Author: Sage Weil <sage@redhat.com>
684 Date: Tue Aug 26 13:13:08 2014 -0700
686 Merge pull request #2244 from dachary/wip-9044-use-ruleset-firefly
688 erasure-code: OSDMonitor::crush_ruleset_create_erasure needs ruleset (firefly)
690 commit 21e1faa75da4e9cab8e139a8e2a1c4fc0ee30867
691 Author: Samuel Just <sam.just@inktank.com>
692 Date: Tue Aug 12 16:41:38 2014 -0700
694 ReplicatedPG::cancel_copy: clear cop->obc
696 Otherwise, an objecter callback might still be hanging
697 onto this reference until after the flush.
700 Introduced: 589b639af7c8834a1e6293d58d77a9c440107bc3
701 Signed-off-by: Samuel Just <sam.just@inktank.com>
702 (cherry picked from commit 5040413054e923d6d5a2b4928162dba140d980e0)
704 commit e6da7323655f3bfa4f669fb8361fcaa96f633456
705 Author: Samuel Just <sam.just@inktank.com>
706 Date: Tue Aug 26 12:02:52 2014 -0700
708 PG: recover from each osd at most once
710 Signed-off-by: Samuel Just <sam.just@inktank.com>
712 commit 8a1723f67f995253c9e6da95ab433743ef23f61c
713 Author: Samuel Just <sam.just@inktank.com>
714 Date: Tue Aug 26 11:38:53 2014 -0700
716 PG: make the reservation sets more descriptively named
718 These sets won't precisely be the backfill_targets or actingbackfill
721 %s/sorted_backfill_set/remote_shards_to_reserve_backfill/g
722 %s/acting_osd_it/remote_recovery_reservation_it/g
723 %s/sorted_actingbackfill_set/remote_shards_to_reserve_recovery/g
725 Signed-off-by: Samuel Just <sam.just@inktank.com>
727 commit 938d735697e2af0b01dd2613a9152b3caef5db57
728 Merge: 2f11631 9ca451a
729 Author: Samuel Just <sam.just@inktank.com>
730 Date: Tue Aug 26 10:30:14 2014 -0700
732 Merge pull request #2203 from ceph/wip-scrub-firefly
734 backport scrub throttling to firefly
736 Reviewed-by: Samuel Just <sam.just@inktank.com>
738 commit 2f11631f3144f2cc0e04d718e40e716540c8af19
739 Author: Sage Weil <sage@redhat.com>
740 Date: Sat Aug 16 12:42:33 2014 -0700
742 os/FileStore: fix mount/remount force_sync race
747 - sync_entry is doing some work
749 - set force_sync = true
751 - sync_entry exits (due to done)
752 - ..but does not set force_sync = false
754 - journal replay starts
755 - sync_entry sees force_sync and does a commit while op_seq == 0
759 Backport: firefly, dumpling
760 Signed-off-by: Sage Weil <sage@redhat.com>
761 (cherry picked from commit dd11042f969b94f7a461d02e1475794031c79f61)
766 commit a38cf1bd503a09843d439f65937e7d60aed163e2
767 Author: Haomai Wang <haomaiwang@gmail.com>
768 Date: Thu Jul 10 10:32:17 2014 +0800
770 Add random_cache.hpp to Makefile.am
772 Signed-off-by: Haomai Wang <haomaiwang@gmail.com>
773 (cherry picked from commit a3e5c6d632119febd2150944a6f2cbce33cfda3a)
775 commit a2c1532b5610207eccd24b8253643eec6bb05294
776 Author: Sage Weil <sage@redhat.com>
777 Date: Tue Aug 26 06:42:12 2014 -0700
779 os/KeyValueStore, MemStore: fix warning
781 os/MemStore.cc: In member function 'void MemStore::_do_transaction(ObjectStore::Transaction&)':
782 os/MemStore.cc:956:18: warning: unused variable 'expected_object_size' [-Wunused-variable]
783 os/MemStore.cc:957:18: warning: unused variable 'expected_write_size' [-Wunused-variable]
784 os/KeyValueStore.cc: In member function 'unsigned int KeyValueStore::_do_transaction(ObjectStore::Transaction&, KeyValueStore::BufferTransaction&, ThreadPool::TPHandle*)':
785 os/KeyValueStore.cc:1426:18: warning: unused variable 'expected_object_size' [-Wunused-variable]
786 os/KeyValueStore.cc:1427:18: warning: unused variable 'expected_write_size' [-Wunused-variable]
788 Signed-off-by: Sage Weil <sage@redhat.com>
790 commit 2b86bf4f4c1be2603368d5031ee7d706056e2d1e
791 Author: Sage Weil <sage@inktank.com>
792 Date: Tue Apr 29 11:23:58 2014 -0700
794 osd: automatically scrub PGs with invalid stats
796 If a PG has recnetly split and has invalid stats, scrub it now, even if
797 it has scrubbed recently. This helps the stats become valid again soon.
800 Signed-off-by: Sage Weil <sage@inktank.com>
801 (cherry picked from commit 68b440d66539e820c9ce86a6942c3188be4ee1ec)
803 commit 271ca7d9bc41e3a2ac387b05ca9f30d8a250fb62
804 Merge: 54333a4 87cd3a8
805 Author: Sage Weil <sage@redhat.com>
806 Date: Tue Aug 26 06:38:34 2014 -0700
808 Merge pull request #2328 from dachary/wip-9209-round-up-to-firefly
810 common: ROUND_UP_TO accepts any rounding factor (firefly)
812 commit 54333a49c9e99a398bf3695845b248ba76ee3930
813 Merge: 4bff6f3 58d0871
814 Author: Sage Weil <sage@redhat.com>
815 Date: Tue Aug 26 06:09:17 2014 -0700
817 Merge pull request #2326 from yuyuyu101/wip-kvstore-firefly
819 Backport from master to Firefly(KeyValueStore)
821 commit 87cd3a8f6e190aed06e361595afc4ec83148da98
822 Author: Loic Dachary <loic-201408@dachary.org>
823 Date: Mon Aug 25 17:05:04 2014 +0200
825 common: ROUND_UP_TO accepts any rounding factor
827 The ROUND_UP_TO function was limited to rounding factors that are powers
828 of two. This saves a modulo but it is not used where it would make a
829 difference. The implementation is changed so it is generic.
831 http://tracker.ceph.com/issues/9209 Fixes: #9209
833 Signed-off-by: Loic Dachary <loic-201408@dachary.org>
834 (cherry picked from commit 9449520b121fc6ce0c64948386d4ff77f46f4f5f)
836 commit 58d08714340049a5165ad682ec5b54292525b45d
837 Author: Haomai Wang <haomaiwang@gmail.com>
838 Date: Thu Mar 20 14:09:49 2014 +0800
840 Remove exclusive lock on GenericObjectMap
842 Now most of GenericObjectMap interfaces use header as argument not the union of
843 coll_t and ghobject_t. So caller should be responsible for maintain the
846 Signed-off-by: Haomai Wang <haomaiwang@gmail.com>
848 commit 98df982a6f830154ff6880f5ee00350ba3375b6b
849 Author: Haomai Wang <haomaiwang@gmail.com>
850 Date: Wed Jul 23 11:26:18 2014 +0800
852 common/RandomCache: Fix inconsistence between contents and count
854 The add/clear method may cause count inconsistent with the real size of
857 Signed-off-by: Haomai Wang <haomaiwang@gmail.com>
859 commit c8ec7721526ccf4bfd91a220fee0dd75bef8fb21
860 Author: Haomai Wang <haomaiwang@gmail.com>
861 Date: Tue Aug 26 04:41:28 2014 +0000
863 Add random cache and replace SharedLRU in KeyValueStore
865 SharedLRU plays pool performance in KeyValueStore with large header cache size,
866 so a performance optimized RandomCache could improve it.
868 RandomCache will record the lookup frequency of key. When evictint element,
869 it will randomly compare several elements's frequency and evict the least
872 Signed-off-by: Haomai Wang <haomaiwang@gmail.com>
876 src/common/config_opts.h
877 src/os/KeyValueStore.cc
879 commit 33ca26a103a0d99627fddc30f2916c5aeb37c691
880 Author: Haomai Wang <haomaiwang@gmail.com>
881 Date: Tue Aug 26 04:40:16 2014 +0000
883 Add Header cache to KeyValueStore
885 In the performance statistic recently, the header lookup becomes the main time
886 consuming for the read/write operations. Most of time it occur 50% to deal with
887 header lookup, decode/encode logics.
889 Now adding header cache using SharedLRU structure which will maintain the header
890 cache and caller will get the pointer to the real header. It also avoid too much
891 header copy operations overhead.
893 Signed-off-by: Haomai Wang <haomaiwang@gmail.com>
897 src/os/KeyValueStore.cc
898 src/os/KeyValueStore.h
900 commit e8dcb8cc603cff64bafc11ee42c5c04a9cf85d12
901 Author: Haomai Wang <haomaiwang@gmail.com>
902 Date: Wed Feb 26 17:46:07 2014 +0800
904 Fix write operation on a deleted object in the same transaction
906 If the following op happened:
911 KeyValueStore will fail at "write" operation.
913 Signed-off-by: Haomai Wang <haomaiwang@gmail.com>
915 commit e0ae6f9de685b60eee5efa5ca960c7a15481e41a
916 Author: Haomai Wang <haomaiwang@gmail.com>
917 Date: Tue Aug 26 04:35:57 2014 +0000
919 Remove SequencerPosition from KeyValueStore
921 Now KeyValueStore expects kv backend to ensure consistency and there is unusable
922 for KeyValueStore to store a SequencerPosition.
924 Signed-off-by: Haomai Wang <haomaiwang@gmail.com>
928 src/os/KeyValueStore.cc
929 src/os/KeyValueStore.h
931 commit 44721f71266e3098169ed00f1ca7e86496b76b07
932 Author: Haomai Wang <haomaiwang@gmail.com>
933 Date: Wed Jun 4 12:58:07 2014 +0800
935 Fix keyvaluestore fiemap bug
937 The result of fiemap is wrong and the offset get from
938 "StripObjectMap::file_to_extents" need to multiply by sequence number
940 Signed-off-by: Haomai Wang <haomaiwang@gmail.com>
942 commit 4bff6f3476e5df400db3dc1f0ad7b35a2a41830e
943 Author: Sage Weil <sage@redhat.com>
944 Date: Thu Aug 21 13:05:35 2014 -0700
946 mon: fix occasional message leak after session reset
950 - we get a message, put it on a wait list
951 - the client session resets
952 - we go back to process the message later and discard
953 - _ms_dispatch returns false, but nobody drops the msg ref
955 Since we call _ms_dispatch() a lot internally, we need to always return
956 true when we are an internal caller.
959 Backport: firefly, dumpling
960 Signed-off-by: Sage Weil <sage@redhat.com>
961 (cherry picked from commit 19df386b2d36d716be2e6d02de0386fac9e7bc1f)
963 commit ca3ac907aa73c5d77760e12da6f1509ea860e894
964 Merge: fdbab46 8d7e77b
965 Author: Sage Weil <sage@redhat.com>
966 Date: Thu Aug 21 10:14:18 2014 -0700
968 Merge pull request #2298 from dachary/wip-9153-jerasure-upgrade-firefly
970 erasure-code: preload the jerasure plugin variant (sse4,sse3,generic)
972 Reviewed-by: Sage Weil <sage@redhat.com>
974 commit 8d7e77b9747f1b3dc3c10d1b0877a4b9c899ba86
975 Author: Loic Dachary <loic-201408@dachary.org>
976 Date: Thu Aug 21 14:41:55 2014 +0200
978 erasure-code: preload the jerasure plugin variant (sse4,sse3,generic)
980 The preloading of the jerasure plugin ldopen the plugin that is in
981 charge of selecting the variant optimized for the
982 CPU (sse4,sse3,generic). The variant plugin itself is not loaded because
983 it does not happen at load() but when the factory() method is called.
985 The JerasurePlugin::preload method is modified to call the factory()
986 method to load jerasure_sse4 or jerasure_sse3 or jerasure_generic as a
989 Indirectly loading another plugin in the factory() method is error prone
990 and should be moved to the load() method instead. This change should be
991 done in a separate commit.
993 http://tracker.ceph.com/issues/9153 Fixes: #9153
995 Signed-off-by: Loic Dachary <loic-201408@dachary.org>
997 commit fdbab46852e74d405b5c747da98564a5866ec8a7
998 Author: Haomai Wang <haomaiwang@gmail.com>
999 Date: Tue May 20 14:32:18 2014 +0800
1001 Fix set_alloc_hint op cause KeyValueStore crash problem
1003 Now KeyValueStore doesn't support set_alloc_hit op, the implementation of
1004 _do_transaction need to consider decoding the arguments. Otherwise, the
1005 arguments will be regarded as the next op.
1007 Fix the same problem for MemStore.
1011 Reported-by: Xinxin Shu <xinxin.shu5040@gmail.com>
1012 Signed-off-by: Haomai Wang <haomaiwang@gmail.com>
1013 (cherry picked from commit c08adbc98ff5f380ecd215f8bd9cf3cab214913c)
1015 commit 922e24e658f09a8a772309f18a8d54c560fad653
1016 Merge: ae787cf 164f1a1
1017 Author: Sage Weil <sage@redhat.com>
1018 Date: Wed Aug 20 10:10:08 2014 -0700
1020 Merge pull request #2286 from dachary/wip-9153-jerasure-upgrade-firefly
1022 erasure-code: preload the jerasure plugin (firefly)
1024 Reviewed-by: Sage Weil <sage@redhat.com>
1026 commit 164f1a1959a863848319585fa752250c7b261381
1027 Author: Loic Dachary <loic-201408@dachary.org>
1028 Date: Tue Aug 19 01:30:15 2014 +0200
1030 erasure-code: preload the jerasure plugin
1032 Load the jerasure plugin when ceph-osd starts to avoid the following
1035 * ceph-osd-v1 is running but did not load jerasure
1037 * ceph-osd-v2 is installed being installed but takes time : the files
1038 are installed before ceph-osd is restarted
1040 * ceph-osd-v1 is required to handle an erasure coded placement group and
1041 loads jerasure (the v2 version which is not API compatible)
1043 * ceph-osd-v1 calls the v2 jerasure plugin and does not reference the
1044 expected part of the code and crashes
1046 Although this problem shows in the context of teuthology, it is unlikely
1047 to happen on a real cluster because it involves upgrading immediately
1048 after installing and running an OSD. Once it is backported to firefly,
1049 it will not even happen in teuthology tests because the upgrade from
1050 firefly to master will use the firefly version including this fix.
1052 While it would be possible to walk the plugin directory and preload
1053 whatever it contains, that would not work for plugins such as jerasure
1054 that load other plugins depending on the CPU features, or even plugins
1055 such as isa which only work on specific CPU.
1057 http://tracker.ceph.com/issues/9153 Fixes: #9153
1060 Signed-off-by: Loic Dachary <loic-201408@dachary.org>
1061 (cherry picked from commit 9b802701f78288ba4f706c65b853415c69002d27)
1064 src/test/erasure-code/test-erasure-code.sh
1065 src/common/config_opts.h
1067 commit ae787cfa88dfd0f5add5932b297258c46af4e333
1068 Author: Matt Benjamin <matt@linuxbox.com>
1069 Date: Thu May 29 10:34:20 2014 -0400
1071 Work around an apparent binding bug (GCC 4.8).
1073 A reference to h->seq passed to std::pair ostensibly could not bind
1074 because the header structure is packed. At first this looked like
1075 a more general unaligned access problem, but the only location the
1076 compiler rejects is a false positive.
1078 Signed-off-by: Matt Benjamin <matt@linuxbox.com>
1079 (cherry picked from commit c930a1f119069a424af28a618b0abff4947c221f)
1081 commit 486deefdc1496b72cc680e432a61d71e5fa265c9
1082 Author: Sage Weil <sage@redhat.com>
1083 Date: Sun Aug 17 20:54:28 2014 -0700
1085 qa/workunits/rbd/qemu-iotests: touch common.env
1087 This seems to be necessary on trusty.
1089 Backport: firefly, dumpling
1090 Signed-off-by: Sage Weil <sage@redhat.com>
1091 (cherry picked from commit 055be68cf8e1b84287ab3631a02e89a9f3ae6cca)
1093 commit cc4e6258d67fb16d4a92c25078a0822a9849cd77
1094 Author: Joao Eduardo Luis <joao.luis@inktank.com>
1095 Date: Fri May 23 16:52:08 2014 +0100
1097 unittest_strtol: fix compilation warning
1099 Was fixed in master by a4923f5bc373d530d1ffdf6c58a4d88139daedd2
1101 Signed-off-by: Sage Weil <sage@redhat.com>
1103 commit 7f9fe22a1c73d5f2783c3303fb9f3a7cfcea61c5
1104 Author: huangjun <hjwsm1989@gmail.com>
1105 Date: Tue Jun 17 13:12:58 2014 +0800
1107 Fix EINVAL err when use "ceph tell osd.* bench"
1109 Signed-off-by: huangjun <hjwsm1989@gmail.com>
1110 (cherry picked from commit 7dc93a9651f602d9c46311524fc6b54c2f1ac595)
1112 commit 14607fc045aa169f1e8fa6872b412fba8f09d645
1113 Author: Ma Jianpeng <jianpeng.ma@intel.com>
1114 Date: Wed Jul 16 17:48:34 2014 -0700
1116 qa/workunits/cephtool/test.sh: fix get erasure_code_profile test
1118 Manual backport of 4d6899c7560e990650959b442980a7249f0ba4c1
1120 Signed-off-by: Sage Weil <sage@redhat.com>
1122 commit 068b906ee396f1ad717968f214610ea86dd8d6b4
1123 Author: Joao Eduardo Luis <joao.luis@inktank.com>
1124 Date: Fri Jun 27 21:41:18 2014 +0100
1126 mon: OSDMonitor: add 'osd pool get-quota' command
1128 Enables us to obtain current quotas for a given pool.
1132 Signed-off-by: Joao Eduardo Luis <joao.luis@inktank.com>
1133 (cherry picked from commit 714a9bb5a058b2553f3be3e4cfb7e7f30150e75a)
1135 commit 2ed0fa034aa8e49ab703bdb798bd6211a83e3a9d
1136 Author: John Spray <jspray@redhat.com>
1137 Date: Tue Jun 3 10:12:41 2014 +0100
1139 mon: name instead of id in "has tiers" message
1141 Instead of "Pool foo has tiers 1 2" print
1142 "Pool foo has tiers bar baz".
1144 Signed-off-by: John Spray <jspray@redhat.com>
1145 (cherry picked from commit 97772c2f53f726bd71710d0d3e34159d2679390a)
1147 commit 7e509b1b8b5a3679094f794cd9334598aef4441e
1148 Author: Joao Eduardo Luis <joao.luis@inktank.com>
1149 Date: Fri May 23 17:01:38 2014 +0100
1151 common/config.cc: allow integer values to be parsed as SI units
1153 We are allowing this for all and any integer values; that is, OPT_INT,
1154 OPT_LONGLONG, OPT_U32 and OPT_U64.
1156 It's on the user to use appropriate units. For instance, the user should
1157 not use 'E(xabyte)' when setting a signed int, and use his best judgment
1158 when setting options that, for instance, ought to receive seconds.
1161 Signed-off-by: Joao Eduardo Luis <joao.luis@inktank.com>
1162 (cherry picked from commit 5500437e064cd6b4b45d63ee9396193df87f4d44)
1164 commit 4e34d83e337d0febf20c86b051721878feef49a8
1165 Author: Joao Eduardo Luis <joao.luis@inktank.com>
1166 Date: Fri May 23 16:52:08 2014 +0100
1168 test/strtol.cc: Test 'strict_strtosi()'
1170 Signed-off-by: Joao Eduardo Luis <joao.luis@inktank.com>
1171 (cherry picked from commit 40587d4792fd55db72d33870aae8b6a806c9baaf)
1173 commit 3b06b8f4893bd18440ecb9705414831a827cef90
1174 Author: Joao Eduardo Luis <joao.luis@inktank.com>
1175 Date: Fri May 23 16:51:37 2014 +0100
1177 common/strtol.cc: strict_strtosi() converts str with SI units to uint64_t
1179 Accepts values with a suffix (B, K, M, G, T, P, E) and returns the
1180 appropriate byte value.
1182 E.g., 10B = 10, while 10K = 10240.
1184 Signed-off-by: Joao Eduardo Luis <joao.luis@inktank.com>
1185 (cherry picked from commit 67dc5751ba9a4e527ff12ea65000d1ba45d956f6)
1187 commit 92ab6ecf321ab68ea8fe3e8c2b310303a5726af3
1188 Author: Alfredo Deza <alfredo.deza@inktank.com>
1189 Date: Wed Aug 13 15:50:20 2014 -0400
1191 ceph-disk: linter cleanup
1193 Signed-off-by: Alfredo Deza <alfredo.deza@inktank.com>
1194 (cherry picked from commit d74ed9d53fab95f27a9ad8e9f5dab7192993f6a3)
1196 commit 08772fd888fb0eca4570bd6b4bb25f8122691cb4
1197 Author: Sage Weil <sage@redhat.com>
1198 Date: Wed Aug 13 12:00:50 2014 -0700
1200 ceph-disk: warn about falling back to sgdisk (once)
1202 This way the user knows something funny might be up if dmcrypt is in use.
1204 Signed-off-by: Sage Weil <sage@redhat.com>
1205 (cherry picked from commit 6f7798e37e098de38fbc73f86c4c6ee705abbe38)
1207 commit 50166efd205f46fa325dec9636d817387e5d4d3b
1208 Author: Sage Weil <sage@redhat.com>
1209 Date: Wed Aug 13 11:40:34 2014 -0700
1211 ceph-disk: only fall back to sgdisk for 'list' if blkid seems old
1213 If the blkid doesn't show us any ID_PART_ENTRY_* fields but we know it is
1214 a GPT partition, *then* fallback. Otherwise, don't bother.
1216 Signed-off-by: Sage Weil <sage@redhat.com>
1217 (cherry picked from commit b1651afb34d9d2c324db3bf5f54ac9ce001c6af9)
1219 commit a45e815cb00f1a807691cbf687990839995f7629
1220 Author: Sage Weil <sage@redhat.com>
1221 Date: Wed Aug 13 11:39:47 2014 -0700
1223 ceph-disk: add get_partition_base() helper
1225 Return the base devices/disk for a partition device.
1227 Signed-off-by: Sage Weil <sage@redhat.com>
1228 (cherry picked from commit b75e8a340c49cbc067baa19790b994a5f904bb4f)
1230 commit 8c04e475efb4471c605e3c921ec8a03b67934899
1231 Author: Sage Weil <sage@redhat.com>
1232 Date: Tue Aug 12 17:26:07 2014 -0700
1234 ceph-disk: display information about dmcrypted data and journal volumes
1236 Signed-off-by: Sage Weil <sage@redhat.com>
1237 (cherry picked from commit c7a1ceba441fa99a82e19ed2cd3c6782a5d77636)
1239 commit b09e659226b26002b58c7697f8c600d06fe78c58
1240 Author: Sage Weil <sage@redhat.com>
1241 Date: Tue Aug 12 17:25:42 2014 -0700
1243 ceph-disk: move fs mount probe into a helper
1245 Signed-off-by: Sage Weil <sage@redhat.com>
1246 (cherry picked from commit f80ed26d2403ba12e80da6459fc45c22584f72de)
1248 commit 6775f46e4a5862ddbf81bef11016bc036222fda9
1249 Author: Sage Weil <sage@redhat.com>
1250 Date: Tue Aug 12 17:25:10 2014 -0700
1252 ceph-disk: use partition type UUIDs, and blkid
1254 Use blkid to give us the GPT partition type. This lets us distinguish
1255 between dmcrypt and non-dmcrypt partitions. Fake it if blkid doesn't
1256 give us what we want and try with sgdisk. This isn't perfect (it can't
1257 tell between dmcrypt and not dmcrypt), but such is life, and we are better
1260 Signed-off-by: Sage Weil <sage@redhat.com>
1261 (cherry picked from commit 6c77f5f2f994c881232d76ce9c69af80d10772bd)
1263 commit 1804a83400b7c3dc07650e09da5224bb999d0940
1264 Author: Sage Weil <sage@redhat.com>
1265 Date: Tue Aug 12 13:53:16 2014 -0700
1267 ceph-disk: fix log syntax error
1269 File "/usr/sbin/ceph-disk", line 303, in command_check_call
1270 LOG.info('Running command: %s' % ' '.join(arguments))
1271 TypeError: sequence item 2: expected string, NoneType found
1274 Signed-off-by: Sage Weil <sage@redhat.com>
1275 (cherry picked from commit 1088d6cd11b476cd67ed30e07edd363c4057a003)
1277 commit 8a8a9c35349d3f58d899afa1a9e219b99586bc9f
1278 Author: Sage Weil <sage@redhat.com>
1279 Date: Mon Aug 11 15:58:15 2014 -0700
1281 Revert "Fix for bug #6700"
1283 This reverts commit 673394702b725ff3f26d13b54d909208daa56d89.
1285 This appears to break things when the journal and data disk are *not* the same.
1286 And I can't seem to reproduce the original failure...
1288 Signed-off-by: Sage Weil <sage@redhat.com>
1289 (cherry picked from commit 2edf01ffa4a7425af2691b4e94bc5fd0bfab1e5b)
1291 commit 50b700fadf6ce6e0bf14b5daa2288786426d7359
1292 Author: Sage Weil <sage@redhat.com>
1293 Date: Mon Aug 11 15:57:52 2014 -0700
1295 ceph-disk: fix verify_no_in_use check
1297 We only need to verify that partitions aren't in use when we want to
1298 consume the whole device (osd data), not when we want to create an
1299 additional partition for ourselves (osd journal).
1302 Signed-off-by: Sage Weil <sage@redhat.com>
1303 (cherry picked from commit d6e6ba198efc4b3afff0c70af53497a70c6b3f19)
1305 commit c9847ef059f182ad15ef027c8bdfae6c99d91867
1306 Author: Alfredo Deza <alfredo@deza.pe>
1307 Date: Thu May 22 17:04:28 2014 -0400
1309 better error reporting on incompatible device requirements
1311 Signed-off-by: Alfredo Deza <alfredo@deza.pe>
1312 (cherry picked from commit 1ac3a503a15ddf7f7c1a33310a468fac10a1b7b6)
1314 commit 40649902891244242d5861ccd9538aec026cae3a
1315 Author: Stuart Longland <stuartl@vrt.com.au>
1316 Date: Tue May 6 14:06:36 2014 -0700
1318 ceph-disk: fix list for encrypted or corrupt volume
1320 Continue gracefully if an fs type is not detected, either because it is
1321 encrypted or because it is corrupted.
1323 Signed-off-by: Sage Weil <sage@inktank.com>
1324 (cherry picked from commit 09beebe3f1fd1b179547743648049b891cb8bc56)
1326 commit 80896dcf2c3ccba0b9b0b1bd9f6e075585e3d502
1327 Author: Alfredo Deza <alfredo.deza@inktank.com>
1328 Date: Fri Jun 13 09:37:33 2014 -0400
1330 support dmcrypt partitions when activating
1332 Signed-off-by: Alfredo Deza <alfredo.deza@inktank.com>
1333 (cherry picked from commit ef8a1281512c4ee70a3764b28891da691a183804)
1335 commit 1b0da81d733a36de4a9dc80263058a6cdf05c745
1336 Author: Sage Weil <sage@redhat.com>
1337 Date: Fri Aug 15 16:41:43 2014 -0700
1339 init-ceph: don't use bashism
1342 the length of STRING is zero
1344 Signed-off-by: Sage Weil <sage@redhat.com>
1345 (cherry picked from commit 0d6d1aa7e0c5e0b5f99c9b548a1f890c511b4299)
1347 commit 9005f33d0d80d6d2b93493b38866b7f74f87d1eb
1348 Author: Sage Weil <sage@redhat.com>
1349 Date: Fri Aug 15 14:28:57 2014 -0700
1351 osd: fix feature requirement for mons
1353 These features should be set on the client_messenger, not
1357 Signed-off-by: Sage Weil <sage@redhat.com>
1358 (cherry picked from commit ae0b9f17760eda9a7e09a6babac50bfe8ebb4b36)
1360 commit 9e46c5ad7dadd1c342c82629577b5a1bf037fe75
1361 Author: Sage Weil <sage@redhat.com>
1362 Date: Fri Aug 15 13:54:11 2014 -0700
1364 unittest_osdmap: test EC rule and pool features
1366 TODO: tiering feature bits.
1368 Signed-off-by: Sage Weil <sage@redhat.com>
1369 (cherry picked from commit 2f0e2951d773b6acce781b4b991d6d8e817ee2f9)
1371 commit dbbe0c627bcd4a60c1346bb120698b5a0c2192c2
1372 Author: Sage Weil <sage@redhat.com>
1373 Date: Fri Aug 15 14:04:05 2014 -0700
1375 unittest_osdmap: create an ec pool in test osdmap
1377 This is part of 7294e8c4df6df9d0898f82bb6e0839ed98149310.
1379 Signed-off-by: Sage Weil <sage@redhat.com>
1381 commit c2aa74a22718d9a59d60201d71213aaff2492ff8
1382 Author: Sage Weil <sage@redhat.com>
1383 Date: Fri Aug 15 08:55:10 2014 -0700
1385 osd: only require crush features for rules that are actually used
1387 Often there will be a CRUSH rule present for erasure coding that uses the
1388 new CRUSH steps or indep mode. If these rules are not referenced by any
1389 pool, we do not need clients to support the mapping behavior. This is true
1390 because the encoding has not changed; only the expected CRUSH output.
1394 Signed-off-by: Sage Weil <sage@redhat.com>
1395 (cherry picked from commit 16dadb86e02108e11a970252411855d84ab0a4a2)
1397 commit 12430fed85b2cfeaeb33a443e6a4c6538d0833f8
1398 Author: Sage Weil <sage@redhat.com>
1399 Date: Fri Aug 15 08:52:37 2014 -0700
1401 crush: add is_v[23]_rule(ruleid) methods
1403 Add methods to check if a *specific* rule uses v2 or v3 features. Refactor
1404 the existing checks to use these.
1406 Signed-off-by: Sage Weil <sage@redhat.com>
1407 (cherry picked from commit 1d95486780a54c85a5c88936a4da4bdc3576a7b8)
1409 commit 2427ea8f153caf55b3bce77cd2fc88935563c7b3
1410 Author: Samuel Just <sam.just@inktank.com>
1411 Date: Mon Jun 30 13:40:07 2014 -0700
1413 PGLog: fix clear() to avoid the IndexLog::zero() asserts
1416 c5b8d8105d965da852c79add607b69d5ae79a4d4
1417 ac11ca40b4f4525cbe9b1778b1c5d9472ecb9efa
1418 Signed-off-by: Samuel Just <sam.just@inktank.com>
1420 (cherry picked from commit 959f2b25910360b930183fbf469ce984a48542dd)
1422 commit 9ca451a6bf2f29e75f72cbd3dbcff0bf8986e462
1423 Author: Sage Weil <sage@redhat.com>
1424 Date: Thu Jun 19 12:34:36 2014 -0700
1426 osd: allow io priority to be set for the disk_tp
1428 The disk_tp covers scrubbing, pg deletion, and snap trimming
1430 Signed-off-by: Sage Weil <sage@inktank.com>
1431 (cherry picked from commit d9073f486527ca13cdb2774745c4c63c218333ad)
1433 commit 11858d7e7a493c9493f7039b7e45dc03fc4feb90
1434 Author: Sage Weil <sage@redhat.com>
1435 Date: Wed Jun 18 11:02:09 2014 -0700
1437 common/WorkQueue: allow io priority to be set for wq
1439 Signed-off-by: Sage Weil <sage@inktank.com>
1440 (cherry picked from commit dd6badcb5eedfec6748b3e6ca4d46e3b266038f6)
1444 src/common/WorkQueue.cc
1446 commit b75f85a2c4dd9807947862f7b89a5f25dfa1defe
1447 Author: Sage Weil <sage@redhat.com>
1448 Date: Wed Jun 18 11:01:42 2014 -0700
1450 common/Thread: allow io priority to be set for a Thread
1452 Ideally, set this before starting the thread. If you set it after, we
1453 could potentially race with create() itself.
1455 Signed-off-by: Sage Weil <sage@inktank.com>
1456 (cherry picked from commit 1b8741022c5a2ebae38905215dadee696433e931)
1458 commit 9f29788fbc19aea341b4fe997a567aa3054b9d1c
1459 Author: Sage Weil <sage@redhat.com>
1460 Date: Wed Jun 18 11:01:09 2014 -0700
1462 common/io_priority: wrap ioprio_set() and gettid()
1464 Signed-off-by: Sage Weil <sage@inktank.com>
1465 (cherry picked from commit a2b49110ef65efd526c3430ad03c988ca9dde768)
1467 commit fac5bfafef1eb82f8cf23c369fb5f50502e2e6f3
1468 Author: Sage Weil <sage@inktank.com>
1469 Date: Tue Jun 17 10:47:24 2014 -0700
1471 osd: introduce simple sleep during scrub
1473 This option is similar to osd_snap_trim_sleep: simply inject an optional
1474 sleep in the thread that is doing scrub work. This is a very kludgey and
1475 coarse knob for limiting the impact of scrub on the cluster, but can help
1476 until we have a more robust and elegant solution.
1478 Only sleep if we are in the NEW_CHUNK state to avoid delaying processing of
1479 an in-progress chunk. In this state nothing is blocked on anything.
1480 Conveniently, chunky_scrub() requeues itself for each new chunk.
1482 Backport: firefly, dumpling
1483 Signed-off-by: Sage Weil <sage@inktank.com>
1484 (cherry picked from commit c4e8451cc5b4ec5ed07e09c08fb13221e31a7ac6)
1486 commit ad82f28122dc41095d5e703279a8a5c494d35913
1487 Author: Sage Weil <sage@inktank.com>
1488 Date: Sat Jun 14 10:30:50 2014 -0700
1490 osd: add sanity check/warning on a few key configs
1492 Warn when certain config values are set to bad values.
1494 Backport: firefly, dumpling
1495 Signed-off-by: Sage Weil <sage@inktank.com>
1496 (cherry picked from commit f3ec7d0b23fdee39a34bda7595cd2a79c08daf8a)
1498 commit 289360c5dc57a4788289472f90d63781143539be
1499 Author: Sage Weil <sage@inktank.com>
1500 Date: Thu May 1 17:24:48 2014 -0700
1502 osd: prevent pgs from getting too far ahead of the min pg epoch
1504 Bound the range of PG epochs between the slowest and fastest pg
1505 (epoch-wise) with 'osd map max advance'. This value should be set to
1506 something less than 'osd map cache size' so that the maps we are
1507 processing will be in memory as many PGs advance forward in time in
1510 This is part of the solution to #7576.
1512 Signed-off-by: Sage Weil <sage@inktank.com>
1513 (cherry picked from commit cf25bdf6b0090379903981fe8cee5ea75efd7ba0)
1515 commit 662af7ac83b67af8c9edc554bbd505de717f9709
1516 Author: Sage Weil <sage@redhat.com>
1517 Date: Thu Aug 7 17:42:06 2014 -0700
1519 osd: fix pg epoch floor tracking
1521 If you call erase() on a multiset it will delete all instances of a value;
1522 we only want to delete one of them. Fix this by passing an iterator.
1525 Signed-off-by: Sage Weil <sage@redhat.com>
1526 (cherry picked from commit a52a855f6c92b03dd84cd0cc1759084f070a98c2)
1528 commit 229a346d824d84c361d7963a356af5251adc9e2e
1529 Author: Sage Weil <sage@inktank.com>
1530 Date: Wed Apr 2 14:29:08 2014 -0700
1532 osd: track per-pg epochs, min
1534 Add some simple tracking so that we can quickly determine what the min
1537 Signed-off-by: Sage Weil <sage@inktank.com>
1538 (cherry picked from commit 81e4c47722255ac3d46f701a80e104cc390e766c)
1540 commit 38c3a3c0b0d1903aea2fdc6224fc4f70bbc39329
1541 Author: Sage Weil <sage@redhat.com>
1542 Date: Wed Aug 13 13:31:10 2014 -0700
1544 mon: fix divide by zero when pg_num adjusted and no osds
1547 Backport: firefly, dumpling
1548 Signed-off-by: Sage Weil <sage@redhat.com>
1549 (cherry picked from commit 239401db7b51541a57c59a261b89e0f05347c32d)
1551 commit 87bf00a2cb93ce2dc904a42c45cc9f43023725c3
1552 Author: Sage Weil <sage@redhat.com>
1553 Date: Sun Aug 10 12:48:29 2014 -0700
1555 ceph_test_rados_api_tier: fix cache cleanup (ec too)
1557 Signed-off-by: Sage Weil <sage@redhat.com>
1558 (cherry picked from commit d7fb7bf5f2059f411633751e376c2270e6040fba)
1560 commit 3be0d731cbe158d837d05c907aefe16aa95977be
1561 Author: Sage Weil <sage@redhat.com>
1562 Date: Sun Aug 10 12:15:38 2014 -0700
1564 ceph_test_rados_api: fix cleanup of cache pool
1566 We can't simply try to delete everything in there because some items may
1567 be whiteouts. Instead, flush+evict everything, then remove overlay, and
1568 *then* delete what remains.
1571 Signed-off-by: Sage Weil <sage@redhat.com>
1572 (cherry picked from commit ebbe8aab17cfa2281902f167a706639535da4010)
1574 commit 7f511ef7fd55cca8f1babaf696ca0faa0e1e411d
1575 Author: Sage Weil <sage@redhat.com>
1576 Date: Sun Aug 10 11:41:23 2014 -0700
1578 librados/TestCase: inheret cleanup_default_namespace
1580 No need to duplicate this code.
1582 Signed-off-by: Sage Weil <sage@redhat.com>
1583 (cherry picked from commit 1d199fb1173dc9685dba96c22b83e4e9edf51a11)
1585 commit 4a5c93a90986de4510c50f0fba72ae7d5aed8a15
1586 Author: Sage Weil <sage@redhat.com>
1587 Date: Wed Aug 13 10:34:53 2014 -0700
1589 osd/ReplicatedPG: only do agent mode calculations for positive values
1591 After a split we can get negative values here. Only do the arithmetic if
1592 we have a valid (positive) value that won't through the floating point
1596 Tested-by: Karan Singh <karan.singh@csc.fi>
1597 Signed-off-by: Sage Weil <sage@redhat.com>
1598 (cherry picked from commit 5be56ff86d9f3ab2407a258a5285d0b8f52f041e)
1600 commit 39bcafa6f2b14e4e11e1541cf01f24525e2b6449
1601 Merge: 7da121d cb48cd4
1602 Author: Sage Weil <sage@inktank.com>
1603 Date: Tue Aug 12 21:15:26 2014 -0700
1605 Merge pull request #2231 from ceph/wip-8944-firefly
1609 Reviewed-by: Sage Weil <sage@redhat.com>
1611 commit 7da121d2aa1ea5c5c8accef92d7304912d4b2eb3
1612 Author: Greg Farnum <greg@inktank.com>
1613 Date: Mon Jul 28 18:33:56 2014 -0700
1615 OSD: add require_same_peer_inst(OpRequestRef&,OSDMap&) helper
1617 Signed-off-by: Greg Farnum <greg@inktank.com>
1618 (cherry picked from commit e99acf9810976b1fc74b84ad289773af43be973f)
1624 commit 8595e9bed3689933c03a8f3443052a36ff1d62f5
1625 Author: Greg Farnum <greg@inktank.com>
1626 Date: Mon Jul 28 14:19:59 2014 -0700
1628 OSD: introduce require_self_aliveness(OpRequestRef&,epoch_t) function
1630 Take the self-aliveness checks out of require_same_or_newer_map() and use
1631 the new function for that and for require_up_osd_peer().
1633 Signed-off-by: Greg Farnum <greg@inktank.com>
1634 (cherry picked from commit e179e9227b4a4482d8359682092fd7f426b9a919)
1640 commit d0f2c4891bbcc1938bee15b3fc0bfb796ce4c7de
1641 Author: Greg Farnum <greg@inktank.com>
1642 Date: Mon Jul 28 14:08:30 2014 -0700
1644 OSD: use OpRequestRef& for a few require_* functions
1646 Signed-off-by: Greg Farnum <greg@inktank.com>
1647 (cherry picked from commit eb2f1ea2c33647934af878b504383829f5a198ce)
1649 commit 8d395f1cbbc63f4cd5d166fd0a667938c83f303c
1650 Author: Greg Farnum <greg@inktank.com>
1651 Date: Tue Jul 22 16:57:00 2014 -0700
1653 OSD: introduce require_up_osd_peer() function for gating replica ops
1655 This checks both that a Message originates from an OSD, and that the OSD
1656 is up in the given map epoch.
1657 We use it in handle_replica_op so that we don't inadvertently add operations
1658 from down peers, who might or might not know it.
1660 Signed-off-by: Greg Farnum <greg@inktank.com>
1661 (cherry picked from commit ccd0eec50103b919b3eb6eea96f7dc6438520ed3)
1663 commit 72c2474e05b918535047181d876678a6a077aa1d
1664 Author: Alfredo Deza <alfredo.deza@inktank.com>
1665 Date: Fri Aug 8 10:16:20 2014 -0400
1667 init-ceph: conditionally update after argparsing
1669 Signed-off-by: Alfredo Deza <alfredo.deza@inktank.com>
1670 (cherry picked from commit 44ed1885fb588aad7e4729373e06c5933178b5e5)
1672 commit 8d7fb538e0163f828289d7da719692bd36a66f83
1673 Author: Alfredo Deza <alfredo.deza@inktank.com>
1674 Date: Thu Aug 7 11:11:57 2014 -0400
1676 make ceph-disk use the new init flag for cluster
1678 Signed-off-by: Alfredo Deza <alfredo.deza@inktank.com>
1679 (cherry picked from commit d0ccb1cbef86b8ccd8967af892e3582ee80edd55)
1681 commit 260a9c84630ca6956eb772d87abdc3bd4c4bb8fc
1682 Author: Alfredo Deza <alfredo.deza@inktank.com>
1683 Date: Thu Aug 7 10:48:09 2014 -0400
1685 allow passing a --cluster flag to the init script
1687 It will fallback to looking at /etc/ceph/$cluster.conf
1690 Signed-off-by: Alfredo Deza <alfredo.deza@inktank.com>
1691 (cherry picked from commit 23b4915fa252852cf02cc42a1b6143e2cd61b445)
1693 commit cb5ecdce68dcd4e4bc32b5c7c029cfd7ffe7c80e
1694 Author: Alfredo Deza <alfredo.deza@inktank.com>
1695 Date: Wed Aug 6 15:30:30 2014 -0400
1697 use cluster name when checking the host
1699 Signed-off-by: Alfredo Deza <alfredo.deza@inktank.com>
1700 (cherry picked from commit bdf5f30010117c71a227855a47fe22ab55b3e0b8)
1702 commit 79452540826d0aed7b94d374cc61b5002e89e4a4
1703 Author: Alfredo Deza <alfredo.deza@inktank.com>
1704 Date: Wed Aug 6 15:30:09 2014 -0400
1706 use name in /var/lib/ceph paths
1708 Signed-off-by: Alfredo Deza <alfredo.deza@inktank.com>
1709 (cherry picked from commit 3fb25c28e8e3cc62bc541538cc11f5ed41471611)
1711 commit 92755d1060ea06f13577cd2fffda6c7984a610b2
1712 Author: Alfredo Deza <alfredo.deza@inktank.com>
1713 Date: Wed Aug 6 15:29:42 2014 -0400
1715 when starting an OSD, pass in the cluster name
1717 Signed-off-by: Alfredo Deza <alfredo.deza@inktank.com>
1718 (cherry picked from commit 9e08e3a7794ce007c80c8fdb56e96900eb5240a3)
1720 commit 272b26f5aab71c22a1954ff92548b535a1f018d8
1721 Author: Samuel Just <sam.just@inktank.com>
1722 Date: Fri Aug 1 14:04:35 2014 -0700
1724 osd_types: s/stashed/rollback_info_completed and set on create
1726 Originally, this flag indicated that the object had already been stashed and
1727 that therefore recording subsequent changes is unecessary. We want to set it
1728 on create() as well since operations like [create, writefull] should not need
1729 to stash the object.
1732 Signed-off-by: Samuel Just <sam.just@inktank.com>
1733 (cherry picked from commit 4260767973d96978e808cb87ef6ae991104b4f8d)
1735 commit 7b5bcb943578b99f5a2aeeae0bbc67f371c30233
1736 Author: Joao Eduardo Luis <joao.luis@inktank.com>
1737 Date: Thu Jun 19 18:27:49 2014 +0100
1739 osd: use appropriate json types instead of stream when dumping info
1741 Signed-off-by: Joao Eduardo Luis <joao.luis@inktank.com>
1742 (cherry picked from commit 363496ad878c685008ccd87f26f121d62a48c24c)
1744 commit daec3226e073b8d6047eb9111f3deb39b1001647
1745 Author: Joao Eduardo Luis <joao.luis@inktank.com>
1746 Date: Thu Jun 19 16:52:06 2014 +0100
1748 osd: have 'tid' dumped as a json unsigned int where appropriate
1750 Most of the places dumping it would dump it as a string. Others as an int.
1751 Just a couple would dump it as an unsigned int.
1753 Signed-off-by: Joao Eduardo Luis <joao.luis@inktank.com>
1754 (cherry picked from commit 3f0ea9512be78d10ac0bc605a38e83cfbab7ffee)
1756 commit 439bce34bc25cd7c8eba7ffc124185d810d3711f
1757 Author: Yehuda Sadeh <yehuda@redhat.com>
1758 Date: Sat Aug 2 13:01:05 2014 -0700
1760 rgw: need to pass need_to_wait for throttle_data()
1762 need_to_wait wasn't passed into processor->throttle_data(). This was
1763 broken in fix for #8937.
1765 CID 1229541: (PW.PARAM_SET_BUT_NOT_USED)
1769 Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>
1770 (cherry picked from commit e93818df33286a2a7f73b593dc20da412db4e0a6)
1771 (cherry picked from commit 8519e9ab06caf6999e1243514a93eaffb155ab2e)
1773 commit fc1a4b5fdce7b92ba1b321dec37c4c75501e3521
1774 Author: Sylvain Munaut <s.munaut@whatever-company.com>
1775 Date: Thu Jun 5 11:28:27 2014 +0200
1777 rgw: Don't send error body when it's a HEAD request
1779 The main 'operation' know not to send any body for HEAD requests.
1781 However for errors, this was not the case, the formatter would be
1782 flushed and would send the error 'message' in the body in all cases.
1784 For the FastCGI case it doesn't seem to be an issue, it's possible
1785 that the webserver (apache/lighttpd/...) cleans up the response
1788 But when using the new civetweb frontend this cause invalid HTTP.
1793 Reviewed-by: Yehuda Sadeh <yehuda@inktank.com>
1794 Signed-off-by: Sylvain Munaut <s.munaut@whatever-company.com>
1795 (cherry picked from commit 0a2b4c25541bbd15776d3d35986518e37166910f)
1797 commit e3bc15343720d0455e32e429e7cf865037bd90d7
1798 Author: Sage Weil <sage@redhat.com>
1799 Date: Wed Jul 30 13:57:34 2014 -0700
1801 osd: prevent old clients from using tiered pools
1803 If the client is old and doesn't understand tiering, don't let them use a
1804 tiered pool. Reply with EOPNOTSUPP.
1808 Signed-off-by: Sage Weil <sage@redhat.com>
1809 (cherry picked from commit 0190df53056834f219e33ada2af3a79e8c4dfb77)
1811 commit f3113e9adde54fa50a64010381f5b6c113f5cda0
1812 Author: Sage Weil <sage@redhat.com>
1813 Date: Thu Jul 31 11:02:55 2014 -0700
1815 mon/OSDMonitor: warn when cache pools do not have hit_sets configured
1817 Give users a clue when cache pools are enabled but the hit_set is not
1818 configured. Note that technically this will work, but not well, so for
1819 now let's just steer them away.
1821 Signed-off-by: Sage Weil <sage@redhat.com>
1822 (cherry picked from commit 383536a2ef2506c5591b25c0fd61dc8c181fdab9)
1824 NOTE: removed readforward state for backport
1826 commit aaf732f7475c075f8bf8e79bc08b11a020e5fadf
1827 Author: Sage Weil <sage@redhat.com>
1828 Date: Thu Jul 31 09:28:12 2014 -0700
1830 osd/ReplicatedPG: improve agent_choose_mode args
1832 Signed-off-by: Sage Weil <sage@redhat.com>
1833 (cherry picked from commit caf554bf2f752b50670d118c99843b7c4a2967b4)
1835 commit 76a519b215e591e0948d627fb1de826ed0bf415d
1836 Author: Sage Weil <sage@redhat.com>
1837 Date: Thu Jul 31 09:26:03 2014 -0700
1839 osd/ReplicatedPG: evict blindly if there is no hit_set
1841 If there is no hit set for a PG, blindly evict objects. This avoids an
1842 assert(hit_set) in agent_estimate_atime_temp().
1845 Signed-off-by: Sage Weil <sage@redhat.com>
1846 (cherry picked from commit ea4996d3ef15c5d5a776d162b08d9fad6859c976)
1848 commit 8ff48d79697e98dca04799ee5982a5308a4efc92
1849 Author: Sage Weil <sage@redhat.com>
1850 Date: Thu Jul 31 09:13:11 2014 -0700
1852 osd/ReplicatedPG: check agent_mode if agent is enabled but hit_sets aren't
1854 It is probably not a good idea to try to run the tiering agent without a
1855 hit_set to inform its actions, but it is technically possible. For
1856 example, one could simply blindly evict when we reach the full point.
1857 However, this doesn't work because the agent mode is guarded by a hit_set
1858 check, even though agent_setup() is not. Fix that.
1860 Signed-off-by: Sage Weil <sage@redhat.com>
1861 (cherry picked from commit 5d1c76f641310f5f65600f70ae76945b2aa472d7)
1863 commit f98a6194a9a9f5794c38814c1c33d13b7fcea538
1864 Author: Sage Weil <sage@redhat.com>
1865 Date: Tue Aug 12 07:05:34 2014 -0700
1867 common/LogClient: fix sending dup log items
1869 We need to skip even the most recently sent item in order to get to the
1870 ones we haven't sent yet.
1873 Backport: firefly, dumpling
1874 Signed-off-by: Sage Weil <sage@redhat.com>
1875 (cherry picked from commit 057c6808be5bc61c3f1ac2b956c1522f18411245)
1877 commit bbb1e6a3a84b2d853f935687f01016e7c6a67951
1878 Author: Pavan Rallabhandi <pavan.rallabhandi@sandisk.com>
1879 Date: Wed Aug 6 15:10:14 2014 +0530
1881 RadosClient: Fixing potential lock leaks.
1883 In lookup_pool and pool_delete, a lock is taken
1884 before invoking wait_for_osdmap, but is not
1885 released for the failure case of the call. Fixing the same.
1888 Signed-off-by: Pavan Rallabhandi <pavan.rallabhandi@sandisk.com>
1889 (cherry picked from commit f1aad8bcfc53f982130dbb3243660c3c546c3523)
1891 commit e767254ce9a1610f3d640d8157fcb0d1fdf8edf2
1892 Author: Josh Durgin <josh.durgin@inktank.com>
1893 Date: Mon Aug 11 16:41:26 2014 -0700
1895 librbd: fix error path cleanup for opening an image
1897 If the image doesn't exist and caching is enabled, the ObjectCacher
1898 was not being shutdown, and the ImageCtx was leaked. The IoCtx could
1899 later be closed while the ObjectCacher was still running, resulting in
1900 a segfault. Simply use the usual cleanup path in open_image(), which
1904 Backport: dumpling, firefly
1905 Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
1906 (cherry picked from commit 3dfa72d5b9a1f54934dc8289592556d30430959d)
1908 commit dec028d113ab4bcc7106fea0bf413b54829f67e7
1909 Merge: e087862 c7fbf27
1910 Author: Sage Weil <sage@redhat.com>
1911 Date: Tue Aug 12 14:38:33 2014 -0700
1913 Merge remote-tracking branch 'gh/firefly-next' into firefly
1915 commit 07e1ceb55ac3cf0bea58cec27737e45391de9106
1916 Author: Loic Dachary <loic@dachary.org>
1917 Date: Sun Aug 10 17:10:04 2014 +0200
1919 erasure-code: ErasureCodeJerasure::create_ruleset must return a ruleset
1921 CrushWrapper::add_simple_ruleset does not return a ruleset, it returns
1922 a ruleid that must be converted into a ruleset before being returned.
1924 http://tracker.ceph.com/issues/9044 Fixes: #9044
1926 Signed-off-by: Loic Dachary <loic@dachary.org>
1927 (cherry picked from commit 0029a35872d3fc15f9a0d60d095b2e111d6e98a6)
1929 commit 96047f1de4e0591a294e31ed6af6edfb3b2357e4
1930 Author: Loic Dachary <loic@dachary.org>
1931 Date: Sun Aug 10 17:06:33 2014 +0200
1933 erasure-code: OSDMonitor::crush_ruleset_create_erasure needs ruleset
1935 When OSDMonitor::crush_ruleset_create_erasure checks the ruleset for
1936 existence, it must convert the ruleid into a ruleset before assigning it
1937 back to the *ruleset parameter.
1939 http://tracker.ceph.com/issues/9044 Fixes: #9044
1941 Signed-off-by: Loic Dachary <loic@dachary.org>
1942 (cherry picked from commit 04a484a4d5800c9bcf9805d3e87ed0c8dee01c9a)
1944 commit e087862d5bd894572a6c223b45f4275cadd4b480
1945 Author: Haomai Wang <haomaiwang@gmail.com>
1946 Date: Mon Jul 14 14:27:17 2014 +0800
1948 Add rbdcache max dirty object option
1950 Librbd will calculate max dirty object according to rbd_cache_max_size, it
1951 doesn't suitable for every case. If user set image order 24, the calculating
1952 result is too small for reality. It will increase the overhead of trim call
1953 which is called each read/write op.
1955 Now we make it as option for tunning, by default this value is calculated.
1957 Signed-off-by: Haomai Wang <haomaiwang@gmail.com>
1958 (cherry picked from commit 3c7229a2fea98b30627878c86b1410c8eef2b5d7)
1960 commit 5f74fb294a0cd9b973aae45de17f38436bac6d61
1961 Author: Danny Al-Gaaf <danny.al-gaaf@bisect.de>
1962 Date: Wed Jun 4 23:22:18 2014 +0200
1964 librbd/internal.cc: check earlier for null pointer
1966 Fix potential null ponter deref, move check for 'order != NULL'
1967 to the beginning of the function to prevent a) deref in ldout() call
1968 and b) to leave function as early as possible if check fails.
1970 [src/librbd/internal.cc:843] -> [src/librbd/internal.cc:865]: (warning)
1971 Possible null pointer dereference: order - otherwise it is redundant
1972 to check it against null.
1974 Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>
1975 (cherry picked from commit 3ee3e66a9520a5fcafa7d8c632586642f7bdbd29)
1977 commit b2189465f36fc375b864f88a47d25c7808c08901
1978 Author: Josh Durgin <josh.durgin@inktank.com>
1979 Date: Thu Apr 24 14:47:24 2014 -0700
1981 librbd: add an interface to invalidate cached data
1983 This is useful for qemu to guarantee live migration with caching is
1984 safe, by invalidating the cache on the destination before starting it.
1986 Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
1987 (cherry picked from commit 5d340d26dd70192eb0e4f3f240e3433fb9a24154)
1989 commit f7dc58552a9ee228b7fa2a10094e962ed66fe8fd
1990 Author: Josh Durgin <josh.durgin@inktank.com>
1991 Date: Thu Apr 24 14:43:35 2014 -0700
1993 librbd: check return code and error out if invalidate_cache fails
1995 This will only happen when shrinking or rolling back an image is done
1996 while other I/O is in flight to the same ImageCtx. This is unsafe, so
1997 return an error before performing the resize or rollback.
1999 Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
2000 (cherry picked from commit e08b8b66c77be3a3d7f79d91c20b1619571149ee)
2002 commit c7fbf272a202cd8a72d91b7a55e53fbb5cbbbadd
2003 Author: Sage Weil <sage@redhat.com>
2004 Date: Sat Aug 9 13:23:06 2014 -0700
2006 os/FileStore: dump open fds before asserting
2008 Backport: firefly, dumpling
2009 Signed-off-by: Sage Weil <sage@redhat.com>
2010 (cherry picked from commit 4e8de1792731cf30f2744ab0659d036adc0565a3)
2012 commit 978f5749f3d08bb4c96529ca2d1b10760ee2606d
2013 Author: Sage Weil <sage@redhat.com>
2014 Date: Fri Jul 25 13:51:45 2014 -0700
2016 ceph_test_rados_api_tier: do fewer writes in HitSetWrite
2018 We don't need to do quite so many writes. It can be slow when we are
2019 thrashing and aren't doing anything in parallel.
2022 Signed-off-by: Sage Weil <sage@redhat.com>
2023 (cherry picked from commit c5f766bb16c0ab3c3554e73791ad0b74077ad35c)
2025 commit 22aa81413e7db43418e613aea037e978031f5b8a
2026 Merge: b65cef6 7025075
2027 Author: Sage Weil <sage@redhat.com>
2028 Date: Fri Aug 8 19:05:37 2014 -0700
2030 Merge remote-tracking branch 'gh/firefly-next' into firefly
2032 commit cb48cd47a26d27869884af83012d6a8a684abf25
2033 Author: Dan Mick <dan.mick@inktank.com>
2034 Date: Wed Jul 30 14:50:37 2014 -0700
2036 qa/workunits/cephtool/test_daemon.sh: verify ceph -c works with daemon
2038 Signed-off-by: Dan Mick <dan.mick@inktank.com>
2039 (cherry picked from commit aa9ae1f270293778aa937e7f7e4bcaee3099b9b2)
2041 commit 1ac0e5436da68e441f42a1016b12481046147850
2042 Author: Dan Mick <dan.mick@inktank.com>
2043 Date: Wed Jul 30 14:49:02 2014 -0700
2045 qa/workunits/cephtool/test_daemon.sh: typo
2047 Signed-off-by: Dan Mick <dan.mick@inktank.com>
2048 (cherry picked from commit 22d20f39b7355966554319d5a1aa888967607569)
2050 commit a144950624208d8ed2901fd7421060a706ac2016
2051 Author: Dan Mick <dan.mick@inktank.com>
2052 Date: Wed Jul 30 14:48:28 2014 -0700
2054 qa/workunits/cephtool/test_daemon.sh: allow local ceph command
2056 (cherry picked from commit 97a8d5a9fdbd3a25cc922c242ee57da58c57d0bc)
2058 commit 76b593f6268d77e6cf5a9931f6f3524b3f574c09
2059 Author: Dan Mick <dan.mick@inktank.com>
2060 Date: Mon Jul 28 21:28:42 2014 -0700
2062 ceph.in: Pass global args to ceph-conf for proper lookup
2065 Signed-off-by: Dan Mick <dan.mick@inktank.com>
2066 (cherry picked from commit 6d89a99648630f81b85ad115fe7662dba6b08a55)
2068 commit b65cef678777c1b87d25385595bf0df96168703e
2069 Author: Sage Weil <sage@inktank.com>
2070 Date: Mon Jul 14 08:04:16 2014 -0700
2072 ceph_test_rados_api_tier: fix [EC] HitSet{Read,Write,Trim} tests
2074 The hit_set_ fields can only be set on tier pools as of
2075 f131dfbaedf6f451572e7aa3a83f653912122953.
2078 Signed-off-by: Sage Weil <sage@inktank.com>
2079 (cherry picked from commit e17e9d857722ee478abda10adb32e15b11fff2ff)
2081 commit 70250755e4b2c114f1f0f463b42930dd99751d81
2082 Merge: 6f1a544 f8494c8
2083 Author: Samuel Just <sam.just@inktank.com>
2084 Date: Sun Aug 3 12:47:23 2014 -0700
2086 Merge branch 'wip-8701-firefly' into firefly-next
2088 commit f8494c87e5276f0587c3341ddef9d7a933dfea11
2089 Author: Sage Weil <sage@redhat.com>
2090 Date: Tue Jul 22 06:53:41 2014 -0700
2092 ceph_test_objectstore: clean up on finish of MoveRename
2094 Otherwise, we leave collections around, and the next test fails.
2096 Signed-off-by: Sage Weil <sage@redhat.com>
2097 (cherry picked from commit d4faf747b73e70dff9cb5c98ee6aaa4ecec215fc)
2099 commit 94fee997204d439abfd642be20a7e33df5979342
2100 Author: Sage Weil <sage@redhat.com>
2101 Date: Mon Jul 21 13:45:21 2014 -0700
2103 os/LFNIndex: use FDCloser for fsync_dir
2105 This prevents an fd leak when maybe_inject_failure() throws an exception.
2107 Signed-off-by: Sage Weil <sage@redhat.com>
2108 (cherry picked from commit 3ec9a42b470422b1fe72b6294d82d9efcaca7f53)
2110 commit b213866eac422a6ac6aeacd3ab4741a6421ce20c
2111 Author: Sage Weil <sage@redhat.com>
2112 Date: Fri Jul 18 23:16:09 2014 -0700
2114 os/LFNIndex: only consider alt xattr if nlink > 1
2116 If we are doing a lookup, the main xattr fails, we'll check if there is an
2117 alt xattr. If it exists, but the nlink on the inode is only 1, we will
2118 kill the xattr. This cleans up the mess left over by an incomplete
2119 lfn_unlink operation.
2121 This resolves the problem with an lfn_link to a second long name that
2122 hashes to the same short_name: we will ignore the old name the moment the
2126 Signed-off-by: Sage Weil <sage@redhat.com>
2127 (cherry picked from commit 6fb3260d59faab1e20ebf1e44f850f85f6b8342a)
2129 commit 4fc72d8d936ac499fbb53a1d73a3f7c05ebe596a
2130 Author: Sage Weil <sage@redhat.com>
2131 Date: Fri Jul 18 17:28:18 2014 -0700
2133 os/LFNIndex: remove alt xattr after unlink
2135 After we unlink, if the nlink on the inode is still non-zero, remove the
2136 alt xattr. We can *only* do this after the rename or unlink operation
2137 because we don't want to leave a file system link in place without the
2138 matching xattr; hence the fsync_dir() call.
2140 Note that this might leak an alt xattr if we happen to fail after the
2141 rename/unlink but before the removexattr is committed. We'll fix that
2144 Signed-off-by: Sage Weil <sage@redhat.com>
2145 (cherry picked from commit ec36f0a130d67df6cbeefcc9c2d83eb703b6b28c)
2147 commit 816ed7e951b32bb9ac57466c0affaec9781ac7f4
2148 Author: Sage Weil <sage@redhat.com>
2149 Date: Mon Jul 21 13:43:42 2014 -0700
2151 os/LFNIndex: FDCloser helper
2153 Add a helper to close fd's when we leave scope. This is important when
2154 injecting failures by throwing exceptions.
2156 Signed-off-by: Sage Weil <sage@redhat.com>
2157 (cherry picked from commit a320c260a9e088ab0a4ea3d5298c06a2d077de37)
2159 commit cbfbe637851c7ebe4a9ec1fd6e429cdf85aef608
2160 Author: Sage Weil <sage@redhat.com>
2161 Date: Fri Jul 18 17:09:07 2014 -0700
2163 os/LFNIndex: handle long object names with multiple links (i.e., rename)
2165 When we rename an object (collection_move_rename) to a different name, and
2166 the name is long, we run into problems because the lfn xattr can only track
2167 a single long name linking to the inode. For example, suppose we have
2169 foobar -> foo_123_0 (attr: foobar) where foobar hashes to 123.
2171 At first, collection_add could only link a file to another file in a
2172 different collection with the same name. Allowing collection_move_rename
2173 to rename the file, however, means that we have to convert:
2175 col1/foobar -> foo_123_0 (attr: foobar)
2179 col1/foobaz -> foo_234_0 (attr: foobaz)
2181 This is a problem because if we link, reset xattr, unlink we end up with
2183 col1/foobar -> foo_123_0 (attr: foobaz)
2185 if we restart after we reset the attr. This will cause the initial foobar
2186 lookup to since the attr doesn't match, and the file won't be able to be
2189 Fix this by allow *two* (long) names to link to the same inode. If we
2190 lfn_link a second (different) name, move the previous name to the "alt"
2191 xattr and set the new name. (This works because link is always followed
2192 by unlink.) On lookup, check either xattr.
2194 Don't even bother to remove the alt xattr on unlink. This works as long
2195 as the old name and new name don't hash to the same shortname and end up
2196 in the same LFN chain. (Don't worry, we'll fix that next.)
2199 Signed-off-by: Sage Weil <sage@redhat.com>
2201 (cherry picked from commit b2cdfce6461b81f4926602a8c63b54aa92684e6c)
2203 commit 5db6c12b61e00b0bc8084ead5976a912ece0fc65
2204 Author: Sage Weil <sage@redhat.com>
2205 Date: Fri Jul 18 15:46:58 2014 -0700
2207 ceph_test_objectstore: fix warning
2209 Signed-off-by: Sage Weil <sage@redhat.com>
2210 (cherry picked from commit cf98805c09a38cce78ac08317899dc4152ae55a5)
2212 commit e18777ed6eb492e8421d53090b3c9af6e75bd792
2213 Author: Samuel Just <sam.just@inktank.com>
2214 Date: Tue Jul 15 14:50:33 2014 -0700
2216 store_test: add long name collection_move_rename tests
2220 Signed-off-by: Samuel Just <sam.just@inktank.com>
2221 (cherry picked from commit 6aa48a485e03ca100f3d9ebec77cc06f99756cd7)
2224 src/test/objectstore/store_test.cc
2226 commit 4e03d5b512c8d2f7fa51dda95c6132e676529f9b
2227 Author: Sage Weil <sage@redhat.com>
2228 Date: Sun Aug 3 08:40:57 2014 -0700
2230 Revert "enforce rados put aligment"
2232 This reverts commit 7a58da53ebfcaaf385c21403b654d1d2f1508e1a.
2234 This was alread backported in dece65064d949b5afcc359cd408615883b5e002a.
2237 Signed-off-by: Sage Weil <sage@redhat.com>
2239 commit 6f1a54436c63a4cb63712936fccade3267d03db8
2240 Author: Yehuda Sadeh <yehuda@redhat.com>
2241 Date: Tue Jul 29 15:25:47 2014 -0700
2243 rgw: fix crash in swift CORS preflight request
2247 This fixes error handling, in accordance with commit 6af5a537 that fixed
2248 the same issue for the S3 case.
2250 Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>
2251 (cherry picked from commit 18ea2a869791b4894f93fdafde140285f2e4fb65)
2253 commit dfcd99567e925bd6565ef4676c560f6c8c89e683
2254 Author: devicenull <dn@devicenull.org>
2255 Date: Fri Jul 18 10:25:51 2014 -0400
2257 rgw: fix decoding + characters in URL
2262 Only decode + characters to spaces if we're in a query argument. The +
2263 query argument. The + => ' ' translation is not correct for
2264 file/directory names.
2266 Resolves http://tracker.ceph.com/issues/8702
2268 Reviewed-by: Yehuda Sadeh <yehuda@redhat.com>
2269 Signed-off-by: Brian Rak <dn@devicenull.org>
2270 (cherry picked from commit 4a63396ba1611ed36cccc8c6d0f5e6e3e13d83ee)
2272 commit f9716214a35e3cddd3006f3839f9688c41f0c7ca
2273 Author: Yehuda Sadeh <yehuda@redhat.com>
2274 Date: Fri Jul 25 20:33:52 2014 -0700
2276 rgw: call processor->handle_data() again if needed
2280 Following the fix to #8928 we end up accumulating pending data that
2281 needs to be written. Beforehand it was working fine because we were
2282 feeding it with the exact amount of bytes we were writing.
2284 Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>
2285 (cherry picked from commit 0553890e79b43414cc0ef97ceb694c1cb5f06bbb)
2290 commit 4cd1b60c043089358ea8e70a51310005d5b68383
2291 Author: Yehuda Sadeh <yehuda@redhat.com>
2292 Date: Thu Jul 24 15:30:27 2014 -0700
2294 rgw: object write should not exceed part size
2298 This can happen if the stripe size is not a multiple of the chunk size.
2302 Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>
2303 (cherry picked from commit 14cad5ece7d1de9d93e72acca6d4c3b4a9cfcfa2)
2305 commit f68720a6a61364f72f4e0c62aabea65eabe33f98
2306 Author: Yehuda Sadeh <yehuda@redhat.com>
2307 Date: Tue Jul 22 15:30:11 2014 -0700
2309 rgw: align object chunk size with pool alignment
2313 Data pools might have strict write alignment requirements. Use pool
2314 alignment info when setting the max_chunk_size for the write.
2316 Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>
2317 (cherry picked from commit fc83e197ab85355e385c13f2a64957cad7481298)
2320 src/rgw/rgw_rados.cc
2322 commit 67d73451c893b58634f06ed141792da1bf0fa77c
2323 Author: Yehuda Sadeh <yehuda@redhat.com>
2324 Date: Wed Jul 30 11:53:16 2014 -0700
2326 cls_rgw: fix object name of objects removed on object creation
2329 Backport: firefly, dumpling
2331 Reported-by: Patrycja Szabłowska <szablowska.patrycja@gmail.com>
2332 Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>
2333 (cherry picked from commit 0f8929a68aed9bc3e50cf15765143a9c55826cd2)
2335 commit d90148da293b4a3e2a44d4486a6b48e27d35b79b
2336 Merge: 223a0a9 d48a737
2337 Author: Samuel Just <sam.just@inktank.com>
2338 Date: Sat Aug 2 11:28:14 2014 -0700
2340 Merge remote-tracking branch 'origin/wip-8438' into firefly-next
2342 Backport of c5b8d8105d965da852c79add607b69d5ae79a4d4
2344 commit 223a0a9e7ed76fbc114c2f058d3aff93556244e9
2345 Merge: 84bd3cd d72eec0
2346 Author: Samuel Just <sam.just@inktank.com>
2347 Date: Sat Aug 2 11:22:47 2014 -0700
2349 Merge remote-tracking branch 'origin/wip-7999' into firefly-next
2351 Backport of 830940bf242a73403ec1882a489e31f7694b7f7e
2353 commit 84bd3cd79961d1ddce12048f9ae980fc197e6be4
2354 Author: Sage Weil <sage@redhat.com>
2355 Date: Mon Jul 28 17:17:23 2014 -0700
2357 unittest_crush_wrapper: fix build
2359 Signed-off-by: Sage Weil <sage@redhat.com>
2360 (cherry picked from commit f36cffc986c973014c89aa37ca73740b2fc194ca)
2362 commit 4983aa32fe8c64c78957b921f5d31ab07be9f850
2363 Merge: a2681c9 b6d8fea
2364 Author: Sage Weil <sage@inktank.com>
2365 Date: Fri Aug 1 19:26:44 2014 -0700
2367 Merge pull request #2178 from dachary/wip-erasure-code-profile-default-firefly
2369 erasure-code: create default profile if necessary (firefly)
2371 commit d89a5a37f2515cc9af1322fd18357e445b94256f
2372 Author: Sage Weil <sage@redhat.com>
2373 Date: Wed Jul 30 12:26:44 2014 -0700
2377 Signed-off-by: Sage Weil <sage@redhat.com>
2378 (cherry picked from commit d700076a42a5a5ebe769a8311fd3b52bf2e98cd2)
2380 commit da0de4ee973d9a9c8ee6f9a5c12eb37c83b4d453
2381 Author: Sage Weil <sage@redhat.com>
2382 Date: Wed Jul 30 14:52:06 2014 -0700
2384 atomic: fix read() on i386, clean up types
2386 Among other things, fixes #8969
2389 Signed-off-by: Sage Weil <sage@redhat.com>
2390 (cherry picked from commit 96863128e6668257f435c6962263caae0d7d10dd)
2392 commit 1e3b927731104629cae9682b55b8d540ef8af12f
2393 Author: Sage Weil <sage@inktank.com>
2394 Date: Thu Jun 5 10:43:16 2014 -0700
2396 include/atomic: make 32-bit atomic64_t unsigned
2400 In file included from test/perf_counters.cc:19:0:
2401 ./common/perf_counters.h: In member function ‘std::pair PerfCounters::perf_counter_data_any_d::read_avg() const’:
2402 warning: ./common/perf_counters.h:156:36: comparison between signed and unsigned integer expressions [-Wsign-compare]
2403 } while (avgcount2.read() != count);
2406 Signed-off-by: Sage Weil <sage@inktank.com>
2407 (cherry picked from commit 2081c992bbe3a83d711f465634d19c011d28ea3e)
2409 commit 34ac3060d408604c7082c583641b90f330f12663
2410 Author: Dmitry Smirnov <onlyjob@member.fsf.org>
2411 Date: Wed May 21 07:56:43 2014 +1000
2413 Define AO_REQUIRE_CAS (fixes FTBFS on 'hppa')
2415 to fix FTBFS due to undeclared atomic functions.
2419 https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=748571
2421 by John David Anglin <dave.anglin@bell.net>
2424 ./include/atomic.h: In member function 'size_t ceph::atomic_t::inc()':
2425 ./include/atomic.h:42:36: error: 'AO_fetch_and_add1' was not declared in this scope
2426 return AO_fetch_and_add1(&val) + 1;
2428 ./include/atomic.h: In member function 'size_t ceph::atomic_t::dec()':
2429 ./include/atomic.h:45:42: error: 'AO_fetch_and_sub1_write' was not declared in this scope
2430 return AO_fetch_and_sub1_write(&val) - 1;
2432 ./include/atomic.h: In member function 'void ceph::atomic_t::add(size_t)':
2433 ./include/atomic.h:48:36: error: 'AO_fetch_and_add' was not declared in this scope
2434 AO_fetch_and_add(&val, add_me);
2436 ./include/atomic.h: In member function 'void ceph::atomic_t::sub(int)':
2437 ./include/atomic.h:52:48: error: 'AO_fetch_and_add_write' was not declared in this scope
2438 AO_fetch_and_add_write(&val, (AO_t)negsub);
2440 ./include/atomic.h: In member function 'size_t ceph::atomic_t::dec()':
2441 ./include/atomic.h:46:5: warning: control reaches end of non-void function [-Wreturn-type]
2444 make[5]: *** [cls/user/cls_user_client.o] Error 1
2447 Signed-off-by: Dmitry Smirnov <onlyjob@member.fsf.org>
2448 (cherry picked from commit 74218f3d6ca8ca9943ff9d08b7926e38fb13b329)
2450 commit a0c12b70d93604cde59fc21b4d42cae6bd0f4f26
2451 Author: Yehuda Sadeh <yehuda@inktank.com>
2452 Date: Mon Mar 31 14:49:50 2014 -0700
2454 atomic_t: add atomic64_t
2456 Signed-off-by: Yehuda Sadeh <yehuda@inktank.com>
2457 (cherry picked from commit bf3ba6001c7b4cf37edfe6551d3ef298ebcbf421)
2459 commit 5a8c8c01cac7f8ef8b898a88cac96af973350a3d
2460 Author: Sage Weil <sage@redhat.com>
2461 Date: Wed Jul 30 13:40:33 2014 -0700
2463 test/cli-integration/rbd: fix trailing space
2465 Newer versions of json.tool remove the trailing ' ' after the comma. Add
2466 it back in with sed so that the .t works on both old and new versions, and
2467 so that we don't have to remove the trailing spaces from all of the test
2472 Signed-off-by: Sage Weil <sage@redhat.com>
2473 (cherry picked from commit 605064dc685aa25cc7d58ec18b6449a3ce476d01)
2476 src/test/cli-integration/rbd/defaults.t
2478 commit 234b2d1933f7f403c094a3c65bf912fcee688181
2479 Author: John Spray <jspray@redhat.com>
2480 Date: Mon Jul 7 15:45:54 2014 +0100
2482 tests: don't depend on 'data' pool in rbd test
2484 Since we removed the default 'data' and 'metadata' pools,
2485 tests which need a pool should create it themselves.
2487 Signed-off-by: John Spray <john.spray@redhat.com>
2488 (cherry picked from commit a7a631d1e284f151e305f770cef2042a1b9f86c0)
2490 commit 751258db76d9f5288ff42e8e961a9ffea1d226de
2491 Author: Xiaoxi Chen <xiaoxi.chen@intel.com>
2492 Date: Mon Jul 28 16:54:48 2014 +0800
2494 PGMonitor: fix bug in caculating pool avail space
2496 Currently for pools with different rules, "ceph df" cannot report
2497 right available space for them, respectively. For detail assisment
2498 of the bug ,pls refer to bug report #8943
2500 This patch fix this bug and make ceph df works correctlly.
2504 Signed-off-by: Xiaoxi Chen <xiaoxi.chen@intel.com>
2505 (cherry picked from commit 04d0526718ccfc220b4fe0c9046ac58899d9dafc)
2507 commit 7b1546cc5c361a1af7f724e3c3b7ca9436bad76f
2508 Author: Sage Weil <sage@inktank.com>
2509 Date: Wed May 28 13:59:47 2014 -0700
2511 mon: set min_size to data chunk count for erasure pools
2513 Make the min_size value meaningful for erasure pools.
2515 Signed-off-by: Sage Weil <sage@inktank.com>
2516 (cherry picked from commit e06c58c9b8f585d2fe7c97d010aa0aa61c09d609)
2518 commit e859587911edddd3a5b7fd96ee2f6a7ed77478dc
2519 Author: Sage Weil <sage@inktank.com>
2520 Date: Sun May 11 13:36:03 2014 -0700
2522 mon: include 'max avail' in df output
2524 Include an estimate of the maximum writeable space for each pool. Note
2525 that this value is a conservative estimate for that pool based on the
2526 most-full OSD. It is also potentially misleading as it is the available
2527 space if *all* new data were written to this pool; one cannot (generally)
2528 add up the available space for all pools.
2530 Signed-off-by: Sage Weil <sage@inktank.com>
2531 (cherry picked from commit 7a9652b58ea70f9a484a135bde20d872616c5947)
2533 commit a5e0ebc36ccbdf36e733e5eab926e4446a48fbd8
2534 Author: Sage Weil <sage@inktank.com>
2535 Date: Sun May 11 13:31:14 2014 -0700
2537 mon: right justify df values
2539 Signed-off-by: Sage Weil <sage@inktank.com>
2540 (cherry picked from commit 2f63a309df4b7086725949bc0a532595cf927edf)
2542 commit 3d578479420f4b0c153dffe96605bf413070ad45
2543 Author: John Spray <john.spray@inktank.com>
2544 Date: Thu May 8 11:04:46 2014 +0100
2546 mon: Fix % escaping (\% should be %%)
2548 Clang's -Wpedantic points this out.
2550 Signed-off-by: John Spray <john.spray@inktank.com>
2551 (cherry picked from commit f0231ef364d531eb60351598c4a0f5fa6efad23c)
2554 src/mon/DataHealthService.cc
2556 commit 6164c3ef8d9ced38d878c191e4de3d84a8bda45e
2557 Author: Sage Weil <sage@inktank.com>
2558 Date: Wed May 28 13:49:52 2014 -0700
2560 crush: add get_rule_weight_map
2562 Calculate a weight map of OSDs for a given rule.
2564 Signed-off-by: Sage Weil <sage@inktank.com>
2565 (cherry picked from commit 297f6169feecd20e121d102e1b63a505c8b3e74a)
2567 commit 2f90bfdb854750990881f2f53492687a07b19546
2568 Author: Guang Yang <yguang@yahoo-inc.com>
2569 Date: Wed Jul 9 11:20:36 2014 +0000
2571 Fix the PG listing issue which could miss objects for EC pool (where there is object shard and generation).
2573 Signed-off-by: Guang Yang (yguang@yahoo-inc.com)
2575 (cherry picked from commit 228760ce3a7109f50fc0f8e3c4a5697a423cb08f)
2577 commit 61c90252c3a496e09962cfda2eb11e6bd112af8a
2578 Author: Sage Weil <sage@redhat.com>
2579 Date: Fri Jul 25 14:48:10 2014 -0700
2581 osd/ReplicatedPG: requeue cache full waiters if no longer writeback
2583 If the cache is full, we block some requests, and then we change the
2584 cache_mode to something else (say, forward), the full waiters don't get
2585 requeued until the cache becomes un-full. In the meantime, however, later
2586 requests will get processed and redirected, breaking the op ordering.
2588 Fix this by requeueing any full waiters if we see that the cache_mode is
2592 Signed-off-by: Sage Weil <sage@redhat.com>
2593 (cherry picked from commit 8fb761b660c268e2264d375a4db2f659a5c3a107)
2595 commit f701688a360fd977b5e1f59518622ab6d2299a08
2596 Author: Sage Weil <sage@redhat.com>
2597 Date: Fri Jul 25 14:43:48 2014 -0700
2599 osd/ReplicatedPG: fix cache full -> not full requeueing when !active
2601 We only want to do this if is_active(). Otherwise, the normal
2602 requeueing code will do its thing, taking care to get the queue orders
2606 Signed-off-by: Sage Weil <sage@redhat.com>
2607 (cherry picked from commit 36aaab9eee7ed41a46a4ac27376d630a29de5eb9)
2609 commit a1ddce36ebf1ce479129def364328486fbcc5c86
2610 Author: Dan Mick <dan.mick@inktank.com>
2611 Date: Wed Jul 30 14:50:37 2014 -0700
2613 qa/workunits/cephtool/test_daemon.sh: verify ceph -c works with daemon
2615 Signed-off-by: Dan Mick <dan.mick@inktank.com>
2616 (cherry picked from commit aa9ae1f270293778aa937e7f7e4bcaee3099b9b2)
2618 commit 50a1c08e46eaedd3d81a52f64923228bc8a81c26
2619 Author: Dan Mick <dan.mick@inktank.com>
2620 Date: Wed Jul 30 14:49:02 2014 -0700
2622 qa/workunits/cephtool/test_daemon.sh: typo
2624 Signed-off-by: Dan Mick <dan.mick@inktank.com>
2625 (cherry picked from commit 22d20f39b7355966554319d5a1aa888967607569)
2627 commit 4cfb7dbc5c8fe1a28fb945fa12d9795943db1db5
2628 Author: Dan Mick <dan.mick@inktank.com>
2629 Date: Wed Jul 30 14:48:28 2014 -0700
2631 qa/workunits/cephtool/test_daemon.sh: allow local ceph command
2633 (cherry picked from commit 97a8d5a9fdbd3a25cc922c242ee57da58c57d0bc)
2635 commit 120b5d5ad51fa47a2bb0a144821635f5f81b5a2e
2636 Author: Dan Mick <dan.mick@inktank.com>
2637 Date: Mon Jul 28 21:28:42 2014 -0700
2639 ceph.in: Pass global args to ceph-conf for proper lookup
2642 Signed-off-by: Dan Mick <dan.mick@inktank.com>
2643 (cherry picked from commit 6d89a99648630f81b85ad115fe7662dba6b08a55)
2645 commit e0c27cd7fde84944fd350f7df89e79d9a39ace98
2646 Author: Sage Weil <sage@inktank.com>
2647 Date: Wed Jul 16 17:55:36 2014 -0700
2649 qa/workunits/cephtool/test.sh: test osd pool get erasure_code_profile
2651 Signed-off-by: Sage Weil <sage@inktank.com>
2652 (cherry picked from commit ce9f12d7a2202948532fed9da4d763ed03f6b8fa)
2655 qa/workunits/cephtool/test.sh
2657 commit cf22759cdf1c0a6d258e5c0a2d32268ce051c742
2658 Author: Ma Jianpeng <jianpeng.ma@intel.com>
2659 Date: Wed Jul 16 17:48:34 2014 -0700
2661 mon: OSDMonitor: add "osd pool get <pool> erasure_code_profile" command
2663 Enable us to obtain the erasure-code-profile for a given erasure-pool.
2665 Signed-off-by: Ma Jianpeng <jianpeng.ma@intel.com>
2666 Signed-off-by: Sage Weil <sage@inktank.com>
2667 (cherry picked from commit e8ebcb79a462de29bcbabe40ac855634753bb2be)
2669 commit b94d2f142db63bdf395d8288f8dcef1558d30eaf
2670 Author: Sage Weil <sage@redhat.com>
2671 Date: Wed Jul 23 18:25:53 2014 -0700
2673 osd/ReplicatedPG: observe INCOMPLETE_CLONES in is_present_clone()
2675 We cannot assume that just because cache_mode is NONE that we will have
2676 all clones present; check for the absense of the INCOMPLETE_CLONES flag
2679 Signed-off-by: Sage Weil <sage@redhat.com>
2680 (cherry picked from commit 63abf11390bb9b8dd604aae2b3e90596f9ab65ac)
2682 commit 22cc1b643f83edbfc0475f6da89cf4bd4c89aae1
2683 Author: Sage Weil <sage@redhat.com>
2684 Date: Wed Jul 23 18:24:51 2014 -0700
2686 osd/ReplicatedPG: observed INCOMPLETE_CLONES when doing clone subsets
2688 During recovery, we can clone subsets if we know that all clones will be
2689 present. We skip this on caching pools because they may not be; do the
2690 same when INCOMPLETE_CLONES is set.
2692 Signed-off-by: Sage Weil <sage@redhat.com>
2693 (cherry picked from commit 41364711a66c89ce2e94435fe0d54eeda6092614)
2695 commit c42232e3539fc4682e8fec3e064b21acbcc996ed
2696 Author: Sage Weil <sage@redhat.com>
2697 Date: Wed Jul 23 18:23:56 2014 -0700
2699 osd/ReplicatedPG: do not complain about missing clones when INCOMPLETE_CLONES is set
2701 When scrubbing, do not complain about missing cloens when we are in a
2702 caching mode *or* when the INCOMPLETE_CLONES flag is set. Both are
2703 indicators that we may be missing clones and that that is okay.
2706 Signed-off-by: Sage Weil <sage@redhat.com>
2707 (cherry picked from commit 956f28721dd98c5fb9eb410f4fe9e320b3f3eed3)
2709 commit bde3a9f0f610e6af30066dd77949a1249cd658bf
2710 Author: Sage Weil <sage@redhat.com>
2711 Date: Wed Jul 23 18:21:38 2014 -0700
2713 osd/osd_types: add pg_pool_t FLAG_COMPLETE_CLONES
2715 Set a flag on the pg_pool_t when we change cache_mode NONE. This
2716 is because object promotion may promote heads without all of the clones,
2717 and when we switch the cache_mode back those objects may remain. Do
2718 this on any cache_mode change (to or from NONE) to capture legacy
2719 pools that were set up before this flag existed.
2721 Signed-off-by: Sage Weil <sage@redhat.com>
2722 (cherry picked from commit 54bf055c5dadc55acf5731e08712d529b180ffc5)
2724 commit f19aa8ef0a2cbad7af3135defc1fe3485d377409
2725 Author: Joao Eduardo Luis <joao.luis@inktank.com>
2726 Date: Fri Jul 11 00:40:47 2014 +0100
2728 qa/workunits: cephtool: adjust pool name where missing as it has changed
2730 Signed-off-by: Joao Eduardo Luis <joao.luis@inktank.com>
2731 (cherry picked from commit 50e93c2138978f7f7c2fbafacc1611c8705a8eab)
2733 commit 163440a764a6fac8f2efbee6a8aee5cbc460dfe6
2734 Author: Joao Eduardo Luis <joao.luis@inktank.com>
2735 Date: Fri Jul 11 00:41:01 2014 +0100
2737 qa/workunits: cephtool: cleanup after pool creation
2739 Signed-off-by: Joao Eduardo Luis <joao.luis@inktank.com>
2740 (cherry picked from commit 6cd345732b15e84de17d743e06bc4d85569b79d4)
2742 commit c71c99ff21063e26ebfef431f916463f5606aae2
2743 Author: Joao Eduardo Luis <joao.luis@inktank.com>
2744 Date: Tue Jul 8 19:22:49 2014 +0100
2746 qa/workunits: cephtool: pool needs to be a tier to be used as such
2748 Signed-off-by: Joao Eduardo Luis <joao.luis@inktank.com>
2749 (cherry picked from commit 704b0a33f2071eabeb8c5b000a6805ef6d498961)
2751 commit ba0a9bf013e8d056a4126b6840621e5a3a8b37fc
2752 Author: Joao Eduardo Luis <joao.luis@inktank.com>
2753 Date: Tue Jul 8 19:22:01 2014 +0100
2755 qa/workunits: cephtool: test erroneous 'tier remove'
2757 Signed-off-by: Joao Eduardo Luis <joao.luis@inktank.com>
2758 (cherry picked from commit 49db6767152092d503ccf8ead6f7cb069e152a22)
2760 commit ffec081a551817288303cf3fa00c016bbffe114a
2761 Author: Joao Eduardo Luis <joao.luis@inktank.com>
2762 Date: Thu Jul 3 15:32:46 2014 +0100
2764 qa/workunits: cephtool: test get/set on both tier and non-tier pools
2766 Make sure gets and sets of tiering-specific variables succeed on tier
2767 pools and fail on non-tier pools.
2769 Signed-off-by: Joao Eduardo Luis <joao.luis@inktank.com>
2770 (cherry picked from commit 9fea033f30aec44a3273c623ec6c93eb1d7dd26b)
2772 commit 55166c3a098d8751ed6dafb3541e45a6439c5491
2773 Author: Joao Eduardo Luis <joao.luis@inktank.com>
2774 Date: Thu Jul 3 15:24:51 2014 +0100
2776 qa/workunits: cephtool: split get/set on tier pools from get/set tests
2778 Signed-off-by: Joao Eduardo Luis <joao.luis@inktank.com>
2779 (cherry picked from commit df5944955d96c041e65964a13b802028e9700904)
2781 commit 0425eb1dbd156d03a51e91c6f3079b8a687d62f3
2782 Author: Joao Eduardo Luis <joao.luis@inktank.com>
2783 Date: Fri Jun 27 22:20:51 2014 +0100
2785 qa/workunits: cephtool: test for 'osd pool {get,set}-quota'
2787 Signed-off-by: Joao Eduardo Luis <joao.luis@inktank.com>
2788 (cherry picked from commit b927c0de7d5c7a78bf3c133be52cbc1d769974bb)
2790 commit 68197bc28c8960db67c4246d50eb5146659593bd
2791 Author: Joao Eduardo Luis <joao.luis@inktank.com>
2792 Date: Mon Jun 30 17:51:47 2014 +0100
2794 mon: OSDMonitor: 'osd pool' - if we can set it, we must be able to get it
2796 Add support to get the values for the following variables:
2797 - target_max_objects
2799 - cache_target_dirty_ratio
2800 - cache_target_full_ratio
2801 - cache_min_flush_age
2802 - cache_min_evict_age
2804 Signed-off-by: Joao Eduardo Luis <joao.luis@inktank.com>
2805 (cherry picked from commit ddc04c83ff6842ca0b2f804b46099ea729b9fb6b)
2807 commit 01498ce8c897f7e3b66c8054a8b144a505a33010
2808 Author: Loic Dachary <loic@dachary.org>
2809 Date: Tue Jul 1 08:42:58 2014 +0200
2811 qa: support running under non privileged user
2813 If the test is run against a cluster started with vstart.sh (which is
2814 the case for make check), the --asok-does-not-need-root disables the use
2815 of sudo and allows the test to run without requiring privileged user
2818 Signed-off-by: Loic Dachary <loic@dachary.org>
2819 (cherry picked from commit 522174b066044e51a7019bd6cad81117e83c394e)
2821 commit 99259dff5bad93dce928c1ec07a16b34e3dac3dc
2822 Author: Sage Weil <sage@inktank.com>
2823 Date: Mon Jun 30 16:38:10 2014 -0700
2825 qa/workunits/cephtool/test.sh: sudo ceph daemon
2827 Signed-off-by: Sage Weil <sage@inktank.com>
2828 (cherry picked from commit bcc09f93761d46202742ca85cce498a352edd494)
2830 commit 4f911d45632fd23c4124bf6b29287ece473d33b4
2831 Author: Joao Eduardo Luis <joao.luis@inktank.com>
2832 Date: Wed Jun 25 21:25:29 2014 +0100
2834 qa/workunits: cephtool: fix 'osd bench' test
2836 Commit 7dc93a9651f602d9c46311524fc6b54c2f1ac595 fixed an incorrect
2837 behavior with the OSD's 'osd bench' value hard-caps. The test wasn't
2838 appropriately modified unfortunately.
2840 Signed-off-by: Joao Eduardo Luis <joao.luis@inktank.com>
2841 (cherry picked from commit 48e38ac6323f4a0e06b0dedd37ecd10dc339b1e3)
2843 commit bad2f513c5bed4602005eaedff9384012103aaee
2844 Author: Joao Eduardo Luis <joao.luis@inktank.com>
2845 Date: Wed Jun 25 01:48:24 2014 +0100
2847 qa/workunits: cephtool: only run heap profiler test if tcmalloc enabled
2849 Signed-off-by: Joao Eduardo Luis <joao.luis@inktank.com>
2850 (cherry picked from commit 4b0809a13eb363a15e52a6a57372a0a31a64cef8)
2852 commit 26f9860d19c38595bfced76fce864fe9ad9f3a27
2853 Author: Joao Eduardo Luis <joao.luis@inktank.com>
2854 Date: Tue Jun 24 23:16:17 2014 +0100
2856 qa/workunits: cephtool: set +e for the tcmalloc tests
2858 Avoids failing the tests when tcmalloc is not present
2860 Signed-off-by: Joao Eduardo Luis <joao.luis@inktank.com>
2861 (cherry picked from commit 5c4616e100297ba8639919aca7a9cb59e4bda54a)
2863 commit a0720b608db208ac766df8eea53bdbba31d79492
2864 Author: Joao Eduardo Luis <joao.luis@inktank.com>
2865 Date: Tue Jun 24 22:47:31 2014 +0100
2867 qa/workunits: cephtool: delete unnecessary function calls
2869 Signed-off-by: Joao Eduardo Luis <joao.luis@inktank.com>
2870 (cherry picked from commit 67255435151627314cc2fc38732d4fb5efddc3cc)
2872 commit 373224729e9ab2a46afb43b1388d37ff951cc42e
2873 Author: Joao Eduardo Luis <joao.luis@inktank.com>
2874 Date: Tue Jun 24 22:46:44 2014 +0100
2876 qa/workunits: cephtool: disable bash debug when not running tests
2878 Signed-off-by: Joao Eduardo Luis <joao.luis@inktank.com>
2879 (cherry picked from commit 946bd0dad4b027326b03c13405782b99ef0f91b2)
2881 commit 52d733c7202b393b3b7da48e3aa6db275946757a
2882 Author: Joao Eduardo Luis <joao.luis@inktank.com>
2883 Date: Tue Jun 24 22:39:34 2014 +0100
2885 qa/workunits: cephtool: allow running individual tests
2887 Signed-off-by: Joao Eduardo Luis <joao.luis@inktank.com>
2888 (cherry picked from commit 5d26575ef2d31d745ec4aa69ca1501cd76e5e8db)
2890 commit 96e58d6d38969ba8b47b1924205eb34850a24ed6
2891 Author: Joao Eduardo Luis <joao.luis@inktank.com>
2892 Date: Tue Jun 24 20:55:27 2014 +0100
2894 qa/workunits: cephtool: cleanup state after erasure-code-profile test
2896 Signed-off-by: Joao Eduardo Luis <joao.luis@inktank.com>
2897 (cherry picked from commit f4184086d0d647e064e34308f678ef889e13c373)
2899 commit 4cd5d179375c06a25f04ef8150e2a53aabf7427e
2900 Author: Joao Eduardo Luis <joao.luis@inktank.com>
2901 Date: Wed Jun 25 01:17:58 2014 +0100
2903 qa/workunits: cephtool: add/remove comments
2905 Signed-off-by: Joao Eduardo Luis <joao.luis@inktank.com>
2906 (cherry picked from commit 780424df3a107c7da57fc28d64f9e7a4bb47f8e8)
2908 commit 5bcf690b1dbf5a2e4388661f9d280b5934db1c9c
2909 Author: Joao Eduardo Luis <joao.luis@inktank.com>
2910 Date: Thu Jun 26 00:55:46 2014 +0100
2912 qa/workunits: cephtool: split into properly indented functions
2914 The test was a big sequence of commands being run and it has been growing
2915 organically for a while, even though it has maintained a sense of
2916 locality with regard to the portions being tested.
2918 This patch intends to split the commands into functions, allowing for a
2919 better semantic context and easier expansion. On the other hand, this
2920 will also allow us to implement mechanisms to run specific portions of
2921 the test instead of always having to run the whole thing just to test a
2922 couple of lines down at the bottom (or have to creatively edit the test).
2924 Signed-off-by: Joao Eduardo Luis <joao.luis@inktank.com>
2925 (cherry picked from commit 3d14a96f4b2b7094d05ead1dec7547d165857e31)
2928 qa/workunits/cephtool/test.sh
2930 commit 2af751807d1b51348302734c46d506db294303fd
2931 Author: Joao Eduardo Luis <joao.luis@inktank.com>
2932 Date: Wed Jun 25 01:12:55 2014 +0100
2934 qa/workunits: cephtool: move test line to where it's more appropriate
2936 Signed-off-by: Joao Eduardo Luis <joao.luis@inktank.com>
2937 (cherry picked from commit 04658b7b2b5f98ae81ffb3f77303745e6d46eb81)
2939 commit 5bf9e3c7b59d7ce5996d4d05312b4a92f34bbcb6
2940 Author: Joao Eduardo Luis <joao.luis@inktank.com>
2941 Date: Wed Jun 25 01:11:52 2014 +0100
2943 qa/workunits: cephtool: split into functions
2945 Signed-off-by: Joao Eduardo Luis <joao.luis@inktank.com>
2946 (cherry picked from commit db6cc133ba4bb38b3c11eb835fd3983dc2f6b00e)
2949 qa/workunits/cephtool/test.sh
2951 commit c898c9736950437256b8c261106f2827fc160e01
2952 Author: Loic Dachary <loic@dachary.org>
2953 Date: Fri Jun 20 09:38:39 2014 +0200
2955 mon: test that pools used in tiers cannot be removed
2957 Signed-off-by: Loic Dachary <loic@dachary.org>
2958 (cherry picked from commit 39a4b78177cb9896ff35ab05bcf8774bfc934f3a)
2960 commit 5b72674238db80e62cf3d5cec1d94be5c0102388
2961 Author: Joao Eduardo Luis <joao.luis@inktank.com>
2962 Date: Fri May 23 17:02:23 2014 +0100
2964 qa/workunits/cephtool: test setting options using SI units
2966 Signed-off-by: Joao Eduardo Luis <joao.luis@inktank.com>
2967 (cherry picked from commit 38405d3554dfb0caf2b0a2c010b95a61bdb99349)
2970 qa/workunits/cephtool/test.sh
2972 commit aacf67d13f81fcc88881091d175fee9ecb0f8954
2973 Author: Joao Eduardo Luis <joao.luis@inktank.com>
2974 Date: Wed Jul 2 20:19:19 2014 +0100
2976 mon: OSDMonitor: be scary about inconsistent pool tier ids
2978 We may not crash your cluster, but you'll know that this is not something
2979 that should have happened. Big letters makes it obvious. We'd make them
2980 red too if we bothered to look for the ANSI code.
2982 Signed-off-by: Joao Eduardo Luis <joao.luis@inktank.com>
2983 (cherry picked from commit 8e5a8daf98052954f3880d2d3516841b5062466b)
2985 commit af59cbf5854bd0fabeb0ef394d7d5b1207bbd6d3
2986 Author: Joao Eduardo Luis <joao.luis@inktank.com>
2987 Date: Wed Jul 2 19:47:52 2014 +0100
2989 osd: pg_pool_t: clear tunables on clear_tier()
2991 Signed-off-by: Joao Eduardo Luis <joao.luis@inktank.com>
2992 (cherry picked from commit 64bdf6c92bc44adad5a49b9dc4f674789cee80b0)
2994 commit 8f788e14d22c913479630bce9f9c502ac900328c
2995 Author: Joao Eduardo Luis <joao.luis@inktank.com>
2996 Date: Wed Jul 2 19:46:48 2014 +0100
2998 mon: OSDMonitor: limit tier-specific pool set/get on non-tier pools
3002 Signed-off-by: Joao Eduardo Luis <joao.luis@inktank.com>
3003 (cherry picked from commit f131dfbaedf6f451572e7aa3a83f653912122953)
3005 commit 4b768a8fe9146e6ec033d4f6780e0ef373fcfd7a
3006 Author: Sage Weil <sage@redhat.com>
3007 Date: Thu Jul 24 10:06:31 2014 -0700
3009 mon/OSDMonitor: improve no-op cache_mode set check
3011 If we have a pending pool value but the cache_mode hasn't changed, this is
3012 still a no-op (and we don't need to block).
3015 Signed-off-by: Sage Weil <sage@redhat.com>
3016 (cherry picked from commit 67d13d76f5692fa20649ea877f254c34094c11f6)
3018 commit c3f403293c7f8d946f66a871aa015a558120ce78
3019 Author: Joao Eduardo Luis <joao.luis@inktank.com>
3020 Date: Sat May 3 03:42:19 2014 +0100
3022 mon: OSDMonitor: disallow nonsensical cache-mode transitions
3026 Signed-off-by: Joao Eduardo Luis <joao.luis@inktank.com>
3027 (cherry picked from commit fd970bbc95d89bf66c9551feca17ac0afbf4d1e2)
3029 commit d9aacd96acfe24b15c4e3b2441838945de1ab720
3030 Author: Joao Eduardo Luis <joao.luis@inktank.com>
3031 Date: Tue Apr 29 18:03:56 2014 +0100
3033 mon: OSDMonitor: return immediately if 'osd tier cache-mode' is a no-op
3035 Signed-off-by: Joao Eduardo Luis <joao.luis@inktank.com>
3036 (cherry picked from commit d01aa5bff30441eec1ffaa3e59a21187f8478475)
3038 commit 981add9cad5f2db06a4932be228eb0ff30254a9b
3039 Author: Sage Weil <sage@redhat.com>
3040 Date: Tue Jul 22 18:01:14 2014 -0700
3042 osd/ReplicatedPG: debug obc locks
3044 Signed-off-by: Sage Weil <sage@redhat.com>
3045 (cherry picked from commit 356af4bf46d6387e2f1a59646548f9a77e49e5f8)
3047 commit e32e2559c8ff408fff81cce0f15d6601dcd4a2f7
3048 Author: Sage Weil <sage@redhat.com>
3049 Date: Tue Jul 22 13:16:11 2014 -0700
3051 osd/ReplicatedPG: greedily take write_lock for copyfrom finish, snapdir
3053 In the cases where we are taking a write lock and are careful
3054 enough that we know we should succeed (i.e, we assert(got)),
3055 use the get_write_greedy() variant that skips the checks for
3056 waiters (be they ops or backfill) that are normally necessary
3057 to avoid starvation. We don't care about staration here
3058 because our op is already in-progress and can't easily be
3059 aborted, and new ops won't start because they do make those
3063 Signed-off-by: Sage Weil <sage@redhat.com>
3064 (cherry picked from commit 6fe27823b8459271bf0c0e807493bb7cf1e4559b)
3066 commit 68651e9137df2b1034a5b470bbb2fc1d19d8debb
3067 Author: Sage Weil <sage@redhat.com>
3068 Date: Tue Jul 22 13:11:42 2014 -0700
3070 osd: allow greedy get_write() for ObjectContext locks
3072 There are several lockers that need to take a write lock
3073 because there is an operation that is already in progress and
3074 know it is safe to do so. In particular, they need to skip
3075 the starvation checks (op waiters, backfill waiting).
3077 Signed-off-by: Sage Weil <sage@redhat.com>
3078 (cherry picked from commit 09626501d7a0ff964027fd7a534465b76bad23cb)
3080 commit a2681c93fb30b912f653b596e27eb8b1bc5c6b78
3081 Author: Sage Weil <sage@inktank.com>
3082 Date: Wed Jul 2 10:38:43 2014 -0700
3084 qa/workunits/rest/test.py: make osd create test idempotent
3086 Avoid possibility that we create multiple OSDs do to retries by passing in
3087 the optional uuid arg. (A stray osd id will make the osd tell tests a
3088 few lines down fail.)
3091 Signed-off-by: Sage Weil <sage@inktank.com>
3092 (cherry picked from commit bb3e1c92b6682ed39968dc5085b69c117f43cbb0)
3094 commit 7a58da53ebfcaaf385c21403b654d1d2f1508e1a
3095 Author: Lluis Pamies-Juarez <lluis@pamies.cat>
3096 Date: Wed Jun 18 10:00:09 2014 -0700
3098 enforce rados put aligment
3100 Signed-off-by: Lluis Pamies-Juarez <lluis.pamies-juarez@hgst.com>
3101 (cherry picked from commit 304b08a23a3db57010078046955a786fe3589ef8)
3102 (cherry picked from commit dece65064d949b5afcc359cd408615883b5e002a)
3104 commit 7ae15549a795ffd9ffcd34ea1d5c101b90eee26d
3105 Author: Sage Weil <sage@redhat.com>
3106 Date: Thu Jul 31 15:39:40 2014 -0700
3108 use llrintl when converting double to micro
3110 This avoids rounding error (noticeable on i386).
3113 Signed-off-by: Sage Weil <sage@redhat.com>
3114 (cherry picked from commit 80911736bd61b6b88eac0974d24f21c15c5385a4)
3117 src/mon/OSDMonitor.cc
3119 commit f5fc24368756c794d186a0990241e0dc0adfc03c
3120 Author: Sage Weil <sage@redhat.com>
3121 Date: Thu Jul 24 18:22:22 2014 -0700
3123 msg/SimpleMessenger: drop local_conneciton priv link on shutdwon
3125 This breaks ref cycles between the local_connection and session, and let's
3126 us drop the explicit set_priv() calls in OSD::shutdown().
3128 Signed-off-by: Sage Weil <sage@redhat.com>
3129 (cherry picked from commit 63c1711a9e237c14d137131f56751970ad1567b1)
3131 commit b6d8feab85fda97bf3a68980249a71846778a157
3132 Author: Loic Dachary <loic@dachary.org>
3133 Date: Thu Jun 19 00:49:13 2014 +0200
3135 erasure-code: create default profile if necessary
3137 After an upgrade to firefly, the existing Ceph clusters do not have the
3138 default erasure code profile. Although it may be created with
3140 ceph osd erasure-code-profile set default
3142 it was not included in the release notes and is confusing for the
3145 The *osd pool create* and *osd crush rule create-erasure* commands are
3146 modified to implicitly create the default erasure code profile if it is
3149 In order to avoid code duplication, the default erasure code profile
3150 code creation that happens when a new firefly ceph cluster is created is
3151 encapsulated in the OSDMap::get_erasure_code_profile_default method.
3153 Conversely, handling the pending change in OSDMonitor is not
3154 encapsulated in a function but duplicated instead. If it was a function
3155 the caller would need a switch to distinguish between the case when goto
3156 wait is needed, or goto reply or proceed because nothing needs to be
3157 done. It is unclear if having a function would lead to smaller or more
3160 http://tracker.ceph.com/issues/8601 Fixes: #8601
3163 Signed-off-by: Loic Dachary <loic@dachary.org>
3164 (cherry picked from commit 4e1405e7720eda71a872c991045ac8ead6f3e7d8)
3166 commit 2fff5f031b2cc00b55bd1ada4f83e4838fe59ddd
3167 Author: Loic Dachary <loic@dachary.org>
3168 Date: Mon Jan 13 17:54:08 2014 +0100
3170 common: s/stringstream/ostream/ in str_map
3172 There is no need to specialize more than ostream : it only makes it
3173 impossible to use cerr or cout as a parameter to str_map.
3175 Reviewed-By: Christophe Courtaut <christophe.courtaut@gmail.com>
3176 Signed-off-by: Loic Dachary <loic@dachary.org>
3177 (cherry picked from commit 6aa45b133956b974a992b372496b90c908d94f12)
3179 commit d48a737a6f8daa463ddf82106455b2efe7715a33
3180 Author: Samuel Just <sam.just@inktank.com>
3181 Date: Mon Jun 2 13:44:49 2014 -0700
3183 osd/: in scrub, check for and remove obsolete rollback objs
3185 Signed-off-by: Samuel Just <sam.just@inktank.com>
3187 commit 953c33edb640f6eb0c3f3a0d07c061c0d324242b
3188 Author: Samuel Just <sam.just@inktank.com>
3189 Date: Fri May 30 16:53:37 2014 -0700
3191 osd/: plumb rollback_trim_to through interfaces and messages
3193 We simply trim the roll back info up to min_last_complete_ondisk
3194 since we cannot correctly rollback past that point anyway!
3197 -> PGBackend::submit_transaction
3199 -> PGBackend::Listener::log_operation
3202 Signed-off-by: Samuel Just <sam.just@inktank.com>
3204 commit ac11ca40b4f4525cbe9b1778b1c5d9472ecb9efa
3205 Author: Samuel Just <sam.just@inktank.com>
3206 Date: Fri May 30 16:27:32 2014 -0700
3208 PGLog: add rollback_info_trimmed_to_riter to IndexedLog and use in PG
3210 Signed-off-by: Samuel Just <sam.just@inktank.com>
3212 commit 2207ed12fa544afc48d6d28f199381ccf0367292
3213 Author: Samuel Just <sam.just@inktank.com>
3214 Date: Fri May 30 15:40:15 2014 -0700
3216 PGLog: fix logging whitespace in trim
3218 Signed-off-by: Samuel Just <sam.just@inktank.com>
3220 commit d72eec0341e02881229d0e4ffeafcae92eb33fb6
3221 Author: Samuel Just <sam.just@inktank.com>
3222 Date: Fri Jun 27 10:51:13 2014 -0700
3224 PG: block outgoing peering messages until the last interval is durable
3226 Signed-off-by: Samuel Just <sam.just@inktank.com>
3228 commit f9f89be38333d3817a0eafc52a961365e21e1f3a
3229 Author: Samuel Just <sam.just@inktank.com>
3230 Date: Wed May 21 13:09:33 2014 -0700
3232 PG: add machinery to temporarily buffer outgoing recovery messages
3234 Signed-off-by: Samuel Just <sam.just@inktank.com>
3236 commit f7498124325fe2269c33e45fae4c8181c8f7be14
3237 Author: Samuel Just <sam.just@inktank.com>
3238 Date: Mon Jun 9 11:00:32 2014 -0700
3240 os/: add async flush_commit() method to Sequencer and implementations
3242 Signed-off-by: Samuel Just <sam.just@inktank.com>
3244 commit 669b60571d6db9a9f74020e6a43b12b82edee41e
3245 Author: Samuel Just <sam.just@inktank.com>
3246 Date: Fri May 30 15:18:53 2014 -0700
3248 PGLog: initialize complete_to
3250 Signed-off-by: Samuel Just <sam.just@inktank.com>
3252 commit 8679cdbeea6350fd860a9419f54578a90482895c
3253 Author: Samuel Just <sam.just@inktank.com>
3254 Date: Thu May 29 12:10:31 2014 -0700
3256 osd_types: add pg_log_t::rollback_info_trimmed_to
3258 Signed-off-by: Samuel Just <sam.just@inktank.com>
3260 commit 3bbe29eb5cca7cc42266a6c5561ce68feb4072de
3261 Author: Samuel Just <sam.just@inktank.com>
3262 Date: Tue May 20 16:05:01 2014 -0700
3264 common/Finisher: add queue(list<Context*>&)
3266 Signed-off-by: Samuel Just <sam.just@inktank.com>