Some bits of the standard Zipper conversions were missed for
the RGWLibContinuedReq case, where the setup is encapsulated in
the request, but execution is broken up in to steps. This
currently affects only RGWWriteRequest.
Fixes: https://tracker.ceph.com/issues/48136
Signed-off-by: Matt Benjamin <mbenjamin@redhat.com>
mgr/dashboard: disable cluster selection in NFS export editing form
Reviewed-by: Laura Paduano <lpaduano@suse.com>
Reviewed-by: Tatjana Dehler <tdehler@suse.com>
for two reasons:
* less typing: no need to press "shift" for inputting "_"
* more consistent with executable names like "ceph-conf"
* simpler to grep when compiling the tests. there is chance
we need to kill the dead jobs on a jenkins worker node
where it happens to be compiling the tests.
Signed-off-by: Kefu Chai <kchai@redhat.com>
* refs/pull/37654/head:
mds: print the METRIC_TYPE for debug info
mds: add XXPayload:print support
mds: only update the requesting metrics
Reviewed-by: Venky Shankar <vshankar@redhat.com>
* define `errorator_t` instead of `futurator` to simplify the code a
little bit.
* use seastar helper function / types when appropriate
Signed-off-by: Kefu Chai <kchai@redhat.com>
also add "unittest-seastar-errorator" to unit test.
this should address the build failure introduced by
688b95798b
Signed-off-by: Kefu Chai <kchai@redhat.com>
* refs/pull/37899/head:
qa: avoid using ceph.dir.entries for empty check
qa: do not test dir size equals rbytes
qa: disable check for value kclient fails to get
qa: leave one standby available to avoid warning
qa: do not check access timestamps
qa: ignore skip errors for kclient
qa: add more clients for test_volume_client
qa: accept client returning pool id for layout
qa: ignore expected warning about client missing features
qa: do not fail on skip
qa: fix mount directory
qa: rename distro link to distros
qa: accept eio for mount failure
qa: unmount volumes before removal
qa: some sessionmap tests require fuse
qa: remove obsolete test
qa: do not fail on test skip
qa: use more OSDs for workload tests
qa: use 4 clients for volumes testing
qa: relocate tasks
qa: rename workload suite
qa: relocate new tasks from fs:basic_functional
qa: remove kcephfs suite
qa: move kcephfs:mixed-clients to fs
qa: remove dead kcephfs:cephfs suite
qa: merge snap tests for kcephfs/fs
qa: add new fs:shell sub-suite
qa: remove task-less kcephfs:recovery suite
qa: add new fs:volumes sub-suite
qa: merge fs:basic_functional and kcephfs:recovery
qa: move kcephfs:thrash to fs:thrash
qa: reorganize thrash tasks
qa: move multifs/failover tests
qa: run norstats workunit for klient
qa: merge workload tests from kcephfs with fs
qa: use consistent syntax
qa: remove inline testing
qa: move log-rotate config to common yaml
qa: remove fuse default permissions config
qa: fix whitespace
Reviewed-by: Kotresh Hiremath Ravishankar <khiremat@redhat.com>
Reviewed-by: Ramana Raja <rraja@redhat.com>
rgw: using strict_strtoll for append position. If not, any invalid value will be decoded as zero which is not correct.
Reviewed-by: Yuval Lifshitz <ylifshit@redhat.com>
rgw: keep syncstopped flag when copying bucket shard headers
Reviewed-by: Casey Bodley <cbodley@redhat.com>
Reviewed-by: J. Eric Ivancich <ivancich@redhat.com>
* do not use unique_ptr<> explicitly, use `seastar::do_with()` for
better readability
* use seastar::do_until() instead of seastar::repeat() for better
readability. plain boolean is simpler than
`seastar::stop_iteration::yes`
* do not capture variables using FuturizedStoreLogReader if we could
pass them by instant parameters.
* rename "start()" to "read()". as "read" is more specific in this
context.
Signed-off-by: Kefu Chai <kchai@redhat.com>
this class is not intended to be inherited by classes, and these methods
are "final", so mark them "final".
Signed-off-by: Kefu Chai <kchai@redhat.com>
instead of returning future<int> where the int represents an errno which
is not checked by any callers, we should just return a future<>, we
could erroratorize it later on, but returning an int is not the right
way to handle this, and it is not consistent with how we handle other
errors elsewhere in crimson.
Signed-off-by: Kefu Chai <kchai@redhat.com>
We should not allow changing an export's cluster because an export ID
might live in one cluster but not in another one. Editing a non-existing
export in a cluster causes an error.
Fixes: https://tracker.ceph.com/issues/47373
Signed-off-by: Kiefer Chang <kiefer.chang@suse.com>
