RepoMirrors/ceph
1
0
Fork 0
mirror of https://github.com/ceph/ceph synced 2025-02-24 11:37:37 +00:00
Code Issues Packages Projects Releases Wiki Activity
79,910 Commits 216 Branches 395 Tags 1.4 GiB
f94322066f
Commit Graph

79910 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Patrick Donnelly
6809dbed6b
Merge PR #18109 into master 
* refs/pull/18109/head:
	mds: clean up non-existent data pools in MDSMap
	mds: reduce variable scope

Reviewed-by: John Spray <john.spray@redhat.com>
 2017-10-24 12:17:25 -07:00
Patrick Donnelly
db5df3c605
Merge PR #18366 into master 
* refs/pull/18366/head:
	mon: s/mds_blacklist_interval/mon_mds_blacklist_interval/
	mon: tidy up mds_blacklist_interval

Reviewed-by: Patrick Donnelly <pdonnell@redhat.com>
 2017-10-24 12:17:19 -07:00
Patrick Donnelly
710cf45e48
Merge PR #18384 into master 
* refs/pull/18384/head:
	mds: disable early reply for dir layout and quota related requests
	mds: make mksnap/setlayout wait for unsafe requests on other mds
	mds: don't rdlock locks in replica object while auth mds is recovering
	mds: move ScatterLock::state_flags into SimpleLock

Reviewed-by: Patrick Donnelly <pdonnell@redhat.com>
 2017-10-24 12:17:14 -07:00
Patrick Donnelly
3e9d8155e9
Merge PR #18393 into master 
* refs/pull/18393/head:
	ceph_context: re-expand admin_socket metavariables in child process

Reviewed-by: Patrick Donnelly <pdonnell@redhat.com>
 2017-10-24 12:17:08 -07:00
Mykola Golub
b0c0e956ce Merge pull request #18490 from dillaman/wip-21894 
rbd-mirror: strip environment/CLI overrides for remote cluster

Reviewed-by: Mykola Golub <to.my.trociny@gmail.com>
 2017-10-24 22:14:13 +03:00
Patrick Donnelly
b897ef9445
qa: use fs rm_data_pool 
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
 2017-10-24 11:51:02 -07:00
Patrick Donnelly
0e4027efdc
qa: fix mdsmap lookup 
We were returning the file system structure and not the mdsmap.

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
 2017-10-24 11:36:06 -07:00
Patrick Donnelly
3a5f090a1e
qa: remove usage of mds dump 
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
 2017-10-24 11:32:43 -07:00
Andrew Schoen
0d1163bd60 Merge pull request #18514 from ceph/wip-fix-cv-create-path 
ceph-volume lvm.prepare update to use create_osd_path

Reviewed-by: Andrew Schoen <aschoen@redhat.com>
 2017-10-24 12:40:01 -05:00
Alfredo Deza
bafb721446 ceph-volume lvm.prepare update to use create_osd_path 
Signed-off-by: Alfredo Deza <adeza@redhat.com>
 2017-10-24 13:29:18 -04:00
Kefu Chai
e7539a9a3e Merge pull request #18500 from scienceluo/wip-luo-fix-doc-log-level-branch 
doc: complete and update the subsystem logging level info table

Reviewed-By: Kefu Chai <kchai@redhat.com>
 2017-10-24 23:03:09 +08:00
Jianpeng Ma
6c81c9bb09 osd/PrimaryLogPG: add condition "is_chunky_scrub_active" to check object in chunky_scrub. 
Avoid every time call scrubber.write_block_by_scrub. Most time scrubber
is inactive. And compare to write_block_by_scrub, is_chunky_scrub_active
is light.

Signed-off-by: Jianpeng Ma <jianpeng.ma@intel.com>
 2017-10-24 22:07:18 +08:00
Sage Weil
2207607f3c osd/PrimaryLogPG: clear DEGRADED at recovery completion even if more backfill 
We may have log recovery *and* backfill to do, but cease to be degraded
as soon as the log recovery portion is done.  If that's the case, clear
the DEGRADED bit so that the PG state is not misleading.

Signed-off-by: Sage Weil <sage@redhat.com>
 2017-10-24 08:59:22 -05:00
Sage Weil
8b93144a83 osd/PG: simplify Recovering state change 
Signed-off-by: Sage Weil <sage@redhat.com>
 2017-10-24 08:59:18 -05:00
Sage Weil
937af7067e doc/release-notes: fix backticks 
Signed-off-by: Sage Weil <sage@redhat.com>
 2017-10-24 08:46:07 -05:00
Kefu Chai
aee80dca92 Merge pull request #18467 from yuyuyu101/wip-21883 
msg/async/AsyncConnection: state will be NONE if replacing by another one

Reviewed-By: Kefu Chai <kchai@redhat.com>
 2017-10-24 21:19:43 +08:00
Kefu Chai
7de3b20fba Merge pull request #18469 from yuyuyu101/wip-remove-legacy 
msg/async/AsyncConnection: remove legacy feature case handle

Reviewed-By: Kefu Chai <kchai@redhat.com>
 2017-10-24 21:18:15 +08:00
Kefu Chai
7f0b878ade Merge pull request #18415 from majianpeng/msg-cleanup 
msg/async: batch handle msg_iovlen

Reviewed-by: Haomai Wang <haomai@xsky.com>
Reviewed-By: Kefu Chai <kchai@redhat.com>
 2017-10-24 21:17:25 +08:00
Kefu Chai
c6b8792677 Merge pull request #18397 from tchaikov/wip-mon-opts 
common/options: set max_background_jobs instead of max_background_compactions

Reviewed-by: Josh Durgin <jdurgin@redhat.com>
 2017-10-24 21:16:30 +08:00
Kefu Chai
d0b7a1f2f5 Merge pull request #18382 from ceph/wip-vstart-mgr-dashboard 
cmake: add cython_rbd as a dependency to vstart target

Reviewed-by: John Spray <john.spray@redhat.com>
Reviewed-by: Kefu Chai <kchai@redhat.com>
 2017-10-24 21:15:10 +08:00
Alfredo Deza
72dba6c716 Merge pull request #18480 from tchaikov/wip-doc-cleanup 
doc: various cleanups

Reviewed-by: Alfredo Deza <adeza@redhat.com>
 2017-10-24 07:03:29 -04:00
Leo Zhang
c3f090c515 OSD: drop unsed parameter passed to check_osdmap_features 
Signed-off-by: Leo Zhang <nguzcf@gmail.com>
 2017-10-24 16:43:22 +08:00
Mykola Golub
f2f893aeb1 Merge pull request #18405 from dillaman/wip-21849 
librbd: default to sparse-reads for any IO operation over 64K

Reviewed-by: Mykola Golub <to.my.trociny@gmail.com>
 2017-10-24 11:02:49 +03:00
Kefu Chai
e479bd0ac1 rpm,debian: use download.ceph.com/tarballs instead of ceph.com/download 
Signed-off-by: Kefu Chai <kchai@redhat.com>
 2017-10-24 16:00:09 +08:00
Kefu Chai
127d25ccbd doc: fix warning reported by "build-doc linkcheck" 
all the HTTP 301 (moved permanently) should be killed.

Signed-off-by: Kefu Chai <kchai@redhat.com>
 2017-10-24 15:58:48 +08:00
xie xingguo
c84d613370 osd/PrimaryLogPG: update modified_ranges correctly on truncating 
This is necessary since we are modifying the object contents.

Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>
 2017-10-24 15:45:27 +08:00
Brad Hubbard
2094fda765 osd: Add a flag to ScrubMap to signal check needed 
If we know none of the scrub maps have errors early in
be_large_omap_check() we can return without doing a lot of unnecessary
work.

Signed-off-by: Brad Hubbard <bhubbard@redhat.com>
 2017-10-24 17:27:57 +10:00
Brad Hubbard
71bf04775b osd: Warn about objects with too many omap entries 
Signed-off-by: Brad Hubbard <bhubbard@redhat.com>
 2017-10-24 17:27:57 +10:00
Brad Hubbard
63ef97b3db osd: Move creation of 'master_set' to scrub_compare_maps 
Moving it to scrub_compare_maps allows us to pass it to multiple
subsequently called functions.

Signed-off-by: Brad Hubbard <bhubbard@redhat.com>
 2017-10-24 17:27:51 +10:00
Kefu Chai
fdc0183871 mon/OSDMonitor: more options in new-style 
Signed-off-by: Kefu Chai <kchai@redhat.com>
 2017-10-24 15:14:09 +08:00
Kefu Chai
ca1d9566d0 common/options: set max_background_jobs instead of max_background_compactions 
Signed-off-by: Kefu Chai <kchai@redhat.com>
 2017-10-24 15:14:09 +08:00
Ziye Yang
e5f3af028a bluestore/NVMEDevice: convert the legacy config opt 
Change all the legacy config opt related with SPDK in NVMEDevice

Signed-off-by: Ziye Yang <optimistyzy@gmail.com>
 2017-10-24 14:56:10 +08:00
Haomai Wang
9f5b36879c Merge pull request #18409 from liu-chunmei/fix_dpdk_assert 
Fix two dpdk assert happened in dpdk library.

Reviewed-by: Haomai Wang <haomai@xsky.com>
 2017-10-24 01:06:53 -05:00
Luo Kexue
351874587e doc: complete and update the subsystem logging level info table 
Signed-off-by: Luo Kexue <luo.kexue@zte.com.cn>
 2017-10-24 14:04:59 +08:00
chenliuzhong
36326dc710 ceph.spec.in,debian/rules: change aio-max-nr to 1048576 
when osd is more than 14 in one host,it report error that aio is not enough.
As the default aio-max-nr is 65536, one OSD needs 4096 aios and other programs may use aios.
This patch change aio-max-nr to 1048576 when install ceph-osd rpm package and debian package

Signed-off-by: chenliuzhong <liuzhong.chen@easystack.cn>
 2017-10-24 13:16:36 +08:00
Brad Hubbard
2e3af697a8 unittest_crypto: Don't exceed limit for getentropy 
CryptoRandom::get_bytes calls getentropy directly if available and it
enforces a limit of 256 bytes.

Signed-off-by: Brad Hubbard <bhubbard@redhat.com>
 2017-10-24 14:34:26 +10:00
Kefu Chai
bc139f02d2 Merge pull request #18460 from liewegas/wip-straw2-command 
mon/OSDMonitor: add 'osd crush set-all-straw-buckets-to-straw2'

Reviewed-by: Kefu Chai <kchai@redhat.com>
 2017-10-24 12:27:15 +08:00
xie xingguo
f82228c4af osd/osd_type.cc: dump extents map object_info_t 
which is good for bug hunting and diagnosing.

Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>
 2017-10-24 11:46:23 +08:00
xie xingguo
9da3f6bac1 osd/PrimaryLogPG: fix incorrect extents update on write_full 
Two problems:
(1) write_full may cause object size to shrink, but
    current code logical won't be able to handle this properly.
(2) update stats in a simpler and consistent way.

Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>
 2017-10-24 11:46:23 +08:00
Sage Weil
fed1938bbb Merge pull request #18428 from liewegas/wip-21878 
os/bluestore/BlueFS: fix race with log flush during async log compaction

Reviewed-by: Varada Kari <varada.kari@gmail.com>
 2017-10-23 22:30:05 -05:00
Sage Weil
a91064af10 Merge pull request #18470 from yuyuyu101/wip-tracker-lock 
osd/OpRequest: reduce overhead when disabling tracking

Reviewed-by: Kefu Chai <kchai@redhat.com>
 2017-10-23 22:29:41 -05:00
Sage Weil
557088e07d Merge pull request #18472 from yuyuyu101/wip-kill-asynccompressor 
compressor: kill AsyncCompressor which is broken

Reviewed-by: Kefu Chai <kchai@redhat.com>
 2017-10-23 22:29:19 -05:00
David Zafman
2b792c5077 osd: Handle objects on non-actingbackfill nodes during recovery 
With this fix we can see more misplaced and no degraded when
doing recovery to a new up/acting set with node(s) containing
objects that are not longer part of up/acting.

Signed-off-by: David Zafman <dzafman@redhat.com>
 2017-10-23 20:20:44 -07:00
Sage Weil
d0732fc96f qa/cephfs: test ec data pool 
Signed-off-by: Sage Weil <sage@redhat.com>
 2017-10-23 21:11:24 -05:00
Ziye Yang
15113037ba bluestore/NVMEDevice: Minor fixes. 
1 After SPDK initialization, release DPDK thread core affinity binding
2 When there are errors for polling, throw error. And if there
are no completed ones, sleep 5 us.

Signed-off-by: Ziye Yang <optimistyzy@gmail.com>
 2017-10-24 09:51:49 +08:00
xie xingguo
46bfe362ed osd/osd_type: kill log.approx_size() 
We used approx_size() because std::list::size() used to be O(n) instead of O(1),
but it is constant now in C++11.

Also counting the precise number of current log entries here is much more
meaningful.

Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>
 2017-10-24 08:37:17 +08:00
xie xingguo
94c7e97c6d osd/osd_type.h: kill pg_missing_set.have_old 
which has no consumers, and removing the unused code should
always be the preferred option.

Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>
 2017-10-24 08:37:17 +08:00
xie xingguo
348b319ad1 os/PG: choose replicated acting by last_update 
Candidate which has a newer last_update should take precedence,
since it can be technically brought back to normal sooner than others.

Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>
 2017-10-24 08:37:17 +08:00
David Zafman
3818e8ed2c osd: Adjust misplaced calculation 
When there are more backfill ndoes then misplaced nodes the
object calculation would cause misplace to drop too quickly.
Assuming that all backfill nodes are at approximate the same
place, we compute backfill per node and only adjust by
the number of misplaced nodes.

Fixes: http://tracker.ceph.com/issues/21898

Signed-off-by: David Zafman <dzafman@redhat.com>
 2017-10-23 17:09:24 -07:00
Patrick Donnelly
a9baa8b33a
Merge PR #18255 into master 
* refs/pull/18255/head:
	os/bluestore: handle compressed extents in blob unsharing checks

Reviewed-by: xie xingguo <xie.xingguo@zte.com.cn>
 2017-10-23 17:07:37 -07:00