The cluster has already multiple the full ratio before returning
the "max_avail".
Fixes: https://tracker.ceph.com/issues/50984
Signed-off-by: Xiubo Li <xiubli@redhat.com>
For kclient, the write() will return -ENOSPC instead of the fsync().
Fixes: https://tracker.ceph.com/issues/45434
Signed-off-by: Xiubo Li <xiubli@redhat.com>
Set the object_size to 1MB to make the objects destributed more even
among the OSDs.
Fixes: https://tracker.ceph.com/issues/45434
Signed-off-by: Xiubo Li <xiubli@redhat.com>
The test case assumes that client does not send cap messages if there
is no operation. Recent change that timeout idle open files break that
assumpation.
Fixes: https://tracker.ceph.com/issues/44382
Signed-off-by: "Yan, Zheng" <zyan@redhat.com>
A first step to do more automatic code checks on the qa/
directory. This is useful while transitioning to python3.
Also use log_exc to top-level to not run into:
error: Argument 1 to "log_exc" has incompatible type
"Callable[[OSDThrasher], Any]"; expected "OSDThrasher"
Signed-off-by: Thomas Bechtold <tbechtold@suse.com>
Rather than verify the latest OSDMap is the same as the rank's
osdmap_epoch_barrier, just use the rank's version. The OSDMap may change
out-of-band with the test startup and thus the epoch's would diverge.
The file system and rank is fresh for each test so there's no reason to
care if the MDS barrier is one epoch behind the latest.
Fixes: https://tracker.ceph.com/issues/43554
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
Previously, errors stuck indelibly to the inode, which
meant that a close call would see an error even if the
user already dutifully fsync()'d and handled it.
We should emit each error only once per file handle.
Signed-off-by: John Spray <john.spray@redhat.com>
