RepoMirrors/ceph
1
0
Fork 0
mirror of https://github.com/ceph/ceph synced 2024-12-19 09:57:05 +00:00
Code Issues Packages Projects Releases Wiki Activity
67,952 Commits 211 Branches 394 Tags 2.2 GiB
64c309d7e1
Commit Graph

67952 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Piotr Dałek
64c309d7e1 OSD: allow client throttler to be adjusted on-fly, without restart 
This patch allows the osd_client_message_cap and
osd_client_message_size_cap to be adjusted on-fly, using admin socket
functionality.

Fixes: http://tracker.ceph.com/issues/18791
Signed-off-by: Piotr Dałek <piotr.dalek@corp.ovh.com>
 2017-02-02 09:32:55 +01:00
Sage Weil
67e48e20fd Merge pull request #13066 from liewegas/wip-aio-read 
os/bluestore: use aio for reads

Reviewed-by: Igor Fedotov <ifedotov@mirantis.com>
 2017-01-30 21:43:36 -06:00
Sage Weil
2cc9fac7ba Merge pull request #13071 from liewegas/wip-omap-getkeys-rados-test 
qa/suites/rados/basic: set low omap limit for rgw workload

Reviewed-by: Josh Durgin <jdurgin@redhat.com>
 2017-01-30 21:43:17 -06:00
Sage Weil
29a7965a0a Merge pull request #13150 from tchaikov/wip-mon-cleanup 
mon: post-jewel cleanups

Reviewed-by: Sage Weil <sage@redhat.com>
 2017-01-30 21:42:48 -06:00
Jason Dillaman
3d894cad24 Merge pull request #12966 from dillaman/wip-18436 
osdc: cache should ignore error bhs during trim

Reviewed-by: John Spray <john.spray@redhat.com>
Reviewed-by: Gregory Farnum <gfarnum@redhat.com>
 2017-01-30 21:24:22 -05:00
John Spray
d4f6385b85 Merge pull request #12800 from jcsp/wip-vstart-qasuite 
Improve vstart_runner to (optionally) create its own cluster

Reviewed-by: Patrick Donnelly <pdonnell@redhat.com>
 2017-01-31 02:02:49 +01:00
Jason Dillaman
f56126afe3 Merge pull request #12817 from liupan1111/wip-list-snapshot-time 
librbd: support to list snapshot timestamp

Reviewed-by: Jason Dillaman <dillaman@redhat.com>
 2017-01-30 18:36:46 -05:00
Josh Durgin
c26f317c7f Merge pull request #13098 from wido/doc-osd-greace 
doc: Document that osd_heartbeat_grace applies to MON and OSD

Reviewed-by: Joao Eduardo Luis <joao@suse.de>
Reviewed-by: Kefu Chai <kchai@redhat.com>
Reviewed-by: Josh Durgin <jdurgin@redhat.com>
 2017-01-30 14:55:45 -08:00
Ilya Dryomov
326a11362f Merge pull request #13141 from dillaman/wip-18653 
librbd: introduce new constants for tracking max block name prefix

Reviewed-by: Ilya Dryomov <idryomov@gmail.com>
 2017-01-30 22:33:25 +01:00
Jason Dillaman
2c08629c99 librbd: introduce new constants for tracking max block name prefix 
Fixes: http://tracker.ceph.com/issues/18653
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
 2017-01-30 15:25:23 -05:00
Sage Weil
084f0c1876 Merge pull request #13188 from liewegas/wip-rest-fs-name 
qa/workunits/rest: use unique pool names for cephfs test

Reviewed-by: Kefu Chai <kchai@redhat.com>
 2017-01-30 13:30:05 -06:00
Pan Liu
66992d1931 test: regolden formatted-output.t for timestamp metadata of rbd snapshot 
Signed-off-by: Pan Liu <pan.liu@istuary.com>
 2017-01-31 03:22:33 +08:00
Pan Liu
bd4e36632a test: Add unit test for getting rbd snapshot timestamp info 
Signed-off-by: Pan Liu <pan.liu@istuary.com>
 2017-01-31 03:20:31 +08:00
Pan Liu
9debf5c98a librbd: support to list snapshot time stamp 
Fixes: http://tracker.ceph.com/issues/808
Signed-off-by: Pan Liu <pan.liu@istuary.com>
 2017-01-31 03:18:32 +08:00
Pan Liu
4dd47b0bac librbd: Add timestamp metadata when creating rbd snapshot 
Signed-off-by: Pan Liu <pan.liu@istuary.com>
 2017-01-31 03:18:14 +08:00
Yehuda Sadeh
fc9a03ec4f Merge pull request #12646 from SirishaGuduru/swift-api-auth-doc-change 
doc: Fixes radosgw-admin ex: in swift auth section

Reviewed-by: Yehuda Sadeh <yehuda@redhat.com>
 2017-01-30 08:36:32 -08:00
Sage Weil
388a7ff8ef qa/workunits/rest: use unique pool names for cephfs test 
Using cephfs_[meta]data collides with the pools that teuthology
already creates if an mds is defined.

This became a (noticeable) problem with 052c3d3f68

Signed-off-by: Sage Weil <sage@redhat.com>
 2017-01-30 11:14:31 -05:00
Casey Bodley
b4446ac582 Merge pull request #13054 from oritwas/wip-rgw-period-update-crash 
rgw: fix period update crash

Reviewed-by: Casey Bodley <cbodley@redhat.com>
 2017-01-30 09:52:13 -05:00
Kefu Chai
e690a1afbf Merge pull request #13186 from Adirl/fix_rdma_compile 
cmake: fix broken RDMA compilation after merge PR #12878

Reviewed-by: Kefu Chai <kchai@redhat.com>
 2017-01-30 20:19:11 +08:00
Radoslaw Zarzynski
034a6f9ba0 Merge pull request #13003 from Jing-Scott/dev/swift-container-acl-info 
rgw: swift: the http referer acl in swift API should be shown

Reviewed-by: Radoslaw Zarzynski <rzarzynski@mirantis.com>
 2017-01-30 12:58:56 +01:00
Radoslaw Zarzynski
fca5b3c3ff Merge pull request #13005 from Jing-Scott/dev/swift-http-url-parse 
rgw: swift: The http referer should be parsed to compare in swift API

Reviewed-by: Radoslaw Zarzynski <rzarzynski@mirantis.com>
 2017-01-30 12:55:35 +01:00
Radoslaw Zarzynski
b8e7a276f6 Merge pull request #12874 from Jing-Scott/dev/swift-ref 
rgw: the swift container acl should support field .ref

Reviewed-by: Radoslaw Zarzynski <rzarzynski@mirantis.com>
 2017-01-30 12:44:21 +01:00
Oren Duer
94a27ecf39 cmake: fix broken RDMA compilation after merge PR #12878 
issue: 965984

Change-Id: I2e401acf204069d3907001d276c58d9fb4cd0d58
Signed-off-by: Oren Duer <oren@mellanox.com>
 2017-01-30 10:55:26 +00:00
Wido den Hollander
f42afa828b
doc: Document that osd_heartbeat_grace applies to MON and OSD 
This value has to be set in both the [mon] and [osd] section since
both daemons read this value.

Signed-off-by: Wido den Hollander <wido@42on.com>
 2017-01-30 10:04:52 +01:00
Haomai Wang
d09fded5ae Merge pull request #13142 from Adirl/remove_compilation_warning 
msg/async/rdma: Remove compilation warning

Reviewed-by: Haomai Wang <haomai@xsky.com>
 2017-01-28 15:52:17 +08:00
Sage Weil
496012e206 Merge pull request #13057 from wido/issue-18635 
systemd: Restart Mon after 10s in case of failure

Reviewd-by: Loic Dachary <ldachary@redhat.com>
 2017-01-27 15:38:24 -06:00
Sage Weil
b88b29ef6e Merge pull request #13140 from liewegas/wip-bluestore-statfs 
os/bluestore: fix statfs to not include DB partition in free space

Reviewed-by: Igor Fedotov <ifedotov@mirantis.com>
 2017-01-27 15:37:25 -06:00
Sage Weil
166735a0b4 Merge pull request #13135 from liewegas/wip-kerneldevice-debug 
os/bluestore/KernelDevice: fix debug message

Reviewed-by: Igor Fedotov <ifedotov@mirantis.com>
 2017-01-27 15:37:13 -06:00
John Spray
81e3a9dd9e Merge pull request #13124 from ukernel/wip-18675 
client: remove request from session->requests when handling forward

Reviewed-by: John Spray <john.spray@redhat.com>
 2017-01-27 17:41:36 +01:00
John Spray
f44795a625 Merge pull request #13095 from ukernel/wip-18660 
mds: skip fragment space check for replayed request

Reviewed-by: John Spray <john.spray@redhat.com>
 2017-01-27 17:40:46 +01:00
John Spray
b54707ddca Merge pull request #13052 from ukernel/wip-16768 
mds: don't modify inode that is not projected

Reviewed-by: Patrick Donnelly <pdonnell@redhat.com>
 2017-01-27 17:39:32 +01:00
John Spray
20db6c6f54 Merge pull request #12973 from ukernel/wip-18578 
mds: fix incorrect assertion in Server::_dir_is_nonempty()

Reviewed-by: John Spray <john.spray@redhat.com>
 2017-01-27 17:37:31 +01:00
John Spray
48adc77476 Merge pull request #12825 from stiopaa1/r11124_refuseToDoFsNew 
mon/MDSMonitor.cc:refuse fs new on pools with obj

Reviewed-by: John Spray <john.spray@redhat.com>
 2017-01-27 17:35:38 +01:00
John Spray
954a26534d Merge pull request #12489 from gregsfortytwo/wip-fix-rename-bounds 
client: fix the cross-quota rename boundary check conditions

Reviewed-by: Yan, Zheng <zyan@redhat.com>
 2017-01-27 17:33:41 +01:00
John Spray
f3c7ab7988 Merge pull request #12440 from runsisi/wip-fix-fp-exception 
include/fs_types: fix unsigned integer overflow

Reviewed-by: Sage Weil <sage@redhat.com>
 2017-01-27 17:32:49 +01:00
Sage Weil
73f5d2b73f os/bluestore: only use aio for read if there are >1 blobs 
If we have a single blob to read it is not worth the context switch.

Signed-off-by: Sage Weil <sage@redhat.com>
 2017-01-27 10:27:04 -05:00
Sage Weil
aca26825dd os/bluestore/BlockDevice: allow sync read to accumulate on bl 
Do not clear the target bl; instead, append to it.  This makes our
behavior consistent with aio_read, which does the same.

Signed-off-by: Sage Weil <sage@redhat.com>
 2017-01-27 10:27:04 -05:00
Sage Weil
9602a34525 os/bluestore: do read io via aio in parallel 
Dispatch all blob reads in parallel via aio.

Signed-off-by: Sage Weil <sage@redhat.com>
 2017-01-27 10:27:03 -05:00
Sage Weil
393747a698 os/bluestore/KernelDevice: avoid possible use-after-free for ioc->priv 
If aio_wake() triggers destruction, ioc->priv might be a
use-after-free (this is the case for the BlueStore read path).

Signed-off-by: Sage Weil <sage@redhat.com>
 2017-01-27 10:26:16 -05:00
Sage Weil
dbaa2b056e os/bluestore/NVMEDevice: implement aio_read 
Signed-off-by: Sage Weil <sage@redhat.com>
 2017-01-27 10:26:16 -05:00
Sage Weil
4aeaaabb67 os/bluestore/BlockDevice: add aio_read API 
NVMEDevice not implemented yet.

Signed-off-by: Sage Weil <sage@redhat.com>
 2017-01-27 10:26:15 -05:00
John Spray
5f67cf97f9 Merge pull request #13088 from jcsp/wip-16397-unpin 
qa: unpin knfs from ubuntu

Reviewed-by: Jeff Layton <jlayton@redhat.com>
 2017-01-27 15:39:41 +01:00
Kefu Chai
e5c256d12e mon: assert if pgmap's format_version < 1 
we bump the pgmap's format_version to 1 and the monitor's protocol
version to 12 in 2d2aa00. and it is required to upgrade to jewel before
upgrading to kraken or luminous. so we can safely assume that the
PGMonitor's format_version is always greater or equal to 1.

Signed-off-by: Kefu Chai <kchai@redhat.com>
 2017-01-27 12:17:38 +08:00
Kefu Chai
ff1fd4b4b6 mon: assert on CEPH_FEATURE_MON_STATEFUL_SUB 
this feature was introduced in jewel. and we are targeting luminous in
master. so it's safe to assume that all OSDs in the cluster have this
feature.

Signed-off-by: Kefu Chai <kchai@redhat.com>
 2017-01-27 12:17:38 +08:00
Dan Mick
24abb20695 Merge pull request #13148 from tchaikov/wip-18692 
packaging: install libceph-common.so* not libceph-common.so.*

Reviewed-by: Dan Mick <dmick@redhat.com>
 2017-01-26 19:43:34 -08:00
Kefu Chai
b641135cf8 packaging: install libceph-common.so* not libceph-common.so.* 
Fixes: http://tracker.ceph.com/issues/18692
Signed-off-by: Kefu Chai <kchai@redhat.com>
 2017-01-27 10:43:42 +08:00
Josh Durgin
fa0311ee52 Merge pull request #13087 from athanatos/wip-17831-18583 
osd/PrimaryLogPG::try_lock_for_read: give up if missing

Reviewed-by: Josh Durgin <jdurgin@redhat.com>
 2017-01-26 15:50:53 -08:00
Josh Durgin
2ebea511ed Merge pull request #12934 from athanatos/wip-18529 
test/pybind/test_rados.py: tolerate TimedOut in test_ping_monitor

Reviewed-by: Kefu Chai <kchai@redhat.com>
Reviewed-by: Josh Durgin <jdurgin@redhat.com>
 2017-01-26 15:14:45 -08:00
Mykola Golub
c720e348d4 Merge pull request #13110 from dillaman/wip-18666 
librbd: prevent self-blacklisting during break lock

Reviewed-by: Mykola Golub <mgolub@mirantis.com>
 2017-01-26 23:44:01 +02:00
Marcus Watts
5f229d6a33 radosgw/swift: clean up flush / newline behavior. 
The current code emits a newline after swift errors, but fails
to account for it when it calculates 'content-length'.  This results in
some clients (go github.com/ncw/swift) producing complaints about the
unsolicited newline such as this,
	Unsolicited response received on idle HTTP channel starting with "\n"; err=<nil>

This logic eliminates the newline on flush.  This makes the content length
calculation correct and eliminates the stray newline.

There was already existing separator logic in the rgw plain formatter
that can emit a newline at the correct point.  It had been checking
"len" to decide if previous data had been emitted, but that's reset to 0
by flush().  So, this logic adds a new per-instance variable to separately
track state that it emitted a previous item (and should emit a newline).

Fixes: http://tracker.ceph.com/issues/18473
Signed-off-by: Marcus Watts <mwatts@redhat.com>
Signed-off-by: Matt Benjamin <mbenjamin@redhat.com>
 2017-01-26 14:58:29 -05:00