* refs/pull/22725/head:
qa/workunits/suites/blogbench.sh: use correct dir name
Reviewed-by: Patrick Donnelly <pdonnell@redhat.com>
Reviewed-by: Josh Durgin <jdurgin@redhat.com>
We had been ignoring the failure to load ceph.conf for all subcommands
but most of them require that a ceph.conf be present. This changes that
so only commands that do not need ceph.conf ignore the failure to find
it.
Fixes: http://tracker.ceph.com/issues/23941
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
* drop the experimental warning of multi-MDS
* update snapshotting blocked off version
* update Multiple filesystems availability in Jewel
Signed-off-by: Jos Collin <jcollin@redhat.com>
otherwise we will have:
[ 83%] Building CXX object src/tools/rbd/CMakeFiles/rbd.dir/Utils.cc.o
../../../lib/librbd_test.a(test_ImageWatcher.cc.o): In function
`boost:🧵:do_try_join_until(timespec const&)':
/home/kchai/ceph/build/boost/include/boost/thread/detail/thread.hpp:789:
undefined reference to
`boost:🧵:do_try_join_until_noexcept(timespec const&, bool&)'
collect2: error: ld returned 1 exit status
src/test/librbd/CMakeFiles/ceph_test_librbd.dir/build.make:137: recipe
for target 'bin/ceph_test_librbd' failed
make[2]: *** [bin/ceph_test_librbd] Error 1
Signed-off-by: Kefu Chai <kchai@redhat.com>
before this change, we guard all read/write access to md_config_t with a
lock. after this change, this policy is optional. and is controled by a
enum named `LockPolicy`. we will use `md_config_impl<LockPolicy::SINGLE>`
to implement the lockless config used by crimson/osd.
Signed-off-by: Kefu Chai <kchai@redhat.com>
Compile with -DWITH_FIO=ON, met the following bug.
/home/ceph/src/test/fio/fio_ceph_objectstore.cc:148:9: error: ‘unique_ptr’ in namespace ‘ceph’ does not name a template type
ceph::unique_ptr<std::mutex> lock;
^~~~~~~~~~
/home/ceph/src/test/fio/fio_ceph_objectstore.cc: In constructor ‘{anonymous}::Collection::Collection(const spg_t&, ObjectStore::CollectionHandle)’:
/home/ceph/src/test/fio/fio_ceph_objectstore.cc:158:9: error: class ‘{anonymous}::Collection’ does not have any field named ‘lock’
lock(new std::mutex) {
^~~~
/home/ceph/src/test/fio/fio_ceph_objectstore.cc: In function ‘fio_q_status {anonymous}::fio_ceph_os_queue(thread_data*, io_u*)’:
/home/ceph/src/test/fio/fio_ceph_objectstore.cc:604:38: error: ‘struct {anonymous}::Collection’ has no member named ‘lock’
std::lock_guard<std::mutex> l(*coll.lock);
This bug introduct by commit da5d156b6b.
Signed-off-by: Jianpeng Ma <jianpeng.ma@intel.com>
* refs/pull/22294/head:
common/performance: Now hobject_t's hash no longer depends directly on oid.
common/performance: improved hash calculation for hobject_t and ghobject_t
Reviewed-by: Paul Emmerich <paul.emmerich@croit.io>
Reviewed-by: Josh Durgin <jdurgin@redhat.com>
add a reference-counted SharedMutexImpl so that lock guards can outlive
the SharedMutex itself. this is required because the lock guards are
passed with async completions, and there is no guarantee that the
executor will process those completions before the SharedMutex
destructs. this case is exercised by the async_destruct unit test
Fixes: http://tracker.ceph.com/issues/24124
Signed-off-by: Casey Bodley <cbodley@redhat.com>
This utilizes the recent feature in teuthology [1] to skip hidden files in
suites when building the job matrix.
Idea of this change is to enable referring to the top-level qa directory in a
position-independent way such that copies of a suite to another location do not
break any symlinks.
[1] https://github.com/ceph/teuthology/pull/1185
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
