test/rgw/test_rgw_iam_policy: do not increase ref when creating intrusive_ptr<CephContext>

before this change, we increment the refcount when constructing
`cct` instrusive_ptr, but nobody owns this smart pointer. also,
`CephContext` 's constructor set its refcount to 1. so, when the
test finishes, the refcount is 1, and this leads to a leakage of
the `CephContext` instance. and LeakSanitizer points this out:

```
Indirect leak of 10880000 byte(s) in 1 object(s) allocated from:
    #0 0xaaaac359c7c8 in operator new(unsigned long) (/home/jenkins-build/build/workspace/ceph-pull-requests-arm64/build/bin/unittest_rgw_iam_policy+0x211c7c8) (BuildId: 060fadb10da261b52fd5757c7b1e9812d34542f1)
    #1 0xffff96f764e4 in __gnu_cxx::new_allocator<ceph::logging::ConcreteEntry>::allocate(unsigned long, void const*) /usr/bin/../lib/gcc/aarch64-linux-gnu/11/../../../../include/c++/11/ext/new_allocator.h:127:27
    #2 0xffff96f757cc in std::allocator<ceph::logging::ConcreteEntry>::allocate(unsigned long) /usr/bin/../lib/gcc/aarch64-linux-gnu/11/../../../../include/c++/11/bits/allocator.h:185:32
    #3 0xffff96f757cc in boost::circular_buffer<ceph::logging::ConcreteEntry, std::allocator<ceph::logging::ConcreteEntry> >::allocate(unsigned long) /home/jenkins-build/build/workspace/ceph-pull-requests-arm64/build/boost/include/boost/circular_buffer/base.hpp:2396:39
    #4 0xffff96f75500 in boost::circular_buffer<ceph::logging::ConcreteEntry, std::allocator<ceph::logging::ConcreteEntry> >::initialize_buffer(unsigned long) /home/jenkins-build/build/workspace/ceph-pull-requests-arm64/build/boost/include/boost/circular_buffer/base.hpp:2494:18
    #5 0xffff96f6ec4c in boost::circular_buffer<ceph::logging::ConcreteEntry, std::allocator<ceph::logging::ConcreteEntry> >::circular_buffer(unsigned long, std::allocator<ceph::logging::ConcreteEntry> const&) /home/jenkins-build/build/workspace/ceph-pull-requests-arm64/build/boost/include/boost/circular_buffer/base.hpp:1039:9
    #6 0xffff96f63528 in ceph::logging::Log::Log(ceph::logging::SubsystemMap const*) /home/jenkins-build/build/workspace/ceph-pull-requests-arm64/src/log/Log.cc:53:5
    #7 0xffff96045300 in ceph::common::CephContext::CephContext(unsigned int, ceph::common::CephContext::create_options const&) /home/jenkins-build/build/workspace/ceph-pull-requests-arm64/src/common/ceph_context.cc:729:16
    #8 0xffff960446ec in ceph::common::CephContext::CephContext(unsigned int, code_environment_t, int) /home/jenkins-build/build/workspace/ceph-pull-requests-arm64/src/common/ceph_context.cc:697:5
    #9 0xaaaac3629238 in IPPolicyTest::IPPolicyTest() /home/jenkins-build/build/workspace/ceph-pull-requests-arm64/src/test/rgw/test_rgw_iam_policy.cc:864:15
    #10 0xaaaac3628da0 in IPPolicyTest_MaskedIPOperations_Test::IPPolicyTest_MaskedIPOperations_Test() /home/jenkins-build/build/workspace/ceph-pull-requests-arm64/src/test/rgw/test_rgw_iam_policy.cc:869:1
    #11 0xaaaac3628d3c in testing::internal::TestFactoryImpl<IPPolicyTest_MaskedIPOperations_Test>::CreateTest() /home/jenkins-build/build/workspace/ceph-pull-requests-arm64/src/googletest/googletest/include/gtest/internal/gtest-internal.h:472:44
```

so, in this change, we do not increase the refcount when creating cct.

Signed-off-by: Kefu Chai <tchaikov@gmail.com>

Merge pull request #44264 from tchaikov/wip-mipel

cmake: do not use GCC extension when detecting 16-byte atomic op

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

Merge pull request #45835 from tchaikov/wip-55189

common/ceph_json: dump bool using f->dump_bool()

Reviewed-by: Casey Bodley <cbodley@redhat.com>
Reviewed-by: Anthony D'Atri <anthony.datri@gmail.com>

Merge pull request #46355 from tchaikov/wip-no-annobin

ceph.spec.in: disable annobin plugin if compile with gcc-toolset

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

common/ceph_json: dump bool using f->dump_bool()

as per https://www.json.org/json-en.html, JSON encodes bool as
"true" or "false", without the quotes. before this change, the quotes
are always added when encoding boolean values.

but this change is not backward compatible.

encode_json()'s bool overload is used by rgw. it uses JSONObj
defined in common/ceph_json.h to decode JSON-encoded structs.
and it does not differentiate bool from str when decoding a boolean
value despite that it could have check the "quoted" member variable
of JSONObj for validating the type of value. so we should be fine.

Fixes: https://tracker.ceph.com/issues/55189
Signed-off-by: Kefu Chai <tchaikov@gmail.com>

Merge pull request #46293 from ronen-fr/wip-rf-sched-test

test/osd: unit-tests for the scrubber scheduler

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

Merge pull request #45467 from phlogistonjohn/jjm-format-exp2

A module and decorator for generically handling format= in python mgr modules

Reviewed-by: Adam King <adking@redhat.com>
Reviewed-by: Ernesto Puerta <epuertat@redhat.com>
Reviewed-by: Redouane Kachach <rkachach@redhat.com>

Merge pull request #45786 from adk3798/staggered-upgrade

mgr/cephadm: staggered upgrade

Reviewed-by: Anthony D'Atri <anthonyeleven@users.noreply.github.com>
Reviewed-by: Redouane Kachach <rkachach@redhat.com>

ceph.spec.in: disable annobin plugin if compile with gcc-toolset

we are seeing FTBFS failures when compiling the packages with seastar
enabled, like:

Package gcc-toolset-9-annobin-9.08-4.el8.x86_64 is already installed.
...
Package redhat-rpm-config-130-1.el8.noarch is already installed.
...
CFLAGS=-O2 -g -pipe -Wall -Werror=format-security  -Wp,-D_GLIBCXX_ASSERTIONS -fexceptions -fstack-protector-strong -grecord-gcc-switches -specs=/usr/lib/rpm/redhat/redhat-hardened-cc1 -specs=/usr/lib/rpm/redhat/redhat-annobin-cc1 -m64 -mtune=generic -fasynchronous-unwind-tables -fstack-clash-protection -fcf-protection
...
-- The CXX compiler identification is GNU 9.2.1
....
    Building CXX object CMakeFiles/cmTC_386fb.dir/testCXXCompiler.cxx.o
    /opt/rh/gcc-toolset-9/root/usr/bin/c++   -O2 -g -pipe -Wall -Werror=format-security  -Wp,-D_GLIBCXX_ASSERTIONS -fexceptions -fstack-protector-strong -grecord-gcc-switches -specs=/usr/lib/rpm/redhat/redhat-hardened-cc1 -specs=/usr/lib/rpm/redhat/redhat-annobin-cc1 -m64 -mtune=generic -fasynchronous-unwind-tables -fstack-clash-protection -fcf-protection  -o CMakeFiles/cmTC_386fb.dir/testCXXCompiler.cxx.o -c /home/jenkins-build/build/workspace/ceph-dev-new-build/ARCH/x86_64/AVAILABLE_ARCH/x86_64/AVAILABLE_DIST/centos8/DIST/centos8/MACHINE_SIZE/gigantic/release/17.2.0-209-g3e01cd41/rpm/el8/BUILD/ceph-17.2.0-209-g3e01cd41/x86_64-redhat-linux-gnu/CMakeFiles/CMakeTmp/testCXXCompiler.cxx
    cc1plus: fatal error: inaccessible plugin file /opt/rh/gcc-toolset-9/root/usr/lib/gcc/x86_64-redhat-linux/9/plugin/gcc-annobin.so expanded from short plugin name gcc-annobin: No such file or directory

the root cause is that redhat-rpm-config expects gcc-annobin:

$ cat /usr/lib/rpm/redhat/redhat-annobin-cc1 *cc1_options: + %{!-fno-use-annobin:%{!iplugindir*:%:find-plugindir()} -fplugin=gcc-annobin}

while this plugin is not installed by gcc-toolset-9-annobin, which installs annobin.so though:

$ rpm -ql gcc-toolset-9-annobin| grep annobin.so /opt/rh/gcc-toolset-9/root/usr/lib/gcc/x86_64-redhat-linux/9/plugin/annobin.so /opt/rh/gcc-toolset-9/root/usr/lib/gcc/x86_64-redhat-linux/9/plugin/annobin.so.0 /opt/rh/gcc-toolset-9/root/usr/lib/gcc/x86_64-redhat-linux/9/plugin/annobin.so.0.0.0

but gcc-toolset-8-annobin provides this file. upgrading to
gcc-toolset-11 does not help. see https://centos.pkgs.org/8-stream/centos-appstream-x86_64/gcc-toolset-11-annobin-plugin-gcc-10.23-1.el8.x86_64.rpm.html
so, the intermediate solution would be to disable the plugin, if
we want to use gcc-toolset to build rpm packages.

in this change, _annotated_build is undefined to prevent the compiler
from adding extra information to the binary. in general this change
shuold be safe, without these information, it'd be hard to tell if
the binary is hardened or what ABI version it expects. see
also https://fedoraproject.org/wiki/Changes/Annobin

Signed-off-by: Kefu Chai <tchaikov@gmail.com>

Merge pull request #46335 from cyx1231st/wip-seastore-onode-fix-unaligned-access

crimson/onode-staged-tree: fix unaligned reference to shard_pool_t::pool

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

Merge pull request #46333 from athanatos/sjust/wip-55705

crimson: generalize journal_seq_t to tolerate different paddr_t types

Reviewed-by: Yingxin Cheng <yingxin.cheng@intel.com>
Reviewed-by: Myoungwon Oh <myoungwon.oh@samsung.com>

Merge pull request #46310 from cyx1231st/wip-seastore-tune-cleaner

crimson/os/seastore: increase journal size and decrease rewrite-dirty size

Reviewed-by: Samuel Just <sjust@redhat.com>
Reviewed-by: Xuehan Xu <xxhdx1985126@gmail.com>

Merge pull request #46122 from tchaikov/wip-pmem

blk/pmem: refactor pmem_check_file_type() using std::filesystem

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

Merge pull request #45562 from tchaikov/wip-vstart-with-pmem

vstart.sh: add options to enable PMEM bluestore backend

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

doc/cephadm: staggered upgrade docs

Signed-off-by: Adam King <adking@redhat.com>

mgr/cephadm: unit test for staggered upgrade param validation

Signed-off-by: Adam King <adking@redhat.com>

qa/suites/orch/cephadm: staggered upgrade test

Signed-off-by: Adam King <adking@redhat.com>

mgr/cephadm: autopep8 changes

Signed-off-by: Adam King <adking@redhat.com>

mgr/cephadm: make use of new upgrade control parameters

Fixes: https://tracker.ceph.com/issues/54135
Signed-off-by: Adam King <adking@redhat.com>

mgr/cephadm: make UpgradeState from_json a bit safer

This way, for downgrades to whatever versions
this lands in onward, having added new parameters to
UpgradeState shouldn't break anything. Can't do much
about downgrades to older versions from this one
but this should help in the future.

Signed-off-by: Adam King <adking@redhat.com>

mgr/cephadm: add new args and validation for staggered upgrade

Signed-off-by: Adam King <adking@redhat.com>

mgr/cephadm: split _do_upgrade into sub functions

This function was around 500 lines and difficult to work
with. Splitting it into sub functions should hopefully make
it a bit easier to understand and make changes to.

Signed-off-by: Adam King <adking@redhat.com>

Merge pull request #45425 from rkachach/fix_issue_54309

mgr/cephadm: Adding prometheus service discovery endpoints

Reviewed-by: Adam King <adking@redhat.com>
Reviewed-by: Paul Cuzner <pcuzner@redhat.com>

Merge pull request #46301 from rkachach/fix_issue_55685

mgr/cephadm: adding support to copy ceph conf to per fsid config location

Reviewed-by: Adam King <adking@redhat.com>

Merge pull request #46054 from rkachach/fix_issue_55466

mgr/cephadm: fix unit.stop for ceph daemons

Reviewed-by: Adam King <adking@redhat.com>

Merge pull request #46129 from rkachach/fix_issue_ssh

mgr/cephadm: Adding an early ssh connectivity check during bootsrap

Reviewed-by: Adam King <adking@redhat.com>

Merge pull request #45951 from rhcs-dashboard/feature-set-login-banner

mgr/dashboard: customizable log-in page text/banner

Reviewed-by: Aashish Sharma <aasharma@redhat.com>
Reviewed-by: Avan Thakkar <athakkar@redhat.com>
Reviewed-by: Ernesto Puerta <epuertat@redhat.com>
Reviewed-by: Nizamudeen A <nia@redhat.com>

Merge pull request #46067 from sumedhak27/fix/user-policy/response-correction

rgw_rest_user_policy: Fix GetUserPolicy & ListUserPolicies responses

Reviewed-by: Daniel Gryniewicz <dang@redhat.com>
Reviewed-by: Pritha Srivastava <prsrivas@redhat.com>

Merge pull request #46008 from soumyakoduri/wip-skoduri-cloudtier-tests

rgw/qa: enable s3-tests related to cloud-transition feature

Reviewed-by: casey Bodley <cbodley@redhat.com>
Reviewed-by: Maredia, Ali <amaredia@redhat.com>

crimson/os/seastore: increase journal size and decrease rewrite-dirty size

Signed-off-by: Yingxin Cheng <yingxin.cheng@intel.com>

Merge pull request #46334 from athanatos/sjust/wip-55706

crimson: clarify use of tm_make_config_t

Reviewed-by: Myoungwon Oh <myoungwon.oh@samsung.com>
Reviewed-by: Yingxin Cheng <yingxin.cheng@intel.com>

rgw/qa: Run tests on multiple cloudtier config

Run cloudtier tests with parameter 'retain_head_object'
set to true and false.

However having multiple cloudtier storage classes in the same task
is increasing the transition time and resulting in spurious failures.
Hence until there is a consistent way of running the tests, without
having to depend on lc_debug_interval, disabled one of the config for
now.

Signed-off-by: Soumya Koduri <skoduri@redhat.com>

crimson: generalize journal_seq_t to tolerate different paddr_t types

Introduced: bdc1c1490ff11269c7808e762893fcbdea3a0503
Fixes: https://tracker.ceph.com/issues/55706
Signed-off-by: Samuel Just <sjust@redhat.com>

crimson/onode-staged-tree: fix unaligned reference to shard_pool_t::pool

../src/crimson/os/seastore/onode_manager/staged-fltree/stages/key_layout.h:844:44:
runtime error: reference binding to misaligned address 0x6250013ee905
for type 'const crimson::os::seastore::onode::pool_t' (aka 'const
  long'), which requires 8 byte alignment

from UndefinedBehaviorSanitizer

Signed-off-by: Yingxin Cheng <yingxin.cheng@intel.com>

crimson: clarify use of tm_make_config_t

Fixes: https://tracker.ceph.com/issues/55706
Introduced: 9fe59429fc2ef17e2bec329109713eb63ffc74c9
Signed-off-by: Samuel Just <sjust@redhat.com>

test/osd: unit-tests for the scrubber scheduler

Adding unittest_scrub_sched, testing the way scrubs are scheduled.

Signed-off-by: Ronen Friedman <rfriedma@redhat.com>

Merge pull request #46329 from ceph/1629

doc: 16.2.9 Release notes

Merge pull request #44233 from myoungwon/wip-add-rbm-inline-outline-alloc

seastore: add CircularBoundedJournal

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

seastore/cbjournal: add a log if crc is mismatched

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

seastore/cbjournal: add assert(initialized), revmoving open_device_read_header()

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

seastore/cbjournal: rename init to initialized

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

seastore/cbjournal: move written_to from the header to CBJournal

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

seastore/cbjournal: remove header_checksum field and add calculating header checkksum

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

seastore: use config value and avoid passing a hint

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

seastore: rename journal_type to journal_type_t

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

test/cbjournal: do rb_device.reset() unconditionally

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

test/seastore: use tm_make_config_t to get default_placement_hint

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

test/seastore: return journal type with less lines

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

seastore: add a log line after set_disable_trim is called

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

seastore: add the comment to explain disable_trim

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

seastore: add space between ) and {

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

seastore: add a comment (blk_paddr_t is absolute)

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

seastore: do assert if paddr type is not SEGMENT in add_relative

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

seastore: rollback a change (paddr_t -> auto)

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

seastore/cbjournal: fix comment to explain the condition in detail

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

seastore/cbjournal: make open_device_read_header not to be called if init is true

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

seastore/cbjournal: rename open_for_write to open_device_read_header

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

seastore/cbjournal: add comment to explain initialized

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

seastore/cbjournal: remove unused field

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

seastore/cbjournal: add writev to submit writes with aligned buffers

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

seastore/cbjournal: use journal_tail instead of applied_to to track journal tail

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

seastore/cbjournal: remove start and end fields in cbjournal header

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

seastore/cbjournal: remove error field

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

seastore/cbjournal: change comment for block_size

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

test/cbjournal: fix wrong device id

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

seastore/cbjournal: remove used_size in header

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

seastore/cbjournal: rename csum to header_checksum and remove csum_type

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

seastore/cbjournal: use cur_segment_seq to track the sequence of written records

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

seastore: use config_t instead of parameter in make_transaction_manager

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

seastore: make journal_seq_t.cmp compatible with blk_paddr_t

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

seastore: add disable_trim parameter to segment_cleaner for cbjournal testing

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

seastore: add add_record_relative to blk_paddr to handle adjust_addrs_from_base correctly

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

seastore: fix maybe_relarive_to in paddr_t to handle the both addrs transparently

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

seastore: add maybe_relative_to in blk_paddr_t

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

seastore: replace set_*_off with add_offset to handle transparently

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

seastore: fix add_relative to work corretly in case of blk addr

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

seastore/cbjournal: remove last_committed_record_base

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

seastore: add more explanation regarding how CBJournal works

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

seastore: recover header's variable correctly without storing it

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

seastore: read records until find invalid entry and recover variables

To do not loss last_committed_record and written_to at shutdown,
this commit update the most recent value correctly during replay.

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

seastore: rename *_super to *_header and use it as helper functions

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

seastore: move value initialization to the default constructor

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

seastore: use const in mkfs()

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

seastore: enable cbjournal when running unittest_transaction_manager

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

test: add unit tests for CircularBoundedJournal

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

seatore: add CircularBoundedJournal

Signed-off-by: Myoungwon Oh <myoungwon.oh@samsung.com>

Merge pull request #46151 from cbodley/wip-rgw-derr-config-store

rgw: remove cerr config_store from rgw_common.cc

Reviewed-by: Daniel Gryniewicz <dang@redhat.com>

Merge pull request #46148 from smanjara/wip-smanjara-metadata-sync-fix

rgw multisite: complete fix for metadata sync issue

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

Merge pull request #46141 from sumedhak27/rgw_rest_bucket/support_max-size_attr

rgw_rest_bucket: Support max-size attr for SetBucketQuota REST API

Reviewed-by: Daniel Gryniewicz <dang@redhat.com>

Merge pull request #46232 from cbodley/wip-qa-rgw-sse

qa/rgw: use 'with-sse-s3' override for s3tests

Reviewed-by: Ali Maredia <amaredia@redhat.com>

Merge pull request #46152 from cbodley/wip-55548

radosgw-admin: don't crash on --placement-id without --storage-class

Reviewed-by: Daniel Gryniewicz <dang@redhat.com>

doc: 16.2.9 Release notes

Signed-off-by: David Galloway <dgallowa@redhat.com>

Merge pull request #46056 from ceph/wip-yuriw-release-16.2.8-master

doc: 16.2.8 Release Notes

Merge pull request #46167 from dang/wip-dang-nosuchtagset

RGW - Fix NoSuchTagSet error

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

Merge pull request #44013 from adk3798/use-old-addr

mgr/cephadm: re-use old ip when re-adding hosts if necessary

Reviewed-by: Redouane Kachach <rkachach@redhat.com>

Merge pull request #46088 from pcuzner/cephadm_mpath_support

cephadm: Add support for mpath devices to gather-facts

Reviewed-by: Adam King <adking@redhat.com>

Merge pull request #46315 from zdover23/wip-doc-2022-05-18-intro-3-to-three

doc/start: s/3/three/ in intro.rst

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

Merge pull request #46306 from rkachach/fix_issue_55638

mgr/cephadm: stripping out / from the end of the url

Reviewed-by: Adam King <adking@redhat.com>

Merge pull request #46291 from adk3798/move-active-mgr

mgr/cephadm: force fail over when we want to remove active mgr

Reviewed-by: Redouane Kachach <rkachach@redhat.com>

Merge pull request #46303 from rkachach/fix_issue_55641

mgr/cephadm: do not use sudo for root user

Reviewed-by: Adam King <adking@redhat.com>