pybind/cephfs: avoid unicode check

py3 does not have the unicode built-in. Instead, simply try to do the encoding
and catch failures appropriately. If it quacks like a duck...

(Note that cython apparently cheats and allows the unicode check but this is
simpler.)

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>

Merge PR #28638 into master

* refs/pull/28638/head:
qa: elide python version config
cephfs-shell: assert python3

Reviewed-by: Rishabh Dave <ridave@redhat.com>

Merge pull request #28498 from kamoltat/wip-mgr-progress-add-event-osd-marked-in

mgr/progress: Add recovery event when OSD marked in

Reviewed-by: Josh Durgin <jdurgin@redhat.com>

Merge pull request #28564 from ceph/store_statfs-log

log: fix store_statfs log line

Reviewed-by: Patrick Donnelly <pdonnell@redhat.com>

Merge pull request #28621 from Aran85/fix-ec-rf

osd: accident of rollforward may need to mark pglog dirty

Reviewed-by: Neha Ojha <nojha@redhat.com>
Reviewed-by: Samuel Just <sjust@redhat.com>

Merge pull request #28656 from dvanders/zstd-upgrade

zstd: upgrade to v1.4.0

Reviewed-by: Kefu Chai <kchai@redhat.com>

Merge pull request #28570 from xiexingguo/wip-objecter-direct-spgid

osdc/Objecter: _calc_target - inline spgid

Reviewed-by: Sage Weil <sage@redhat.com>

Merge pull request #28568 from xiexingguo/wip-off-by-one-max-osd

mon/OSDMonitor: fix off-by-one when updating new_last_in_change

Reviewed-by: Sage Weil <sage@redhat.com>

Merge pull request #28629 from xiexingguo/wip-osd-df-by-pool-name

mgr, osd: 'osd df' by pool

Reviewed-by: Sage Weil <sage@redhat.com>

qa: elide python version config

This test doesn't actually use the config and cephfs-shell is py3 only.

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>

cephfs-shell: assert python3

There's little reason to support this new tool with python2.

Fixes: http://tracker.ceph.com/issues/40418
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>

Merge pull request #28553 from trociny/wip-40071

librbd: mirror image enable/disable should enable/disable journaling

Reviewed-by: Jason Dillaman <dillaman@redhat.com>

Merge pull request #28627 from trociny/wip-40409

journal: properly advance read offset after skipping invalid range

Reviewed-by: Jason Dillaman <dillaman@redhat.com>

Merge pull request #28460 from tchaikov/wip-cmake-detect-libnl

cmake: rewrite Findgenl to support components argument

Reviewed-by: Jason Dillaman <dillaman@redhat.com>

zstd: compat with v1.4.0

In zstd d8e215cbee03b038fffe74aebad63b625c42f23c
ZSTD_compress_generic() is renamed to ZSTD_compressStream2().

Signed-off-by: Dan van der Ster <daniel.vanderster@cern.ch>

zstd: upgrade to v1.4.0

Signed-off-by: Dan van der Ster <daniel.vanderster@cern.ch>

Merge pull request #28631 from cbodley/wip-17882

qa/rgw: extra s3tests tasks use rgw endpoint configuration

Reviewed-by: Kefu Chai <kchai@redhat.com>

debian/control: remove "libnl-3-dev" from build dependencies

what we need is "libnl-genl-3-dev", libnl-3-dev should be a dependency
of "libnl-genl-3-dev". the only reason we would have added it, is if
the package maintainer of "libnl-genl-3-dev" fails to do so. and
the packages of "libnl-genl-3-dev" on both debian and ubuntu just look
correct in this perspective.

so let's remove this.

Signed-off-by: Kefu Chai <kchai@redhat.com>

cmake: rewrite Findgenl to support components argument

* rename genl to nl:
  "genl" is not very specific without more context. and libnl is actually
  a library suite. so it would be better if we can rename the library to
  be found to "libnl", and by following the naming convention of CMake,
  in this change, Findgenl.cmake is renamed to Findnl.cmake
* Findnl.cmake: support the component argument, in our specific case,
  what we want is "libnl-genl" library, which is one of the libraries
  offered by libnl. so let's just make it a component.
* Findnl.cmake: should pass the names of the required variables instead of
  their values to `find_package_handle_standard_args()`. before this
  change, we pass `GENL_LIBRARIES` to this function. it was correct at the
  first glance. but it is not able to handle the case where case where
  libnl-genl is not installed. so the fix is to pass all the names of
  required library paths to this function. in this change, their name
  are concatenated to a single variable -- `nl_LIBRARIES`, and the
  value of this variable is passed to
  `find_package_handle_standard_args()`. and the error message would
  be more specific this way if libnl-genl is not found:
    Could NOT find nl (missing: nl_genl_LIBRARY)
* Findnl.cmake: add nl::<component> as imported library, it helps the
  consumer of these libraries to import them -- no need to
  specify the `target_include_directories()` separately anymore.
* move the find_package() code to where it is used. it helps to improve
  the readability.
* remove `HAVE_GENL` variable: it's not used anywhere.
* drop the messages of "Not using foobar", they do not help.

Signed-off-by: Kefu Chai <kchai@redhat.com>

mgr/dashboard: Ensure E2E tests can be run independently (#28592)

mgr/dashboard: Ensure E2E tests can be run independently

Reviewed-by: Adam King <adking@redhat.com>
Reviewed-by: Laura Paduano <lpaduano@suse.com>
Reviewed-by: Nathan Weinberg <nweinber@redhat.com>
Reviewed-by: Rafael Quintero <rquinter@redhat.com>
Reviewed-by: Tiago Melo <tmelo@suse.com>

Merge PR #28453 into master

* refs/pull/28453/head:
qa/valgrind.supp: be slightly less specific on suppression
msg/async, v2: make the reset_recv_state() unconditional.

Reviewed-by: Sage Weil <sage@redhat.com>

Merge pull request #28566 from liewegas/wip-40377

osd: copy (don't move) pg list when sending beacon

Reviewed-by: xie xingguo <xie.xingguo@zte.com.cn>
Reviewed-by: Kefu Chai <kchai@redhat.com>

qa/rgw: update default port in perl workunits

Signed-off-by: Casey Bodley <cbodley@redhat.com>

qa/rgw: extra s3tests tasks use rgw endpoint configuration

Signed-off-by: Casey Bodley <cbodley@redhat.com>

journal: properly advance read offset after skipping invalid range

Fixes: https://tracker.ceph.com/issues/40409
Signed-off-by: Mykola Golub <mgolub@suse.com>

Merge PR #28212 into master

* refs/pull/28212/head:
mds: trim cache during standby replay

Reviewed-by: Patrick Donnelly <pdonnell@redhat.com>
Reviewed-by: Zheng Yan <zyan@redhat.com>

Merge PR #28551 into master

* refs/pull/28551/head:
mds: don't mark cap NEEDSNAPFLUSH if client has no pending capsnap

Reviewed-by: Patrick Donnelly <pdonnell@redhat.com>

Merge PR #28545 into master

* refs/pull/28545/head:
client: set snapdir's link count to 1

Reviewed-by: Patrick Donnelly <pdonnell@redhat.com>
Reviewed-by: Jeff Layton <jlayton@redhat.com>

Merge pull request #28496 from ifed01/wip-ifed-fix-40080

os/bluestore: fix duplicate allocations in bmap allocator

Reviewed-by: Sage Weil <sage@redhat.com>

Merge pull request #28509 from tchaikov/wip-mon_clean_pg_upmaps_per_chunk

common/options: make mon_clean_pg_upmaps_per_chunk unsigned

Reviewed-by: xie xingguo <xie.xingguo@zte.com.cn>
Reviewed-by: Sage Weil <sage@redhat.com>

Merge pull request #28418 from xiexingguo/wip-kick-recovery-priority

osd: give recovery ops initialized by client op a higher priority

Reviewed-by: Neha Ojha <nojha@redhat.com>
Reviewed-by: Sage Weil <sage@redhat.com>

Merge pull request #25395 from rzarzynski/wip-build-compiletest_cxx11_client

tests: introduce compiletest_cxx11_client for C++11 conformity.

Reviewed-by: Kefu Chai <kchai@redhat.com>

Merge PR #28229 into master

* refs/pull/28229/head:
os/bluestore: do garbage collection if blob count is too high.
common/perf_conters: make dump_formatted_xxx funcs as const.
os/bluestore: store extents for GC within WriteContext.
os/bluestore: GC class, make some members local.
os/bluestore: vector -> interval set in GC to track extents to collect.
tests/store_test: many-many spanning blobs test case

Reviewed-by: Sage Weil <sage@redhat.com>

Merge PR #28630 into master

* refs/pull/28630/head:
os/bluestore: fix duplicative and misleading debug in KernelDevice::open().

Reviewed-by: Igor Fedotov <ifedotov@suse.com>

mgr/dashboard: Ensure E2E tests can be run independently

Introduces the PageHelper.

- Reads class attributes from Helpers to reduce boilerplace code for
  navigation.
- The PageHelper is supposed to be the new class for code that's reused
  across all Helpers for page related code.
- The Helper class is by now meant to be used for non-page specific
  helper code, like used in the `checkConsole` method.
- The Helper class will act as central registry to enable all other
  tests to use helper functions of all derived PageHelper tests.

  Example:

  `Helper.pools.create('foobar', ...);`

Fixes: http://tracker.ceph.com/issues/40397
Signed-off-by: Patrick Nawracay <pnawracay@suse.com>

os/bluestore: fix duplicative and misleading debug in KernelDevice::open().

Signed-off-by: Radoslaw Zarzynski <rzarzyns@redhat.com>

mgr, osd: 'ceph osd df' by pool

Our test admin has been asking for this for the past few years:-)
Besides, this is also useful for operating on large Ceph clusters with
mutliple storage pools possibly spanning over all osds.

Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>

mgr/dashboard: Update translations (#28059)

mgr/dashboard: Update translations

Reviewed-by: Laura Paduano <lpaduano@suse.com>
Reviewed-by: Stephan Müller <smueller@suse.com>

qa/workunits/rbd: tweak mirror tests

(after "image mirror disable" started to disable journaling)

Signed-off-by: Mykola Golub <mgolub@suse.com>

librbd: mirror image enable/disable should enable/disable journaling

Fixes: https://tracker.ceph.com/issues/40071
Signed-off-by: Mykola Golub <mgolub@suse.com>

Merge pull request #28377 from p-na/wip-pna-message-xlf-opt

mgr/dashboard: Improve hints shown when message.xlf is invalid

Reviewed-by: Tiago Melo <tmelo@suse.com>

mgr/dashboard: Log's time from UTC to local (#27777)

mgr/dashboard: Log's time from UTC to local

Reviewed-by: Alfonso Martínez <almartin@redhat.com>
Reviewed-by: Stephan Müller <smueller@suse.com>
Reviewed-by: Tatjana Dehler <tdehler@suse.com>

osd/OSDMap: pass filter directly to the OSDUtilizationDumper

This makes the code easy to maintain and especially to extend,
e.g., adding per pool filters.

Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>

Merge pull request #28264 from rhcs-dashboard/mypy-config-to-file

src/script: extract mypy config to mypy.ini

Reviewed-by: Ernesto Puerta <epuertat@redhat.com>
Reviewed-by: Sebastian Wagner <sebastian.wagner@suse.com>

osd: accident of rollforward may need to mark pglog dirty

refers: https://github.com/ceph/ceph/pull/27015/files#r294114392

Fixes: http://tracker.ceph.com/issues/40403
Signed-off-by: Zengran Zhang <zhangzengran@sangfor.com.cn>

Merge pull request #28605 from cbodley/wip-40304

qa/rgw: fix import error in tasks/swift.py

Reviewed-by: Yuri Weinstein <yweinste@redhat.com>

qa/rgw: fix import error in tasks/swift.py

> ImportError: No module named packaging

replaces packaging.version.Version with distutils.version.LooseVersion

Signed-off-by: Casey Bodley <cbodley@redhat.com>

Merge pull request #28095 from Devp00l/wip-39702

mgr/dashboard: OSD custom action button removal

Reviewed-by: Kanika Murarka <kmurarka@redhat.com>
Reviewed-by: Patrick Nawracay <pnawracay@suse.com>
Reviewed-by: Ricardo Marques <rimarques@suse.com>
Reviewed-by: Tatjana Dehler <tdehler@suse.com>

Merge pull request #27670 from trociny/wip-rbd-mirror-fetch-autotune

 rbd-mirror: adjust journal fetch properties based on memory target

Reviewed-by: Jason Dillaman <dillaman@redhat.com>

Merge pull request #28524 from trociny/wip-qa-rbd-mirror-compare

qa/workunits/rbd: fix compare_images and compare_image_snapshots

Reviewed-by: Jason Dillaman <dillaman@redhat.com>

Merge pull request #28529 from trociny/wip-journal-in_flight_advance_sets

journal: wait for in flight advance sets on stopping recorder

Reviewed-by: Jason Dillaman <dillaman@redhat.com>

mgr/dashboard: Add QA tests for RBD configuration (#27774)

mgr/dashboard: Add QA tests for RBD configuration

Reviewed-by: Tatjana Dehler <tdehler@suse.com>

mgr/dashboard: Cleanup HACKING.rst

Rephrase explanations on how to write E2E tests with Protractor.

Signed-off-by: Patrick Nawracay <pnawracay@suse.com>

Merge pull request #28586 from aclamk/fix-vstart-extra-conf

vstart.sh: Fix problem that all extra_conf got merged into single line.

Reviewed-by: Kefu Chai <kchai@redhat.com>

vstart.sh: Fix problem that all extra_conf got merged into single line.

Signed-off-by: Adam Kupczyk <akupczyk@redhat.com>

log: fix store_statfs log line

Signed-off-by: Mohamad Gebai <mgebai@suse.com>

Merge pull request #28559 from dillaman/wip-40368

librbd: avoid dereferencing an empty container during deep-copy

Reviewed-by: Mykola Golub <mgolub@suse.com>

osdc/Objecter: _calc_target - inline spgid

A follow-up of https://github.com/ceph/ceph/pull/28487.
For erasure pools this can save us an obviously redundant call
of do_crush and other things.

Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>

Merge pull request #28487 from xiexingguo/wip-objecter-pg-mapping-cache

osdc/Objecter: pg-mapping cache

Reviewed-by: Sage Weil <sage@redhat.com>

mon/OSDMonitor: fix off-by-one when updating new_last_in_change

E.g., osdmap.osd_weight[osdmap.max_osd] does not exist.

Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>

Merge PR #28082 into master

* refs/pull/28082/head:
mgr / volumes: improve error handling
py / cephfs: invoke base class Error::__init__() from OSError::__init__()
mgr / volumes: carve out subvolume operations as a separate class
mgr / volumes: introduce subvolume specification class
mgr / volumes: carve of volume as a separate class

osd: copy (don't move) pg list when sending beacon

If we move the list, then a second beacon sent before the pg list refreshes
will be empty.

Fixes: http://tracker.ceph.com/issues/40377
Signed-off-by: Sage Weil <sage@redhat.com>

Merge PR #28198 into master

* refs/pull/28198/head:
qa/tasks: upgrade command arguments checks in vstart_runner.py

Reviewed-by: Patrick Donnelly <pdonnell@redhat.com>

librbd: avoid dereferencing an empty container during deep-copy

While updating the image copy progress, a 'while' loop will pop the
last item from the queue and then deference the empty queue.

Fixes: http://tracker.ceph.com/issues/40368
Signed-off-by: Jason Dillaman <dillaman@redhat.com>

Merge pull request #27773 from istudko/patch-1

doc/releases/nautilus: Correct a systemctl command in an upgrade guide

Reviewed-by: Kefu Chai <kchai@redhat.com>

Merge pull request #27201 from bstillwell/bootstrap-keyring-caps-doc-fix

doc: Add missing mgr cap for the bootstrap keyring

Reviewed-by: Kefu Chai <kchai@redhat.com>

Merge pull request #28526 from Osa1989/patch-1

doc/releases/nautilus.rst: fix command to check min_compat_client

Reviewed-by: Kefu Chai <kchai@redhat.com>

Merge pull request #28522 from neha-ojha/wip-add-client-endpoints

 qa/tasks/cbt.py: add support for client_endpoints

Reviewed-by: Mark Nelson <mnelson@redhat.com>

Merge pull request #28346 from smithfarm/wip-39447-alt

qa/rbd: add cram-based snap diff test

Reviewed-by: Jason Dillaman <dillaman@redhat.com>
Reviewed-by: Mykola Golub <mgolub@suse.com>

mgr/dashboard: OSD custom action button removal

Now the custom action dropdown is removed from the OSD component. It was
nearly a clone of the table actions button component, but it had a
different behavior as it did not show the main action inside the dropdown.

Fixes: https://tracker.ceph.com/issues/39702
Signed-off-by: Stephan Müller <smueller@suse.com>

Merge pull request #28532 from cbodley/wip-40304

qa/rgw: skip swift tests on rhel 7.6+

mgr/progress: Add recovery event when OSD marked in

Added the check for osd_id to account for
the acting set of both the old and new maps
to cover both osd marked out and in
respectively

Signed-off-by: Kamoltat (Junior) Sirivadhna <ksirivad@redhat.com>

doc/releases/nautilus.rst: fix command to check min_compat_client

Command should be "ceph osd dump | grep min_compat_client"

Signed-off-by: Osama Elswah <osama.ibrahim.89@gmail.com>

Merge pull request #27951 from ifed01/wip-ifed-fix-slowop-log

os/bluestore: fix origin reference in logging slow ops.

Merge pull request #28490 from xiexingguo/wip-40193

mon/OSDMonitor: report pg[pgp]_num_target instead of pg[pgp]_num

Reviewed-by: Sage Weil <sage@redhat.com>

Merge pull request #28322 from liewegas/wip-preextend

os/bluestore: bluefs_preextend_wal_files=true

Reviewed-by: Igor Fedotov <ifedotov@suse.com>

Merge pull request #28260 from s0nea/wip-dashboard-pool-delete-flag

mgr/dashboard: consider 'mon_allow_pool_delete' flag

Reviewed-by: Patrick Nawracay <pnawracay@suse.com>
Reviewed-by: Stephan Müller <smueller@suse.com>
Reviewed-by: Ricardo Marques <rimarques@suse.com>

mds: don't mark cap NEEDSNAPFLUSH if client has no pending capsnap

This avoid getting wrlocks and setup client_snap_caps on snap inode.
Because there is no further snapflush that triggers releasing these
wrlocks.

Fixes: https://tracker.ceph.com/issues/40361
Signed-off-by: "Yan, Zheng" <zyan@redhat.com>

Merge pull request #28534 from tchaikov/wip-crimson-mkkey

crimson/osd: add --mkkey support

Reviewed-by: Radoslaw Zarzynski <rzarzyns@redhat.com>

Merge pull request #28464 from majianpeng/fio-ceph-messenger-multi-client

test/fio/fio_ceph_messenger: make exec multi client on the same host.

Reviewed-by: Casey Bodley <cbodley@redhat.com>
Reviewed-by: Kefu Chai <kchai@redhat.com>

Merge pull request #28404 from xiexingguo/wip-osd-noup-stall

osd/OSD: keep synchronizing with mon if stuck at booting

Reviewed-by: Sage Weil <sage@redhat.com>

crimson/osd: ignore exception when stopping osd

OSD::stop() is a top-level method in the sense that it will be called
directly when engine stops. so it should not throw any exceptions.
instead, we will print out the error and ignore the exception if any.

Signed-off-by: Kefu Chai <kchai@redhat.com>

crimson/osd: add --mkkey support

see qa/tasks/ceph.py for how it is used, and why we need it.

Signed-off-by: Kefu Chai <kchai@redhat.com>

crimson/os: use buffer::write_file() helper

instead of using the blocking calls

so we can write to file without blocking the reactor

Signed-off-by: Kefu Chai <kchai@redhat.com>

crimson: add buffer_io helpers

Signed-off-by: Kefu Chai <kchai@redhat.com>

Merge pull request #28507 from tchaikov/wip-cmake-dpdk-boost

cmake: update SPDK to build with GCC-9

Reviewed-by: Casey Bodley <cbodley@redhat.com>

qa/rbd: add cram-based snap diff test

Fixes: http://tracker.ceph.com/issues/39447
Signed-off-by: Shyukri Shyukriev <shshyukriev@suse.com>
Signed-off-by: Nathan Cutler <ncutler@suse.com>

auth/KeyRing: remove KeyRing::create_empty()

it never fails unless out of memory. and it would be simpler to just
allocate it on stack.

Signed-off-by: Kefu Chai <kchai@redhat.com>

osdc/Objecter: pg-mapping cache

The CRUSH-based addressing is some kind of CPU intensive task
and hence should be avoid whenever possible.

The patch introduces a per objecter pg-mapping cache, which as
a result can saved us 10%+ CPU (fio, 4k randow read, 24k IOPS):

Was:
$ top -Hp 415864
top - 14:45:39 up 6 days,  5:45,  3 users,  load average: 9.67, 8.66, 8.45
Threads:  23 total,   3 running,  20 sleeping,   0 stopped,   0 zombie
%Cpu(s): 21.6 us, 14.5 sy,  0.0 ni, 59.6 id,  1.4 wa,  0.0 hi,  3.0 si,  0.0 st
KiB Mem : 19616576+total, 12773412+free, 50964508 used, 17467128 buff/cache
KiB Swap:  4194300 total,  4194300 free,        0 used. 14352892+avail Mem

    PID USER      PR  NI    VIRT    RES    SHR S %CPU %MEM     TIME+ COMMAND
 415942 root      20   0 1957800 492820 123880 R 80.0  0.3   0:13.75 tp_librbd

Now:
$ top -Hp 475779
top - 10:22:05 up  1:18,  4 users,  load average: 2.65, 1.44, 1.60
Threads:  23 total,   2 running,  21 sleeping,   0 stopped,   0 zombie
%Cpu(s):  6.1 us,  2.2 sy,  0.0 ni, 91.4 id,  0.0 wa,  0.0 hi,  0.2 si,  0.0 st
KiB Mem : 19616576+total, 13919555+free, 27280820 used, 29689392 buff/cache
KiB Swap:  4194300 total,  4194300 free,        0 used. 16798102+avail Mem

    PID USER      PR  NI    VIRT    RES    SHR S %CPU %MEM     TIME+ COMMAND
 476231 root      20   0 1957808 491712 123836 S 71.1  0.3   0:41.47 tp_librbd

And below is a more detailed report from the perf tool:

Was:
$ perf report -g graph
  Children      Self  Command    Shared Object        Symbol
+   29.33%     0.03%  tp_librbd  librados.so.2.0.0    [.] Objecter::op_submit
+   28.83%     0.09%  tp_librbd  librados.so.2.0.0    [.] Objecter::_op_submit_with_budget
+   27.12%     0.17%  tp_librbd  librados.so.2.0.0    [.] Objecter::_op_submit
+   16.47%     0.26%  tp_librbd  librados.so.2.0.0    [.] Objecter::_calc_target
+   15.04%     0.10%  tp_librbd  libceph-common.so.0  [.] OSDMap::_pg_to_up_acting_osds
+   13.52%     0.16%  tp_librbd  libceph-common.so.0  [.] OSDMap::_pg_to_raw_osds

Now:
$ perf report -g graph
  Children      Self  Command    Shared Object        Symbol
+   17.84%     0.04%  tp_librbd  librados.so.2.0.0    [.] Objecter::op_submit
+   17.34%     0.06%  tp_librbd  librados.so.2.0.0    [.] Objecter::_op_submit_with_budget
+   15.80%     0.17%  tp_librbd  librados.so.2.0.0    [.] Objecter::_op_submit
+    6.11%     2.02%  tp_librbd  librados.so.2.0.0    [.] Objecter::_calc_target

Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>

Merge pull request #28535 from tchaikov/wip-crimson-packaging

debian/: use ceph-osd for packaging crimson-osd

Reviewed-by: Radoslaw Zarzynski <rzarzyns@redhat.com>

mgr / volumes: improve error handling

This cleans up lots of return-fu statements and make the source
look much more pythonic. Functions should raise an instance of
VolumeException() class wherever necessary (error handling).

Fixes: http://tracker.ceph.com/issues/39969
Signed-off-by: Venky Shankar <vshankar@redhat.com>

py / cephfs: invoke base class Error::__init__() from OSError::__init__()

This will allow access to `args` when catching `Error`.

Signed-off-by: Venky Shankar <vshankar@redhat.com>

mgr / volumes: carve out subvolume operations as a separate class

Move (and refactor) subvolume specific operations from main module
source to SubVolume class in a new subvolume source. Also, provide
hooks in VolumeClient to forward subvolume specific opertaions to
this class.

Signed-off-by: Venky Shankar <vshankar@redhat.com>

mgr / volumes: introduce subvolume specification class

Specifications class that represents a unique subvolume
identified by (subvolume-id, group-id) tuple. Provide
heleper functions for fetching  various subvolume specific
attributes such as subvolume path, group path etc.

Signed-off-by: Venky Shankar <vshankar@redhat.com>

mgr / volumes: carve of volume as a separate class

Move volume operations (create/delete) from main module source
to VolumeClient class in new volume source.

Signed-off-by: Venky Shankar <vshankar@redhat.com>

client: set snapdir's link count to 1

Otherwise nfs-ganesha treats snapdir as stale.

Fixes: https://tracker.ceph.com/issues/40101
Signed-off-by: "Yan, Zheng" <zyan@redhat.com>

mgr/progress: Add recovery event when OSD marked in

Combine the functions _osd_in and _osd_out in
to a single function call _osd_in_out, which
has a parameter called marked (string) that
will indicate if the event is when an osd is
marked in or out. Consider only the 'acting'
set of an osd for both when 'in' and 'out'
event of the osd to calculate affected PGs

Signed-off-by: Kamoltat (Junior) Sirivadhna <ksirivad@redhat.com>

Merge PR #28456 into master

* refs/pull/28456/head:
mds: ignore sessionmap version mismatch if mds_wipe_sessions is set
mds: fix corner case of replaying open sessions

Reviewed-by: Patrick Donnelly <pdonnell@redhat.com>

Merge pull request #28521 from alimaredia/wip-qa-rgw-ssl

qa: use curl in wait_for_radosgw() in util/rgw.py

Reviewed-by: Casey Bodley <cbodley@redhat.com>

Merge pull request #28415 from linuxbox2/wip-full-odsock

rgw/OutputDataSocket: actually discard data on full buffer