RepoMirrors/ceph
1
0
Fork 0
mirror of https://github.com/ceph/ceph synced 2024-12-18 09:25:49 +00:00
Code Issues Packages Projects Releases Wiki Activity
67,157 Commits 211 Branches 394 Tags 2.1 GiB
845f432dc1
Commit Graph

67157 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Sage Weil
845f432dc1 osd/PG: do not queue ops for replay 
We no longer send ACKs, so this can't happen anyway.

Signed-off-by: Sage Weil <sage@redhat.com>
 2016-12-29 10:30:07 -05:00
Sage Weil
1addadc40a osd: we can discard "replay" ops 
Signed-off-by: Sage Weil <sage@redhat.com>
 2016-12-29 10:30:07 -05:00
Sage Weil
b9a3698e4c osd/ReplicatedPG: do not send ACK separate from COMMIT 
This never comes before COMMIT, making it useless to clients.  And it's
just unnecessary overhead.

Signed-off-by: Sage Weil <sage@redhat.com>
 2016-12-29 10:30:07 -05:00
Sage Weil
e94521638b osd: remove 'rlat' read latency metrics 
These are measuring how long of a delay we see before ops are readable.
They are calculated in the on_applied path, which we'd like to remove,
and weren't reliably calculated anyway.

Signed-off-by: Sage Weil <sage@redhat.com>
 2016-12-29 10:30:07 -05:00
Sage Weil
46cdcf8229 osd/ReplicatedPG: kill hit_set_flushing 
This was added before we started taking the ondisk lock in order to allow
reading a just-written hitset.  Since this is a rare path anyway, just
take the ondisk lock.

Signed-off-by: Sage Weil <sage@redhat.com>
 2016-12-29 10:30:07 -05:00
Sage Weil
2c05a968fe Merge pull request #12522 from hjwsm1989/ecwrite 
osd/ECBackend: Send write message to peers first, then do local write

Reviewed-by: Sage Weil <sage@redhat.com>
 2016-12-29 09:29:24 -06:00
Sage Weil
12a0f96f07 Merge pull request #12524 from yuyuyu101/wip-bluestore-flush 
os/bluestore: prevent lock for almost "flush" calls

Reviewed-by: Sage Weil <sage@redhat.com>
 2016-12-29 09:28:39 -06:00
Sage Weil
2bdb569267 Merge pull request #12528 from kylinstorage/wip-JournalingObjectStore-cleanup 
os/filestore/JournalingObjectStore cleanup

Reviewed-by: Sage Weil <sage@redhat.com>
 2016-12-29 09:27:51 -06:00
Sage Weil
13c01a0350 Merge pull request #12670 from liewegas/wip-nits 
mon: fix a few nits

Reviewed-by: Kefu Chai <kchai@redhat.com>
 2016-12-29 09:27:32 -06:00
Sage Weil
5c1c3c0845 Merge pull request #12684 from xiexingguo/xxg-wip-fix-onodesizetracking 
os/bluestore: fix OnodeSizeTracking testing

Reviewed-by: Sage Weil <sage@redhat.com>
 2016-12-29 09:26:21 -06:00
Haomai Wang
020fd24d7e Merge pull request #12707 from tchaikov/wip-dpdk-v16.07 
builds with dpdk v16.07

Reviewed-by: Haomai Wang <haomai@xsky.com>
 2016-12-29 15:02:59 +08:00
Kefu Chai
3e559b2194 msg/async: bump dpdk to v16.07 
as some distros include it already. so sync it up to avoid building
failures on them with dpdk-dev package(s) shipped by distro.

Signed-off-by: Kefu Chai <kchai@redhat.com>
 2016-12-29 14:09:20 +08:00
Kefu Chai
09b4978cfd msg/async: be compatible with dpdk v16.07 
Signed-off-by: Kefu Chai <kchai@redhat.com>
 2016-12-29 14:06:41 +08:00
Kefu Chai
27d9cf8fd9 Merge pull request #12698 from yuyuyu101/wip-dpdk-fix 
msg/async/dpdk: fix compile errors

Reviewed-by: Kefu Chai <kchai@redhat.com>
 2016-12-29 14:06:00 +08:00
Kefu Chai
c54f568a1f Merge pull request #12699 from wjwithagen/patch-15 
qa/test_rados_tool.sh: POSIX dd only accepts 'k' as multiplier

Reviewed-by: Kefu Chai <kchai@redhat.com>
 2016-12-29 12:04:23 +08:00
Sage Weil
1611090173 Merge pull request #12703 from kylinstorage/wip-remove-unused-variable 
os/bluestore/BlueStore.cc: remove unused variable

Reviewed-by: Igor Fedotov <ifedotov@mirantis.com>
 2016-12-28 11:56:55 -06:00
Sage Weil
eff4c1c5f3 Merge pull request #12594 from wjwithagen/wip-wjw-freebsd-rebind-timing 
common/config_opts.h: FreeBSD timing changed due to no SO_REUSEADDR

Reviewed-by: Sage Weil <sage@redhat.com>
 2016-12-28 11:54:53 -06:00
Li Wang
2590b97b4d os/bluestore/BlueStore.cc: remove unused variable 
Signed-off-by: Li Wang <li.wang@kylin-cloud.com>
 2016-12-28 15:23:51 +00:00
Willem Jan Withagen
4b9b562465 qa/test_rados_tool.sh: POSIX dd only accepts 'k' as multiplier 
FreeBSD does not want KB as 1k multiplier.
see also: http://www.unix.com/man-page/posix/1posix/dd/

Signed-off-by: Willem Jan Withagen <wjw@digiware.nl>
 2016-12-28 12:13:29 +01:00
Haomai Wang
8ca15b1a4e test_userspace_event: fix missing std 
Signed-off-by: Haomai Wang <haomai@xsky.com>
 2016-12-28 17:56:32 +08:00
Haomai Wang
410002de5a msg/dpdk/TCP: remove unneeded CephContext argument 
Signed-off-by: Haomai Wang <haomai@xsky.com>
 2016-12-28 17:55:45 +08:00
Haomai Wang
b486adb692 Merge pull request #12688 from Adirl/clean_line_end 
msg/async/rdma: clean line endings

Reviewed-by: Haomai Wang <haomai@xsky.com>
 2016-12-28 10:50:15 +08:00
xie xingguo
1d97862f71 os/bluestore: fix OnodeSizeTracking testing 
[  FAILED  ] 1 test, listed below:
[  FAILED  ] ObjectStore/StoreTest.OnodeSizeTracking/2, where GetParam() = "bluestore"

 1 FAILED TEST

The above test failure happens as the bluestore mount() process
will try to load all collections and put them into the coll_map,
which will be also tracked as mempool::bluestore_meta_other.
So total_bytes from mempool won't be equal to zero.

Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>
 2016-12-28 09:56:04 +08:00
Sage Weil
075325c3a4 Merge pull request #12336 from kylinstorage/wip-cleanup-use-string-ref 
osd: cleanup: use string & to avoid unnecessary copy

Reviewed-by: Sage Weil <sage@redhat.com>
Reviewed-by: Kefu Chai <kchai@redhat.com>
 2016-12-27 16:38:17 -06:00
Sage Weil
c67a23592d Merge pull request #12569 from liewegas/wip-bluestore-raw-unused 
os/bluestore/bluestore_types: drop std::bitset for blob unused

Reviewed-by: Igor Fedotov <ifedotov@mirantis.com>
Reviewed-by: Varada Kari <varada.kari@sandisk.com>
 2016-12-27 16:37:15 -06:00
Sage Weil
bb281925f7 Merge pull request #12612 from fangyuxiangGL/master 
osd/ReplicatedBackend: do not set omap header if it is empty

Reviewed-by: Kefu Chai <kchai@redhat.com>
Reviewed-by: Haomai Wang <haomai@xsky.com>
 2016-12-27 16:36:44 -06:00
Sage Weil
56b04966f6 Merge pull request #12668 from hjwsm1989/pg-unfound 
osd/PG: add new have_unfound() function in MissingLoc

Reviewed-by: Sage Weil <sage@redhat.com>
 2016-12-27 16:36:18 -06:00
Sage Weil
2e1a096a82 Merge pull request #12674 from Liuchang0812/tools-rados-offset-support 
tools/rados: add a parameter "--offset" to rados put command

Reviewed-by: Kefu Chai <kchai@redhat.com>
 2016-12-27 16:35:50 -06:00
Sage Weil
c467d7fb23 Merge pull request #12675 from Liuchang0812/fix-test-case-in-rados 
test: replace hard-code binary names with varibles

Reviewed-by: Kefu Chai <kchai@redhat.com>
 2016-12-27 16:35:29 -06:00
Sage Weil
88ae4bb067 mon/OSDMonitor: ceph osd pool get/set crush_rule by name 
This means users don't have to manually translate a rule
they just created to a ruleset ID in order to map a pool
to it.

Signed-off-by: Sage Weil <sage@redhat.com>
 2016-12-27 11:02:19 -05:00
Sage Weil
88d85ad726 mon/PGMap: remove CATEGORY column from 'df detail' 
Signed-off-by: Sage Weil <sage@redhat.com>
 2016-12-27 11:02:19 -05:00
Sage Weil
f198ba81cc mon/PGMap: fix alignment of 'cache io' in ceph -s output 
Currently it looks like

  client io 3508 B/s wr, 0 op/s rd, 5 op/s wr
  cache io 0 op/s promote

but should be

  client io 3508 B/s wr, 0 op/s rd, 5 op/s wr
   cache io 0 op/s promote

Signed-off-by: Sage Weil <sage@redhat.com>
 2016-12-27 11:02:19 -05:00
Sage Weil
52bd5bbb36 mon: 'fs set-default' not 'fs set_default' 
The convention is to use - for command names and
_ for variable names.

Signed-off-by: Sage Weil <sage@redhat.com>
 2016-12-27 11:02:19 -05:00
Sage Weil
b1811e3d55 Merge pull request #12655 from ceph/wip-fix-fio 
test/fio_ceph_objectstore: fix fio plugin build failure caused by rec…

Reviewed-by: Sage Weil <sage@redhat.com>
 2016-12-27 09:34:59 -06:00
Sage Weil
3323230d51 Merge pull request #12682 from renhwztetecs/renhw-wip-mon-osdmonitor-sortbitwise 
mon/OSDMonitor: fixup sortbitwise flag warning

Reviewed-by: Sage Weil <sage@redhat.com>
Reviewed-by: Kefu Chai <kchai@redhat.com>
 2016-12-27 09:30:59 -06:00
Sage Weil
22adda23ec Merge pull request #12654 from yuyuyu101/wip-fix-async-subop 
messages/MOSDSubOp: Make encode_payload can be reentrant

Reviewed-by: Sage Weil <sage@redhat.com>
 2016-12-27 09:26:00 -06:00
Kefu Chai
f3d453d872 Merge pull request #12662 from Liuchang0812/doc-fix 
doc: fix document about rados mon

Reviewed-by: Kefu Chai <kchai@redhat.com>
 2016-12-27 21:09:06 +08:00
Kefu Chai
d0a2651a3b Merge pull request #12198 from tchaikov/wip-pybind-freebsd 
pybind: conditional compile the linux specific constant

Reviewed-by: Willem Jan Withagen <wjw@digiware.nl>
 2016-12-27 20:52:07 +08:00
huanwen ren
c25ee187e2 mon/OSDMonitor: fixup sortbitwise flag warning 
"ceph -s" does not report warning when using
command "ceph osd unset sortbitwise" to drop
sortbitwise flag.
we should use "osdmap.get_up_osd_features() &
CEPH_FEATURE_OSD_BITWISE_HOBJ_SORT"
instead of "(osdmap.get_features(CEPH_ENTITY_TYPE_OSD, NULL) &
CEPH_FEATURE_OSD_BITWISE_HOBJ_SORT)",
because osdmap.get_features only get local "features"

Signed-off-by: huanwen ren <ren.huanwen@zte.com.cn>
 2016-12-27 11:07:54 +00:00
fangyuxiang
c0ddd399ef ReplicatedBackend: do not set omap header if it is empty 
In replicatedpg backfill, objects have no omap header or keys previously, but an empty omap header will be inserted when they are pushed to replica, which leads leveldb performance degrade seriously when cluster has huge number of objects.
so my opinion is that we does't insert omap header if is size is zero when handle PushOp in recovery.

Signed-off-by: fang yuxiang <fang.yuxiang@eisoo.com>
 2016-12-27 16:53:44 +08:00
liuchang0812
fe3d3821cb tools/rados: add parameter offset for put cmd and test case for it 
add a parameter "--offset" for put command
add a test case for rados put command

Signed-off-by: liuchang0812 <liuchang0812@gmail.com>
 2016-12-27 16:06:41 +08:00
liuchang0812
52db747f32 doc: fix document about rados mon 
Signed-off-by: liuchang0812 <liuchang0812@gmail.com>
 2016-12-27 14:59:04 +08:00
Kefu Chai
b5acc288d7 Merge pull request #12663 from badone/wip-ceph_test_rados_api_c_write_operations-valgrind-fixes 
c_write_operations.cc: Fix trivial memory leak

Reviewed-by: Kefu Chai <kchai@redhat.com>
 2016-12-27 14:44:35 +08:00
liuchang0812
14de3a1ee1 test: replace hard-code binary names with varibles 
This test case couldn't run under customize env CEPH_BIN.

Signed-off-by: liuchang0812 <liuchang0812@gmail.com>
 2016-12-27 14:18:29 +08:00
Haomai Wang
610a97abf8 Merge pull request #12672 from liupan1111/wip-fix-nvme-driver-compile-issue 
bluestore: fixed compilation error when enable spdk

Reviewed-by: Haomai Wang <haomai@xsky.com>
 2016-12-27 14:07:07 +08:00
Pan Liu
1ed6a37611 bluestore: fixed compilation error when enable spdk 
Signed-off-by: Pan Liu <pan.liu@istuary.com>
 2016-12-27 12:42:44 +08:00
Adir Lev
131062f70e msg/async/rdma: clean line endings 
Change-Id: I59359ae585b66018fe54d5e841a6506c077c606f
Signed-off-by: Adir Lev <adirl@mellanox.com>
 2016-12-27 01:53:49 +02:00
Sage Weil
11c30a86fc Merge pull request #12659 from kylinstorage/wip-remove-obsolete-configuration-option 
common/config_opts.h: remove obsolete configuration option

Reviewed-by: Sage Weil <sage@redhat.com>
 2016-12-26 12:31:47 -06:00
huangjun
7d9fd21e79 osd/PG: add new have_unfound() function in MissingLoc 
No need to lookup all objects in needs_recovery_map if only want to
          know whether if there are unfound objects.

  Signed-off-by: huangjun <hjwsm1989@gmail.com>
 2016-12-26 19:43:38 +08:00
Haomai Wang
32378addb2 Merge pull request #12648 from Adirl/fix_RoCEv2_deafult_value 
msg/async/rdma: fix RoCE v2 deafult value

Reviewed-by: Haomai Wang <haomai@xsky.com>
 2016-12-26 16:25:30 +08:00