commit ad85b8bfafea6232d64cb7ba76a8b6e8252fa0c7 Author: Gary Lowell Date: Thu Oct 3 22:41:31 2013 +0000 v0.67.4 commit 5cd66d3b4bca92b402c95ab256fbc3f0329c446f Author: Yehuda Sadeh Date: Fri Sep 20 14:04:47 2013 -0700 rgw: fix keystone token expiration test Fixes: #6360 The test was inverted, need expiration to be greater than current time in order for token to be valid. Signed-off-by: Yehuda Sadeh Reviewed-by: Sage Weil commit e0203c61a3f45fdd6d3d3ece26fef6152bdc036d Author: David Zafman Date: Wed Sep 11 16:55:06 2013 -0700 osd/OSD.cc: Use MIN() so that we don't exceed osd_recovery_max_active Caused by 944f3b73531af791c90f0f061280160003545c63 Fixes: #6291 Backport: dumpling Signed-off-by: David Zafman Reviewed-by: Samuel Just (cherry picked from commit 139a714e13aa3c7f42091270b55dde8a17b3c4b8) Conflicts: src/osd/OSD.cc commit c376708358cedb5561fbb43e9b9e622df3ea7a58 Author: Joao Eduardo Luis Date: Wed Sep 25 22:08:24 2013 +0100 mon: OSDMonitor: do not write full_latest during trim On commit 81983bab we patched OSDMonitor::update_from_paxos() such that we write the latest full map version to 'full_latest' each time the latest full map was built from the incremental versions. This change however clashed with OSDMonitor::encode_trim_extra(), which also wrote to 'full_latest' on each trim, writing instead the version of the *oldest* full map. This duality of behaviors could lead the store to an inconsistent state across the monitors (although there's no sign of it actually imposing any issues besides rebuilding already existing full maps on some monitors). We now stop OSDMonitor::encode_trim_extra() from writing to 'full_latest'. This function will still write out the oldest full map it has in the store, but it will no longer write to full_latest, instead leaving it up to OSDMonitor::update_from_paxos() to figure it out -- and it already does. Fixes: #6378 Backport: dumpling Signed-off-by: Joao Eduardo Luis Reviewed-by: Sage Weil (cherry picked from commit bd0f29a2c28cca496ec830eac932477ebf3182ba) commit de40d0b3e35ab0124cd3c4ebfcaa435ab8abfab9 Author: Sage Weil Date: Tue Oct 1 15:53:42 2013 -0700 crush: invalidate rmap on create (and thus decode) If we have an existing CrushWrapper object and decode from a bufferlist, reset build_rmaps so that they get rebuilt. Remove the build_rmaps() all in decode that was useless on a redecode (because have_rmaps == true in that case and it did nothing). Fixes: #6442 Backport: dumpling, maybe cuttlefish Signed-off-by: Sage Weil Reviewed-by: Joao Eduardo Luis (cherry picked from commit 9b7a2ae329b6a511064dd3d6e549ba61f52cfd21) commit 32f5233288c47d95b87c0a9cab5f9c2ffcf15417 Author: Dan Mick Date: Mon Sep 30 14:58:11 2013 -0700 Invoke python with /usr/bin/env python instead of directly Fixes: #6311 Signed-off-by: Dan Mick (cherry picked from commit b9000b314b9166845ff302d4a827a996775d9a14) commit 66aeca5a9079be398403bbff67bd5bf68c6fb111 Author: Sage Weil Date: Wed Sep 25 10:10:21 2013 -0700 qa/workunits/mon/crush_ops.sh: fix test Fix root. Fixes: #6392 Signed-off-by: Sage Weil (cherry picked from commit c8cae87e9e08468cc86145e0fd60c05d12826239) commit beb366302a125dd422c4f092b12eb541cb3bc788 Author: Sage Weil Date: Mon Sep 23 09:04:34 2013 -0700 Revert "ceph: parse CEPH_ARGS environment variable" This reverts commit 67a95b9880c9bc6e858150352318d68d64ed74ad. We now put CEPH_ARGS in the actual args we parse in python, which are passed to rados piecemeal later. This lets you put things like --id ... in there that need to be parsed before librados is initialized. (cherry picked from commit 97f462be4829f0167ed3d65e6694dfc16f1f3243) commit b475ff9576f145d31c053213c699e13df76d2bcb Author: Benoît Knecht Date: Mon Sep 23 15:58:42 2013 +0200 Add CEPH_ARGS at the end of sys.argv This allows, for instance, to pass a different client name to ceph by exporting CEPH_ARGS="--id client_id". Signed-off-by: Benoît Knecht Signed-off-by: Sage Weil (cherry picked from commit 30abe3244c86cbbe1f5b005850c29c9c0eafcad4) commit 94548b4b67cca37366c7d8719209a6d2e7956811 Author: Sage Weil Date: Tue Sep 24 15:26:03 2013 -0700 mon/OSDMonitor: fix 'ceph osd crush reweight ...' The adjust method returns a count of adjusted items. Add a test. Fixes: #6382 Backport: dumpling Signed-off-by: Sage Weil Reviewed-by: Dan Mick (cherry picked from commit 3de32562b55c6ece3a6ed783c36f8b9f21460339) commit 00ff7f5c20e13869d0694379739ba4e61d44b97c Author: Joao Eduardo Luis Date: Tue Sep 10 00:20:41 2013 +0100 qa: workunits: mon: crush_ops: test 'ceph osd crush move' Signed-off-by: Joao Eduardo Luis (cherry picked from commit 3bc618b7b46496c5110edde0da9cae5d3e68e0e1) commit 0ff5b4a96833681e92cc41f019a569134474f4cf Author: Loic Dachary Date: Tue Sep 24 19:04:23 2013 +0200 osd: change warn_interval_multiplier to uint32_t to prevent overflow in OpTracker::check_ops_in_flight when multiplying warn_interval_multiplier *= 2 Backport: cuttlefish, dumpling http://tracker.ceph.com/issues/6370 fixes #6370 Signed-off-by: Loic Dachary (cherry picked from commit 1bce1f009bffd3e28025a08775fec189907a81db) commit fb15040b6cec6221baa550ddfffade823f784c4a Author: David Zafman Date: Mon Sep 9 13:01:12 2013 -0700 crushtool: do not dump core with non-unique bucket IDs Return -EEXIST on duplicate ID BUG FIX: crush_add_bucket() mixes error returns and IDs Add optional argument to return generated ID Fixes: #6246 Signed-off-by: David Zafman Reviewed-by: Sage Weil (cherry picked from commit 8c76f3a0f9cf100ea2c941dc2b61c470aa5033d7) commit 410db3f30c6eb54b807908c1f251ad4026e7d446 Author: Joao Eduardo Luis Date: Fri Sep 20 17:06:30 2013 +0100 qa: workunits: cephtool: check if 'heap' commands are parseable Signed-off-by: Joao Eduardo Luis (cherry picked from commit b1eeaddd5f214c1b0883b44fc8cae07c649be7c4) commit 062060a38bb26ff260cc51accc534413d726de49 Author: Joao Eduardo Luis Date: Fri Sep 20 17:50:27 2013 +0100 osd: OSD: add 'heap' command to known osd commands array Must have been forgotten during the cli rework. Backport: dumpling Signed-off-by: Joao Eduardo Luis (cherry picked from commit 296f2d0db31e9f5a59a3a62a1e95b6c440430fa3) commit 3f32f57b98e0224a1d30b2a81d7d260be0f53800 Author: Joao Eduardo Luis Date: Fri Sep 20 16:43:27 2013 +0100 mds: MDS: pass only heap profiler commands instead of the whole cmd vector The heap profiler doesn't care, nor should it, what our command name is. It only cares about the commands it handles. Backport: dumpling Signed-off-by: Joao Eduardo Luis (cherry picked from commit 238fe272c6bdb62d4e57fd8555c0136de99c8129) commit 46dcc46617d8f35ab8433540b22343ddcbcc3716 Author: Joao Eduardo Luis Date: Fri Sep 20 16:41:14 2013 +0100 perfglue/heap_profiler.cc: expect args as first element on cmd vector We used to pass 'heap' as the first element of the cmd vector when handling commands. We haven't been doing so for a while now, so we needed to fix this. Not expecting 'heap' also makes sense, considering that what we need to know when we reach this function is what command we should handle, and we should not care what the caller calls us when handling his business. Fixes: #6361 Backport: dumpling Signed-off-by: Joao Eduardo Luis (cherry picked from commit c98b910d49bd2b46ceafdc430044a31524c29f5b) commit 9dc5f15fbae22244ad1f62925e17c9d81e856e55 Author: Yehuda Sadeh Date: Mon Sep 16 14:35:25 2013 -0700 rgw: destroy get_obj handle in copy_obj() Fixes: #6176 Backport: dumpling We take different code paths in copy_obj, make sure we close the handle when we exit the function. Move the call to finish_get_obj() out of copy_obj_data() as we don't create the handle there, so that should makes code less confusing and less prone to errors. Also, note that RGWRados::get_obj() also calls finish_get_obj(). For everything to work in concert we need to pass a pointer to the handle and not the handle itself. Therefore we needed to also change the call to copy_obj_data(). Reviewed-by: Josh Durgin Signed-off-by: Yehuda Sadeh (cherry picked from commit 9e98620e4325d15c88440a890b267131613e1aa1) commit 471233e98a9f64ad513a4a196b7661b80534cb00 Author: Joao Eduardo Luis Date: Mon Sep 9 23:14:11 2013 +0100 mon: MonCommands: expect a CephString as 1st arg for 'osd crush move' Fixes: #6230 Signed-off-by: Joao Eduardo Luis (cherry picked from commit 7d3799fde19138f957f26ec6be10a8a0000fc1f0) commit 2908225092bd2aa1b8afcb7848c1cdac5bd9e638 Author: Sage Weil Date: Mon Sep 23 16:23:33 2013 -0700 osd: revert 'osd max xattr size' limit Set it to 0 (unlimited) for now. Backport: dumpling Signed-off-by: Sage Weil Reviewed-by: Yehuda Sadeh (cherry picked from commit abb88d70643c3a76435b7a9d5b04ff29f7502361) commit b3d3b3747c1eef695138dac828e5fcb435309c7b Author: Greg Farnum Date: Wed Sep 11 16:24:32 2013 -0700 mds: be more careful about decoding LogEvents We need to wrap the full decode section or we can abort the process if there's an issue (which we may want to just skip by). Signed-off-by: Greg Farnum Reviewed-by: Sage Weil (cherry picked from commit 73289b34b0be5b6612e38944794d59b5e789f841) commit 06c58132199ed22413b509dfa751321ccdb24225 Author: Joao Eduardo Luis Date: Tue Sep 17 17:58:20 2013 +0100 mon: OSDMonitor: multiple rebuilt full maps per transaction Signed-off-by: Joao Eduardo Luis (cherry picked from commit 0d20cae0be701c5b6151a26ee5e4fe24d89aa20a) commit 65bbcaf4b68790dae4506c1f5db237077e1ff0ae Author: Joao Eduardo Luis Date: Sun Sep 15 21:03:50 2013 +0100 mon: OSDMonitor: update latest_full while rebuilding full maps Not doing so will make the monitor rebuild the osdmap full versions, even though they may have been rebuilt before, every time the monitor starts. This mostly happens when the cluster is left in an unhealthy state for a long period of time and incremental versions build up. Even though we build the full maps on update_from_paxos(), not updating 'full_latest' leads to the situation initially described. Fixes: #6322 Signed-off-by: Joao Eduardo Luis (cherry picked from commit 81983bab3630520d6c7ee9b7e4a747bc17b8c5c3) commit 9b9edb04581cca15e67c567332529f5b3f426743 Author: Joao Eduardo Luis Date: Sun Sep 15 21:00:55 2013 +0100 mon: OSDMonitor: smaller transactions when rebuilding full versions Otherwise, for considerably sized rebuilds, the monitor will not only consume vast amounts of memory, but it will also have troubles committing the transaction. Anyway, it's also a good idea to adjust transactions to the granularity we want, and to be fair we care that each rebuilt full map gets to disk, even if subsequent full maps don't (those can be rebuilt later). Fixes: #6323 Signed-off-by: Joao Eduardo Luis (cherry picked from commit 4ac1570c5cdcd6556dc291cc6d7878fd92d343ae) commit 298811f7a15541b9ec1015c416ad2aa075be5691 Author: Joao Eduardo Luis Date: Wed Aug 28 15:51:01 2013 +0100 mon: OSDMonitor: check if pool is on unmanaged snaps mode on mk/rmsnap Backport: dumpling Fixes: #6047 Signed-off-by: Joao Eduardo Luis (cherry picked from commit fab79543c54c2e446d3f76520d7906645c6b0075) commit a992664435db9dde3745eb7f354cce3fc5400a47 Author: Yehuda Sadeh Date: Thu Sep 12 14:32:17 2013 -0700 lru_map: don't use list::size() replace list::size() with map::size(), which should have a constant time complexity. Reviewed-by: Sage Weil Signed-off-by: Yehuda Sadeh (cherry picked from commit 7c1d2ded8fa8061bf3f14932800998b963745dd1) commit 788546ea71c994ff35323747294ed9c177fe7020 Author: Yehuda Sadeh Date: Thu Sep 12 14:30:19 2013 -0700 common/lru_map: rename tokens to entries This code was originally used in a token cache, now as a generic infrastructure rename token fields. Reviewed-by: Sage Weil Signed-off-by: Yehuda Sadeh (cherry picked from commit 532e41a9985a16b35a6e49cdcba38af0ad166fa8) commit babeb00c42af760b3e7575166479e95365cfcc0a Author: Yehuda Sadeh Date: Wed Sep 18 10:37:21 2013 -0700 rgw: use bufferlist::append() instead of bufferlist::push_back() push_back() expects char *, whereas append can append a single char. Appending a NULL char to push_back is cast as a NULL pointer which is bad. Signed-off-by: Yehuda Sadeh Reviewed-by: Josh Durgin (cherry picked from commit 08fe028bad13096d482454a2f303158727c363ff) commit daf85c45dd4d158bc7c33a2fb784857bc7db35cd Author: Yehuda Sadeh Date: Wed Sep 11 13:46:31 2013 -0700 rgw: NULL terminate buffer before parsing it Fixes: #6175 Backport: dumpling We get a buffer off the remote gateway which might not be NULL terminated. The JSON parser needs the buffer to be NULL terminated even though we provide a buffer length as it calls strlen(). Reviewed-by: Josh Durgin Signed-off-by: Yehuda Sadeh (cherry picked from commit e7f7483192cddca1159aba439ce62b1e78669d51) commit c73040a5518971813b9ebaae1624c5bacef315d0 Author: Yehuda Sadeh Date: Wed Sep 11 22:30:12 2013 -0700 rgw: don't call list::size() in ObjectCache Fixes: #6286 Use an external counter instead of calling list::size() Reviewed-by: Sage Weil Signed-off-by: Yehuda Sadeh (cherry picked from commit 31e3a51e933429d286104fe077e98ea883437ad6) commit a855aba9d18936e9a060119e041518790cd4b831 Author: Yehuda Sadeh Date: Tue Sep 10 12:18:55 2013 -0700 rgw: drain pending requests before completing write Fixes: #6268 When doing aio write of objects (either regular or multipart parts) we need to drain pending aio requests. Otherwise if gateway goes down then object might end up corrupted. Reviewed-by: Josh Durgin Signed-off-by: Yehuda Sadeh (cherry picked from commit 626669afaa333d73707553a85f5c874e99e9cbd8) commit 670db7e80ddc9c26c43a4f66907a5996ce207c4d Author: Yehuda Sadeh Date: Fri Sep 6 22:33:38 2013 -0700 rgw: fix get cors, delete cors Remove a couple of variables that overrode class member. Not really clear how it was working before, might have been a bad merge / rebase. Signed-off-by: Yehuda Sadeh (cherry picked from commit 13872785aeeddbe1b8dd97e49fd6a2d879514f8d) commit a304016fa01b02efd500135c00b9bf3407a9999c Merge: 408cd61 ac0a30f Author: Yehuda Sadeh Date: Wed Sep 11 09:47:10 2013 -0700 Merge branch 'wip-6078-dumpling' into dumpling Reviewed-by: Josh Durgin commit ac0a30feb8c64a3b80d9c519a7b561213403afab Author: Yehuda Sadeh Date: Wed Aug 28 21:25:20 2013 -0700 rgw: fix certain return status cases in CORS Change return values in certain cases, reorder checks, etc. Signed-off-by: Yehuda Sadeh commit 13b28cc3f1eb8ef42875b630c485ee0105cd244a Author: Yehuda Sadeh Date: Wed Aug 28 21:24:36 2013 -0700 rgw: add COPY method to be handled by CORS Was missing this http method. Signed-off-by: Yehuda Sadeh commit d45c87ea738807487e72c0719b0d3d459cbe19e9 Author: Yehuda Sadeh Date: Tue Aug 27 19:38:45 2013 -0700 rgw: fix CORS rule check Signed-off-by: Yehuda Sadeh commit 986fa92a7a1d88111ba28457160adfcfdaabc5d2 Author: Yehuda Sadeh Date: Tue Aug 27 19:38:18 2013 -0700 rgw: don't handle CORS if rule not found (is NULL) Signed-off-by: Yehuda Sadeh commit 71873aba6553492d3ad71596cefd7c841030a277 Author: Yehuda Sadeh Date: Thu Aug 22 13:38:55 2013 -0700 rgw: tie CORS header response to all relevant operations Have the CORS responses on all relevant operations. Also add headers on failure cases. Signed-off-by: Yehuda Sadeh commit 94e7b594d85dbd26e58d823b41f418032e9f163f Author: Yehuda Sadeh Date: Thu Aug 22 10:00:53 2013 -0700 rgw: add a generic CORS response handling Signed-off-by: Yehuda Sadeh commit c3385d8a102faf5379559bb98cf89637ceda1579 Author: Yehuda Sadeh Date: Wed Aug 21 17:22:46 2013 -0700 rgw: OPTIONS request doesn't need to read object info This is a bucket-only operation, so we shouldn't look at the object. Object may not exist and we might respond with Not Exists response which is not what we want. Signed-off-by: Yehuda Sadeh commit a5fdd44e5d8ce4b8d82273d83e27aea19e63aa7c Author: Yehuda Sadeh Date: Wed Aug 21 14:43:28 2013 -0700 rgw: remove use of s->bucket_cors Some old code still tried to use s->bucket_cors, which was abandoned in a cleanup work. Signed-off-by: Yehuda Sadeh