we should chop off the consumed chunk from indata, otherwise the op
consuming data from indata will always read from the beginning of
indata.
Signed-off-by: Kefu Chai <kchai@redhat.com>
mgr/orchestrator: Update docstring of service_action method
Reviewed-by: Sebastian Wagner <sebastian.wagner@suse.com>
Reviewed-by: Stephan Müller <smueller@suse.com>
* refs/pull/36501/head:
qa: add tests for mds_min_caps_working_set
mds: add working set minimum for caps
qa: use config_set/config_get
qa: do not append file names to dirname
qa: add exception for test timeouts
Reviewed-by: Douglas Fuller <dfuller@redhat.com>
A client may hold many inodes pinned in its cache for open files. That
client may be unable to release those caps to respond to cache pressure
from the MDS (or quiescent client cap recall). We should not complain if
that number of capabilities is reasonable (< 10k by default).
Fixes: https://tracker.ceph.com/issues/46830
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
Otherwise the files generated are not actually under the sub-directory!
This is correcting a confusing aspect of the test infrastructure but
doesn't actually require any changes to the tests.
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
This is needed as newer versions of the framework have
the `seastar::compat` compatibility layer stripped.
Signed-off-by: Radoslaw Zarzynski <rzarzyns@redhat.com>
This is needed as newer versions of the framework have
the `seastar::apply` compatibility layer stripped.
Signed-off-by: Radoslaw Zarzynski <rzarzyns@redhat.com>
To make this easier to catch. It is still a RuntimeError so it should
not affect current tests by default.
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
we should redirect stderr for crimson instead for default flavor. this
change addresses a regression introduced by
da76f46461
Signed-off-by: Kefu Chai <kchai@redhat.com>
Clang complains:
/home/jenkins/workspace/ceph-master-compile/src/librbd/cache/ParentCacheObjectDispatch.cc:97:66: error: 'object_off' in capture list does not name a variable
([this, read_data, dispatch_result, on_dispatched, object_no, object_off,
^
/home/jenkins/workspace/ceph-master-compile/src/librbd/cache/ParentCacheObjectDispatch.cc:98:6: error: 'object_len' in capture list does not name a variable
object_len, snap_id, &parent_trace](ObjectCacheRequest* ack) {
^
/home/jenkins/workspace/ceph-master-compile/src/librbd/cache/ParentCacheObjectDispatch.cc:99:41: error: reference to local binding 'object_off' declared in enclosing function 'librbd::cache::ParentCacheObjectDispatch::read'
handle_read_cache(ack, object_no, object_off, object_len, snap_id,
^
/home/jenkins/workspace/ceph-master-compile/src/librbd/cache/ParentCacheObjectDispatch.cc:81:9: note: 'object_off' declared here
auto [object_off, object_len] = extents.front();
^
/home/jenkins/workspace/ceph-master-compile/src/librbd/cache/ParentCacheObjectDispatch.cc:99:53: error: reference to local binding 'object_len' declared in enclosing function 'librbd::cache::ParentCacheObjectDispatch::read'
handle_read_cache(ack, object_no, object_off, object_len, snap_id,
^
/home/jenkins/workspace/ceph-master-compile/src/librbd/cache/ParentCacheObjectDispatch.cc:81:21: note: 'object_len' declared here
auto [object_off, object_len] = extents.front();
^
/home/jenkins/workspace/ceph-master-compile/src/librbd/cache/ParentCacheObjectDispatch.cc:99:41: error: reference to local binding 'object_off' declared in enclosing function 'librbd::cache::ParentCacheObjectDispatch<librbd::ImageCtx>::read'
handle_read_cache(ack, object_no, object_off, object_len, snap_id,
^
/home/jenkins/workspace/ceph-master-compile/src/librbd/cache/ParentCacheObjectDispatch.cc:242:31: note: in instantiation of member function 'librbd::cache::ParentCacheObjectDispatch<librbd::ImageCtx>::read' requested here
template class librbd::cache::ParentCacheObjectDispatch<librbd::ImageCtx>;
^
/home/jenkins/workspace/ceph-master-compile/src/librbd/cache/ParentCacheObjectDispatch.cc:81:9: note: 'object_off' declared here
auto [object_off, object_len] = extents.front();
^
/home/jenkins/workspace/ceph-master-compile/src/librbd/cache/ParentCacheObjectDispatch.cc:99:53: error: reference to local binding 'object_len' declared in enclosing function 'librbd::cache::ParentCacheObjectDispatch<librbd::ImageCtx>::read'
handle_read_cache(ack, object_no, object_off, object_len, snap_id,
^
/home/jenkins/workspace/ceph-master-compile/src/librbd/cache/ParentCacheObjectDispatch.cc:81:21: note: 'object_len' declared here
auto [object_off, object_len] = extents.front();
^
6 errors generated.
fixes: https://github.com/ceph/ceph/pull/36145
Signed-off-by: Willem Jan Withagen <wjw@digiware.nl>
When 'set-user', 'set-pub-key' or 'set-priv-key' are called two
times with the same value, we can ignore the second call.
Signed-off-by: Ricardo Marques <rimarques@suse.com>
Make sure we don't mix daemon names, config entities and
auth entities. Cause all of them don't mix well together.
Signed-off-by: Sebastian Wagner <sebastian.wagner@suse.com>
This will be very helpful when release ceph package(like RPM) with
backporting rados_getaddrs() api to previous version.
In tcmu-runner we need to check whether the rados_getaddrs() is
support or not.
Fixes: https://tracker.ceph.com/issues/46842
Signed-off-by: Xiubo Li <xiubli@redhat.com>
With the RWRef supporting, before _unmount() could continue all the
the sync write(the "readers") must have finished.
Signed-off-by: Xiubo Li <xiubli@redhat.com>
This will help us to break the big client_lock. Such as if we want
to add one separate lock for each inode, it will work like:
inode_lock.lock();
...
client_lock.lock();
client_lock.unlock();
...
inode_lock.unlock();
So we must get rid of the client_lock for all the "readers", which
must hold the client_lock from the beginning until the end.
Fixes: https://tracker.ceph.com/issues/46649
Signed-off-by: Xiubo Li <xiubli@redhat.com>
This is a common read/write reference framework, which will work
like a rw lock, for the readers/writers it won't hold the lock but
will increase the reference instead.
With this we can get rid of big locks in some use case, such as for
the Client.cc, in which such as for the file read()/write() it must
hold the client_lock at the begining until they finish, this will
make sure the _unmount() won't release the resources the read()/write()
are still using. With this it maks breaking the big client_lock to
be possible.
The usage: such as in Client.cc
Readers:
For the read()/writer(), etc, they will be work as "readers", in
the beginning they just need to check the state whether it satisfies
what the "readers" need to be, if not it will return directly, or
will increase the reference and continue.
Writers:
And for the _unmount(), as the "writer" in the beginning it will
just update the state to next stage first and when wait for all
the "readers" to finish. After the "writer" updating the state then
all the new comming "readers" will fail and return directly.
Fixes: https://tracker.ceph.com/issues/46649
Signed-off-by: Xiubo Li <xiubli@redhat.com>
This commit deduplicates the `OpsExexuter::do_osd_op()` by moving
`std::optional`-typed `osd_op_params` instantiation to `do_write_op()`.
It also fixes an issue where `clean_regions` of `osd_op_params` were
reset on every write, writefull or truncate.
Signed-off-by: Radoslaw Zarzynski <rzarzyns@redhat.com>
This simplifies the logic of `OpsExecuter::flush_changes()` and
enables for moving the instantiation code to `do_write_op()` in
the future.
The performance impact seems negligible as the affected ops are
not the ones dominating in our expected workload (read/write, omap).
Signed-off-by: Radoslaw Zarzynski <rzarzyns@redhat.com>
The benefit is no need to instantiate `osd_op_params` when only
non-modifying operation had been handled.
Signed-off-by: Radoslaw Zarzynski <rzarzyns@redhat.com>
