The current implementations are correct and will not link the
throttleing future with unrelated futures.
Signed-off-by: Yingxin Cheng <yingxin.cheng@intel.com>
Return a boolean to notify whether the Dispatcher has claimed the
message. And fix all the ms_dispatch() to return future only for
throttling purposes, which is currently always seastar::now().
Signed-off-by: Yingxin Cheng <yingxin.cheng@intel.com>
rgw: temporarily disable calls to defer_gc() in RGWGetObj
Reviewed-by: Matt Benjamin <mbenjamin@redhat.com>
Reviewed-by: J. Eric Ivancich <ivancich@redhat.com>
Currently the configure script is run on 'liburing_ext' not only for the
initial build but for subsequent incremental builds.
Signed-off-by: Brad Hubbard <bhubbard@redhat.com>
Commit 9536625558 ("qa/tasks/ceph: use Cluster.sh() and Remote.sh()
when appropriate") dropped run.wait(), which waits for all given
processes to exit. This resulted in errors like
INFO:teuthology.orchestra.run.smithi107.stderr:tar: ./objectstore_tool..log: file changed as we read it
INFO:teuthology.orchestra.run.smithi107.stderr:tar: ./ceph-client.admin.175125.log: File removed before we read it
as the task moved on to archiving semi-corrupted and uncompressed logs,
filling up the lab cluster.
Revert that hunk, as Cluster.sh() is useless here -- we don't need
stdout or stderr, but very much need parallel execution and wait for
the compression to finish.
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
cls_rgw_gc_queue_update_entry() is known to cause data loss when called
on objects that have not actually been scheduled for garbage collection
RGWGetObj is the only caller, and uses defer_gc() when reads are taking
a long time compared to rgw_gc_obj_min_wait. if an object has since been
deleted and submitted for garbage collection, this allows RGWGetObj to
defer that gc until the entire read completes
by disabling these calls to defer_gc(), very long reads (longer than 1hr,
with default configuration) may fail if the object gets deleted, and a
retry will result in a 404 Not Found error as expected
Fixes: https://tracker.ceph.com/issues/47866
Signed-off-by: Casey Bodley <cbodley@redhat.com>
Some of the new crypto tests use a generic test name that
doesn't include the full namespace of the class under test.
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
This commit introduces internal (not yet part of the api) librbd functions for:
1. formating an RBD image in LUKS format
2. parsing an RBD image in LUKS format
The actual implementation of the LUKS format is done via libcryptsetup, which is added as a new dependency.
Signed-off-by: Or Ozeri <oro@il.ibm.com>
With CEPHX_V2 authorizer challenges brought back in commit
4a82c72e3b, these need to be bumped again, as two authorizers
(without and then with the challenge) are transmitted and signed
instead of one (without the challenge). See commit 94953dd939
("msg/async/ProtocolV2: allow rxbuf/txbuf get bigger in testing")
for details.
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
At the moment, "rbd-wnbd" says that "help" is an unsupported command,
only "rbd-wnbd -h" works. This isn't very user friendly.
Although "rbd-wnbd" isn't supposed to be used directly in most cases,
users might want to retrieve device specific arguments that can be
passed using "rbd device -o".
Signed-off-by: Lucian Petrut <lpetrut@cloudbasesolutions.com>
