* refs/pull/40903/head:
test: add test case for MDS privated inos accessing
mds: make the lost+found dir accessible from clients
mds: move the inos 1,2 and 3 macros to ceph_fs.h
Reviewed-by: Jeff Layton <jlayton@redhat.com>
Reviewed-by: Patrick Donnelly <pdonnell@redhat.com>
* refs/pull/41120/head:
doc/_ext: ignore desc if it is unavailable
doc/_ext: check "default" for None
doc/_ext: print 0B if option value is 0
doc/cephfs: render options using confval directive
Reviewed-by: Patrick Donnelly <pdonnell@redhat.com>
write_log_entries() will split a log entry at the end of the log, the
remainder is written to the beginning at DATA_RING_BUFFER_OFFSET. On
the read side aio_read_data_block() doesn't handle this case and just
crashes. Unless the workload in use is <= 4K, the image is rendered
unusable sooner or later.
Fixes: https://tracker.ceph.com/issues/50589
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
there is chance that we don't have desc, desc_long or fmt_desc, in that
case, we should just skip desc before checking its length. so, just use
'if desc' which is able to check for None or empty string.
Signed-off-by: Kefu Chai <kchai@redhat.com>
default could be None and we don't reference the default value using
`opt.default`, so this variable is always defined. to avoid printing
a None default value, check for a None value. see
https://jinja.palletsprojects.com/en/2.11.x/templates/#none
Signed-off-by: Kefu Chai <kchai@redhat.com>
before this change, the rendered Default is "0Ei" if the default size is 0,
would be more readable if it is "0B" or "0".
Signed-off-by: Kefu Chai <kchai@redhat.com>
this change addresses the warning of:
/home/jenkins-build/build/workspace/ceph-pr-docs/doc/cephfs/mds-config-ref.rst:2: WARNING: duplicate confval_option description of mds_cache_memory_limit, other instance in cephfs/cache-configuration
/home/jenkins-build/build/workspace/ceph-pr-docs/doc/cephfs/mds-config-ref.rst:2: WARNING: duplicate confval_option description of mds_cache_reservation, other instance in cephfs/cache-configuration
Signed-off-by: Kefu Chai <kchai@redhat.com>
* refs/pull/41091/head:
doc/cephfs/nfs: add section about ganesha logs
doc/cephfs/nfs: Replace volume/nfs with nfs
doc/cephfs/nfs: add note about export management with volume/nfs interface only
Reviewed-by: Patrick Donnelly <pdonnell@redhat.com>
Recently, nfs related code was moved out of volumes plugin[1]. So using the
name volume/nfs for the interface is not appropriate.
[1] https://github.com/ceph/ceph/pull/40526
Signed-off-by: Varsha Rao <varao@redhat.com>
There already is a test to verify the mempool sharding works, in the sense that
it uses at least half of the variables available to count the number of
allocated objects and their total size. This new test verifies that, with
sharding, object counting is at least twice faster than without sharding. It
also collects cacheline contention data with the perf c2c tool. The manual
analysis of this data shows the optimization gain is indeed related to cacheline
contention.
Fixes: https://tracker.ceph.com/issues/49896
Signed-off-by: Loïc Dachary <loic@dachary.org>
Bump "Waiting for allocation" to 5.
"Retiring" is at 20 for rwl and 1 for ssd. Bump the latter to 20 as
well.
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Follow rwl mode and use AbstractWriteLog::m_bytes_allocated_cap
instead of m_log_pool_ring_buffer_size specific to ssd. This fixes
"bytes available" calculation in STATS output.
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Currently if ssd cache is filled to capacity, all future I/O hangs
indefinitely because even though the cache eventually becomes clean
and retires enough entries to get back under RETIRE_HIGH_WATER, this
isn't communicated to AbstractWriteLog::check_allocation().
Fixes: https://tracker.ceph.com/issues/50560
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
mgr/dashboard: Fixed name clash when hostname similar to another
Reviewed-by: Aashish Sharma <aasharma@redhat.com>
Reviewed-by: Avan Thakkar <athakkar@redhat.com>
Reviewed-by: p-se <NOT@FOUND>
os/bluestore: Revert" Optimizing the lock of bluestore writing process"
Reviewed-by: Mark Nelson <mnelson@redhat.com>
Reviewed-by: Adam Kupczyk <akupczyk@redhat.com>
