Commit Graph

71277 Commits

Author SHA1 Message Date
Sage Weil
b2a8207582 Merge pull request #14644 from liewegas/wip-rados-df-json
rados: out json 'df' values as numbers, not strings

Reviewed-by: David Zafman <dzafman@redhat.com>
2017-04-21 08:37:46 -05:00
Sage Weil
cc3c43bc33 Merge pull request #14645 from liewegas/wip-mgr-mon
mon/MgrMonitor: only propose if we updated

Reviewed-by: Kefu Chai <kchai@redhat.com>
2017-04-21 08:37:31 -05:00
Sage Weil
4c1d59c549 Merge pull request #14648 from gregsfortytwo/wip-17916-master
osd: pglog: with config, don't assert in the presence of stale diverg…

Reviewed-by: Sage Weil <sage@redhat.com>
2017-04-21 08:37:18 -05:00
John Spray
0227ab7570 mon: don't call propose_pending in dispatch
This was happening indirectly because the command handling
code was calling tick() as a shortcut to "promote something"
and tick calls propose_pending because.

Fixes: http://tracker.ceph.com/issues/19738
Signed-off-by: John Spray <john.spray@redhat.com>
2017-04-21 06:46:57 -04:00
Willem Jan Withagen
c4616aaf6f msg/async/EventKqueue.cc: set logging from Kqueue higher
Signed-off-by: Willem Jan Withagen <wjw@digiware.nl>
2017-04-21 11:38:06 +02:00
Kefu Chai
cef433eccb mgr: move handle_pyerror() from MgrPyModules to MgrPyModule
so we can reuse it in a more sane way.

Signed-off-by: Kefu Chai <kchai@redhat.com>
2017-04-21 14:39:44 +08:00
Dan Mick
8c249e4b2b mgr/MgrPyModule::handle_command: order stringstreams to match intent
ds is intended for command output (like stdout), to be shown/saved
when there's no error returned; ss is for error status (like stderr)
to further explain the error.  mgr Python tracebacks should go to
the latter.

Signed-off-by: Dan Mick <dan.mick@redhat.com>
2017-04-21 14:17:09 +08:00
Dan Mick
6fc3a46295 mgr/PyModules handle_pyerror: don't add extra newlines to tb
Signed-off-by: Dan Mick <dan.mick@redhat.com>
2017-04-21 14:17:09 +08:00
Dan Mick
71a9856233 mgr/MgrPyModule.cc: get Python traceback in all CallMethod spots
Signed-off-by: Dan Mick <dan.mick@redhat.com>
2017-04-21 14:17:09 +08:00
Tim Serong
3d3ee4bf76 mgr: log module name before PyErr_Print()
If an exception is raised in any of the python module functions
invoked by mgr (for example, ImportError if rest_framework isn't
available when serve() in called in the rest module), print the
module name along with the failed method name.

This commit also fixes what looks to me like a missed call to
PyGILState_Release() in the failure case in MgrPyModule::serve().

Signed-off-by: Tim Serong <tserong@suse.com>
2017-04-21 14:17:09 +08:00
Kefu Chai
f50d9f799d mgr: redirect python stdout,stderr to ceph log
PyErr_Print() uses a file object "stderr" for printing messages.

Signed-off-by: Kefu Chai <kchai@redhat.com>
2017-04-21 14:17:09 +08:00
Jos Collin
e320a4699e test: warning: comparison between signed and unsigned integer expressions
The following warning appears during make:
[ 61%] Building CXX object src/test/CMakeFiles/unittest_simple_spin.dir/simple_spin.cc.o
In file included from ./src/test/simple_spin.cc:1:0:
./src/googletest/googletest/include/gtest/gtest.h: In instantiation of ‘testing::AssertionResult testing::internal::CmpHelperEQ(const char*, const char*, const T1&, const T2&) [with T1 = unsigned int; T2 = int]’:
./src/googletest/googletest/include/gtest/gtest.h:1421:23:   required from ‘static testing::AssertionResult testing::internal::EqHelper<lhs_is_null_literal>::Compare(const char*, const char*, const T1&, const T2&) [with T1 = unsigned int; T2 = int; bool lhs_is_null_literal = false]’
./src/test/simple_spin.cc:49:3:   required from here
./src/googletest/googletest/include/gtest/gtest.h:1392:11: warning: comparison between signed and unsigned integer expressions [-Wsign-compare]
   if (lhs == rhs) {
       ~~~~^~~~~~

Signed-off-by: Jos Collin <jcollin@redhat.com>
2017-04-21 11:37:46 +05:30
Jason Dillaman
528790781b rbd-mirror: remove image name retrieval from pool watcher
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
2017-04-20 22:38:19 -04:00
Jason Dillaman
5363ad5c74 rbd-mirror: eliminate use of image name during bootstrap
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
2017-04-20 22:38:19 -04:00
Jason Dillaman
19dd5a82bb rbd-mirror: supply image id of new images to bootstrap state machine
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
2017-04-20 22:38:18 -04:00
Jason Dillaman
eae3169539 librbd: specify image id for create and clone operations
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
2017-04-20 22:38:18 -04:00
Jason Dillaman
ec1f3bbaa2 rbd-mirror: removed unused pool name from cluster watcher
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
2017-04-20 22:38:18 -04:00
Jason Dillaman
9795f125d6 rbd-mirror: renamed Replayer to PoolReplayer
This is a stepping stone to support multiple peers within a single
pool.

Signed-off-by: Jason Dillaman <dillaman@redhat.com>
2017-04-20 22:38:18 -04:00
Jason Dillaman
dee4f31cab Merge pull request #13978 from trociny/wip-18787
rbd-mirror A/A: proxy InstanceReplayer APIs via InstanceWatcher RPC

Reviewed-by: Jason Dillaman <dillaman@redhat.com>
2017-04-20 21:52:56 -04:00
David Zafman
daeb4d072c Merge pull request #14647 from dzafman/wip-19698
Fix full testing in cephtool/test.sh when used by rados suite
Replace using sleep with new wait_for_health() bash function

Reviewed-by: Loic Dachary <ldachary@redhat.com>
2017-04-20 18:18:16 -07:00
David Zafman
7d579ff12b cephtool.yaml: Add "failsafe" osd error messages to whitelist
Signed-off-by: David Zafman <dzafman@redhat.com>
2017-04-20 15:47:12 -07:00
Willem Jan Withagen
59238fb1db qa/workunits/ceph-helpers.sh: Improve test_mon_pg() by starting healthy
Signed-off-by: Willem Jan Withagen <wjw@digiware.nl>
2017-04-20 15:47:01 -07:00
Willem Jan Withagen
e07f9ccb13 qa/workunits/ceph-helpers.sh: introduce (and use) wait_for_health
Signed-off-by: Willem Jan Withagen <wjw@digiware.nl>
Signed-off-by: David Zafman <dzafman@redhat.com>
2017-04-20 15:39:34 -07:00
David Zafman
1810a17e23 mon: Fix incorrect health output and test code to match
Signed-off-by: David Zafman <dzafman@redhat.com>
2017-04-20 15:39:34 -07:00
David Zafman
86fea9d539 /workunits/cephtool/test.sh: Improve testing of health-output.
Instead of 9ff401e65b use regex to handle:
       HEALTH_ERR 2 pgs stuck unclean; Full ratio(s) out of order

Signed-off-by: David Zafman <dzafman@redhat.com>
2017-04-20 15:39:33 -07:00
David Zafman
1474d0f233 test: Fix test_admin_heap_profiler() and enable it
Signed-off-by: David Zafman <dzafman@redhat.com>
2017-04-20 15:39:33 -07:00
David Zafman
84f2182881 cephtool/test.sh: Get socket for whichever environment this test is running
Need sudo for qa suite runs

Fixes: http://tracker.ceph.com/issues/19698

Signed-off-by: David Zafman <dzafman@redhat.com>
2017-04-20 15:39:09 -07:00
Sage Weil
b3cbe4af22 Merge pull request #14689 from liewegas/wip-jewel-x
qa/suites/upgrade/jewel-x: add mgr.x role
2017-04-20 17:27:29 -05:00
Sage Weil
630e778951 rados: out json 'df' values as numbers, not strings
Fixes: http://tracker.ceph.com/issues/15546
Signed-off-by: Sage Weil <sage@redhat.com
2017-04-20 16:54:33 -04:00
Sage Weil
0486c63354 Merge pull request #14423 from tchaikov/wip-cmake-crc32-cleanup
cmake: should not compile crc32c_ppc.c on intel arch.

Reviewed-by: Casey Bodley <cbodley@redhat.com>
2017-04-20 15:13:40 -05:00
Sage Weil
adada74873 Merge pull request #14434 from ifed01/wip-bluestore-params
os/bluestore: make bluestore_max_blob_size parameter hdd/ssd case dependant

Reviewed-by: Sage Weil <sage@redhat.com>
2017-04-20 15:13:18 -05:00
Sage Weil
9ddd888382 Merge pull request #14448 from majianpeng/bluestore-kv-queued-lat
os/bluestore: Record l_bluestore_state_kv_queued_lat for sync_submit_…

Reviewed-by: Sage Weil <sage@redhat.com>
2017-04-20 15:12:58 -05:00
Sage Weil
d710e06a1b Merge pull request #14549 from tchaikov/wip-kill-clang-warnings
misc: kill clang warnings

Reviewed-by: Haomai Wang <haomai@xsky.com>
2017-04-20 15:12:36 -05:00
Sage Weil
a0a1e66474 Merge pull request #14557 from liewegas/wip-bluefs-balance-interval
os/bluestore: do not balance bluefs on every kv_sync_thread iteration

Reviewed-by: Igor Fedotov <ifedotov@mirantis.com>
2017-04-20 15:12:09 -05:00
Mohamad Gebai
885c88e0b1 init-ceph: add ceph libraries path to environment
These libraries are set in vstart.sh, but not in init-ceph. When
init-ceph is not invoked through vstart.sh, library paths are missing.

Signed-off-by: Mohamad Gebai <mgebai@suse.com>
2017-04-20 15:52:08 -04:00
Casey Bodley
a4fc5c38e5 qa/rgw: don't scan radosgw logs for encryption keys on jewel upgrade test
Signed-off-by: Casey Bodley <cbodley@redhat.com>
2017-04-20 14:49:04 -04:00
Sage Weil
b9dfb3caac qa/suites/upgrade/jewel-x/point-to-point: no xenial
Avoid conflict with upstream repo having a newer version
than what we're installing.

Signed-off-by: Sage Weil <sage@redhat.com>
2017-04-20 14:18:59 -04:00
Sage Weil
0ff16f3c57 upgrade/jewel-x/stress-split: not pg import/export tests during upgrade
These aren't meant to be run on a mixed-version cluster.

Signed-off-by: Sage Weil <sage@redhat.com>
2017-04-20 14:17:19 -04:00
Matt Benjamin
dcbc28fc46 Merge pull request #11796 from linuxbox2/wip-bucket-size-ck
rgw: add bucket size limit check to radosgw-admin
2017-04-20 13:42:03 -04:00
Sage Weil
ab9e8419ec Merge pull request #14690 from ceph/luminous
build/ops: add 12.0.1 release tag in master
2017-04-20 11:55:00 -05:00
Sage Weil
1207f8945d qa/suites/upgrade/jewel-x: add mgr.x role
Signed-off-by: Sage Weil <sage@redhat.com>
2017-04-20 12:32:59 -04:00
Piotr Dałek
1645baf65d ReplicatedPG: use new fiemap interface
Reduce memory allocator load during sparse reads and recovery by not
encoding sparse file extents from map into bufferlist, then back again
into map.

Signed-off-by: Piotr Dałek <piotr.dalek@corp.ovh.com>
2017-04-20 17:24:55 +02:00
Matt Benjamin
7bc144ce36 rgw: add bucket size limit check to radosgw-admin
The change adds a new list of all buckets x all users, with
fields for bucket name, tenant name, current num_objects,
current num_shards, current objects per shard, and the
corresponding fill_status--the latter consisting of 'OK',
'WARN <n>%', or 'OVER <n>%.'

The warning check is relative to two new tunables.  The threshold
max objects per shard is set as rgw_bucket_safe_max_objects_per_shard,
which defaults to 100K.  The value rgw_bucket_warning_threshold is
a percent of the current safe max at which to warn (defaults to
90% of full).

From review:

* fix indentation (rgw_admin)
* if user a user_id is provided, check only buckets for that user
* update shard warn pct to be pct-of-fill (not 100 - pct-of-fill)
* print only buckets near or over per-shard limit, if --warnings-only
* s/bucket limitcheck/bucket limit check */
* sanity shard limit should be 90, not 10 (because that changed)
* fixes for memleaks and other points found by cbodley

Fixes: http://tracker.ceph.com/issues/17925

Signed-off-by: Matt Benjamin <mbenjamin@redhat.com>
2017-04-20 11:22:50 -04:00
Piotr Dałek
52cc9ad09a ObjectStore, *Store: extend fiemap interface
This extends fiemap interface in objectstore and all data stores so
it's possible to pass in the map used later by caller, instead of
encoding internal map into bufferlist, then parsing it back into the
same type of data container, wasting memory and CPU time in the process.

Signed-off-by: Piotr Dałek <piotr.dalek@corp.ovh.com>
2017-04-20 17:19:00 +02:00
Piotr Dałek
dcd1687562 common/interval_set.h: add move_into() and new ctor
move_into() will be used to move data from internal (private)
std::map into other, external std::map, without exposing internal one
to public. New ctor improves moving data from external map to internal
one.

Signed-off-by: Piotr Dałek <piotr.dalek@corp.ovh.com>
2017-04-20 17:18:54 +02:00
John Spray
f854e3769e Merge pull request #14447 from ukernel/wip-19450
mds: drop partial entry and adjust write_pos when opening PurgeQueue

Reviewed-by: John Spray <john.spray@redhat.com>
2017-04-20 15:55:43 +01:00
John Spray
f695a0e30f qa: s/REQUIRE_MGRS/MGRS_REQUIRED/ for consistency
Signed-off-by: John Spray <john.spray@redhat.com>
2017-04-20 15:00:31 +01:00
John Spray
636fc40d90 qa: additions to mgr.test_failover
Reproducers for recent fixes:
http://tracker.ceph.com/issues/19407
http://tracker.ceph.com/issues/19258

Signed-off-by: John Spray <john.spray@redhat.com>
2017-04-20 15:00:31 +01:00
John Spray
8ea98b4cbf qa: fix vstart_runner --create for mgr tests
Signed-off-by: John Spray <john.spray@redhat.com>
2017-04-20 15:00:31 +01:00
John Spray
838f66269f mgr: fix waiting for module shutdown() before destroying
Fixes: http://tracker.ceph.com/issues/19412
Signed-off-by: John Spray <john.spray@redhat.com>
2017-04-20 15:00:30 +01:00