* refs/pull/22999/head:
mds: consider max age of dispatch queue when finding stale client
mds: reset heartbeat map at potential time-consuming places
mds: change MDSRank::finished_queue to deque
Reviewed-by: Patrick Donnelly <pdonnell@redhat.com>
* refs/pull/23370/head:
client: check for unmounted condition before printing debug output
Reviewed-by: Patrick Donnelly <pdonnell@redhat.com>
Reviewed-by: Zheng Yan <zyan@redhat.com>
The CLI `status` method tells you what backend
is selected, and also gives the backend
an opportunity to complain if something
seems wrong with its configuration.
Signed-off-by: John Spray <john.spray@redhat.com>
This is pretty sparse because users don't really
interact with the module, they interact with other
modules that use it.
Signed-off-by: John Spray <john.spray@redhat.com>
For those with running k8s/Rook environments, that want
to inject newly built binaries from their development
tree into their kubernetes environment.
Signed-off-by: John Spray <john.spray@redhat.com>
This is going to be pretty useful for development: calls
into the same places that the dashboard would call into.
Signed-off-by: John Spray <john.spray@redhat.com>
build/ops: install-deps.sh: refrain from installing/using lsb_release, and other cleanup
Reviewed-by: Ken Dreyer <kdreyer@redhat.com>
Reviewed-by: Kefu Chai <kchai@redhat.com>
* refs/pull/22433/head:
common/config: Add description to (near)full ratio settings
Reviewed-by: David Zafman <dzafman@redhat.com>
Reviewed-by: Sage Weil <sage@redhat.com>
If we have a long line that has to be dynamically allocated, we need to
adjust m_log_buf_pos accordingly afterward. In particular,
_write_and_copy does this adjustment, which means the later += line_used
is not only unnecessary but very bad, since line_used is > MAX_LOG_BUF.
Simplify all of this by getting rid of _write_and_copy: if we are writing
a big chunk let's just send the whole thing to m_fd and not bother keeping
the tail of it around for later.
Fix _log_safe_write to write the passed buffer, not m_log_buf.
Replace the paranoid check with a simple assert.
Fixes: 65da5ba216
Fixes: https://github.com/ceph/ceph/pull/23422
Signed-off-by: Sage Weil <sage@redhat.com>
Now we don't need to add a breadcrumb in each new page we create,
just add the necessary extra data in the route module and the
breadcrumb will be automatically created.
I used a modified version of ngx-breadcrumbs from McNull:
https://github.com/McNull/ngx-breadcrumbs
Fixes: http://tracker.ceph.com/issues/24781
Signed-off-by: Tiago Melo <tmelo@suse.com>
Avoid catch-all exception handlers, especially
ones that just `pass`
When catching RADOS object not found type errors,
catch specifically those, and make sure we're still
logging any actual unexpected IO errors from RADOS.
Signed-off-by: John Spray <john.spray@redhat.com>
We don't usually do gratuitious whitespace commits,
but for the python code it's worthwhile as it allows
smart editors to show us real problems without them
getting lost in the style noise.
Signed-off-by: John Spray <john.spray@redhat.com>
* rename Findrocksdb.cmake to FindRocksDB.cmake to match its name
* add RocksDB::RocksDB target to BuildRocksDB.cmake and
FindRocksDB.cmake
* use RocksDB::RocksDB target instead of accessing its property
directly, and do not link against its dependencies explicitly.
let its INTERFACE_LINK_LIBRARIES do the job.
Signed-off-by: Kefu Chai <kchai@redhat.com>
as the higher version of libstdc++ is backward compatible with the lower
ones. so there is no need to statically link against C++ libraries. they
can always use the libstdc++ ships with the distro.
This reverts commit a6c73b6ac1d81cc9f467aa38e5afb146d5dea6c2
Signed-off-by: Kefu Chai <kchai@redhat.com>
