mgr/smb: remove embedded join source and users and groups types

Drop the embedded versions of the join source and usrers and groups
configs from the cluster resource type. This means that potentially
sensitive information is *always* stored in a separate resource
from the less-sensitive general cluster config. Future changes can
then make security related choices about those specific resources
without worrying about having embedded variants inside the cluster.

For convenience I added a new "emtpy" users and groups source type that
can be used for testing. It defines *no* users and groups (an empty set
basically). This wouldn't be useful in a production scenario today as it
would deploy an smbd that you couldn't log into. But for unit testing
it's extremely handy, and maybe it can be used for integration testing
too.

This is a large change because lots of unit tests relied on embedded
join auths or users-and-groups. I tried to handle all the breaking
changes in one go for bisect-ability.

Signed-off-by: John Mulligan <jmulligan@redhat.com>

mgr/smb: convert smb module cluster create to use linked resources

Convert the smb mgr module cli to use linked join auth and
users-and-groups resources when created on the `cluster create` command
line. These linked resources with have semi-ranodmized names and be
linked to the cluster that we're creating on the CLI.

Signed-off-by: John Mulligan <jmulligan@redhat.com>

mgr/smb: add squash method to ResultGroup

The new squash method works similarly to `one` but permits multiple
resources to be reported on by moving them into the report for the
"parent" resource. This is meant for upcoming changes to `cluster
create` when it starts using linked join auths and linked
users-and-groups.

Signed-off-by: John Mulligan <jmulligan@redhat.com>

mgr/smb: add handler tests for new linked_to_cluster field

Signed-off-by: John Mulligan <jmulligan@redhat.com>

mgr/smb: add rand_name function

This will be used later to create resources that are not directly named
by a user.

Signed-off-by: John Mulligan <jmulligan@redhat.com>

mgr/smb: update handler to make use of linked_to_cluster fields

Signed-off-by: John Mulligan <jmulligan@redhat.com>

mgr/smb: add linked_to_cluster fields for JoinAuth and UsersAndGroups

The linked_to_cluster field will be used for resources that must be
life-cycled along side the cluster using them. This is mainly intended
for use with `cluster create` command line commands.

Signed-off-by: John Mulligan <jmulligan@redhat.com>

mgr/smb: add staging area class to handler

This in-memory staging area for new resources being applied to the
config helps simplify validation.

Signed-off-by: John Mulligan <jmulligan@redhat.com>

mgr/smb: extend key fetching support in internal store helper module

Signed-off-by: John Mulligan <jmulligan@redhat.com>

mgr/smb: split ConfigStore protocol listing methods out

Split the ConfigStore protocol into a new base protocol that covers the
methods needed to list keys in a store, called ConfigStoreListing.

Signed-off-by: John Mulligan <jmulligan@redhat.com>

mgr/smb: move handler's resource check logic into functions

There's no need for these to be methods as they're largely
self-contained. Functions should be easier to unit-test independently in
the future too.

Signed-off-by: John Mulligan <jmulligan@redhat.com>

mgr/smb: make handler's resource ordering a top-level function

There's no need for order_resources to be a method. Make order_resources
a proper function.

Signed-off-by: John Mulligan <jmulligan@redhat.com>

Merge pull request #57004 from kchheda3/wip-fix-retry

rgw/multisite-notification: retry storing bucket notification attrs for ECANCELED(ConcurrentModification) errors.

Reviewed-by: Casey Bodley <cbodley@redhat.com>
Reviewed-by: Yuval Lifshitz <ylifshit@ibm.com>

Merge PR #57207 into main

* refs/pull/57207/head:
doc/rados: s/cepgsqlite/cephsqlite/

Reviewed-by: Patrick Donnelly <pdonnell@redhat.com>
Reviewed-by: Anthony D Atri <anthony.datri@gmail.com>

Merge pull request #57143 from adk3798/nvmeof-osd-profile-rbd

mgr/cephadm: set OSD cap for NVMEoF daemon to "profile rbd"

Reviewed-by: Ernesto Puerta <epuertat@redhat.com>
Reviewed-by: Nizamudeen A <nia@redhat.com>

Merge pull request #56575 from cloudbehl/ceph-cluster-json-update

monitoring/ceph-mixin: Add cluster variable to ceph-cluster.json

Reviewed-by: Aashish Sharma <aasharma@redhat.com>

Merge pull request #55495 from frittentheke/issue_64321

monitoring/ceph-mixin: Cleanup of variables, queries and tests (to fix showMultiCluster=True)

Reviewed-by: Aashish Sharma <aasharma@redhat.com>
Reviewed-by: Ankush Behl <cloudbehl@gmail.com>
Reviewed-by: Nizamudeen A <nia@redhat.com>

Merge pull request #57151 from rhcs-dashboard/fix-65698-main

mgr/dashboard: exclude cloned-deleted RBD snaps

Reviewed-by: Ilya Dryomov <idryomov@redhat.com>
Reviewed-by: Nizamudeen A <nia@redhat.com>

doc/rados: s/cepgsqlite/cephsqlite/

As stated in the commit-message line, this corrects the typo "cepg" to
the correct string "ceph".

This typo was discovered by https://github.com/test-erik and this was
brought to our attention way back in
https://github.com/ceph/ceph/pull/50420.

Signed-off-by: Zac Dover <zac.dover@proton.me>

Merge PR #57183 into main

* refs/pull/57183/head:
script/ptl-tool: add switch to update existing qa ticket

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

Merge PR #57059 into main

* refs/pull/57059/head:
mds: do not try fragmenting or exporting a quiesced directory
mds: set/test ALL_LOCKED on fragment_dir request
mds: pass bypassfreezing to parent auth pin req
qa: add quiesce tests during fragmentation
qa: translate empty output from rank_tell to empty dict
qa: move reqid_tostr helper

Reviewed-by: Leonid Usov <leonid.usov@ibm.com>

Merge pull request #57102 from xxhdx1985126/wip-65672

crimson/osd/osdop_params:Unify OpsExecuter::user_modify and osd_op_params_t::user_modify

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

Merge pull request #56995 from athanatos/sjust/wip-65185-scrub-attr-error

osd: only call stat/getattrs once per object during deep-scrub

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

Merge PR #56997 into main

* refs/pull/56997/head:
pybind/mgr: disable sqlite3/python autocommit
qa/tasks/mgr: add tests for sqlite autocommit
qa/tasks/vstart_runner: run daemons in foreground
qa/tasks/vstart_runner: add missing poll method
qa/suites/rados/mgr: add cli/devicehealth tasks
qa: reorganize mgr unit tests
qa: use position-independent link
qa: add missing terminating newline
pybind/mgr: add killpoint for sqlite3 database setup
mgr: allow specifying module option level
mon/MgrMonitor: promote standby when unsetting down flag
mon/MgrMonitor: only drop active if exists

Reviewed-by: Ernesto Puerta <epuertat@redhat.com>

Merge pull request #57144 from adk3798/stray-laundry

qa/cephadm: ignore stray daemon warning during rados_api_tests

Reviewed-by: Laura Flores <lflores@ibm.com>

Merge pull request #56961 from adk3798/agent-check-error-before-json

cephadm: have agent check for errors before json loading mgr response

Reviewed-by: <mfritch@suse.com>

Merge pull request #56933 from adk3798/service-size-cleanup

mgr/cephadm: cleanup service size logic block

Reviewed-by: Michael Fritch <mfritch@suse.com>

Merge pull request #56928 from adk3798/grafana-anon-access-dropped

python-common: handle "anonymous_access: false" in to_json of Grafana spec

Reviewed-by: Nizamudeen A <nia@redhat.com>

Merge pull request #56350 from phlogistonjohn/jjm-mgr-smb-module

mgr: add a new smb cluster & share module

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

Merge pull request #56817 from guits/cephadm-check-apply-spec-file

cephadm: check if file exists when passing `--apply_spec`

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

Merge pull request #56487 from adk3798/bootstrap-custom-alerts

cephadm: allow passing custom prometheus alerts to bootstrap

Reviewed-by: John Mulligan <jmulligan@redhat.com>

script/ptl-tool: add switch to update existing qa ticket

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

mgr/dashboard: exclude cloned-deleted RBD snaps

Fixes: https://tracker.ceph.com/issues/65698
Signed-off-by: Ernesto Puerta <epuertat@redhat.com>

Merge PR #57044 into main

* refs/pull/57044/head:
qa: increase debugging for snap_schedule

Reviewed-by: Milind Changire <mchangir@redhat.com>

Merge PR #57043 into main

* refs/pull/57043/head:
client: clear resend_mds only after sending request

Reviewed-by: Milind Changire <mchangir@redhat.com>

Merge PR #57020 into main

* refs/pull/57020/head:
mds/cache: don't assume non-auth xlocks to be remote locks

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

Merge PR #56935 into main

* refs/pull/56935/head:
mds: encode flags for all inode types
qa: test file inode with F_QUIESCE_BLOCK is replicated

Reviewed-by: Leonid Usov <leonid.usov@ibm.com>

Merge PR #56934 into main

* refs/pull/56934/head:
mds: move drop_locks to directly after rdonly check
qa: test quiesce.block is replicated
qa: test that ceph.dir.subvolume is replicated properly
mds: add debug "lock path" command
qa: move reqid_tostr helper
qa: return run_shell process for waiters

Reviewed-by: Leonid Usov <leonid.usov@ibm.com>

Merge PR #56923 into main

* refs/pull/56923/head:
qa: extend rank 1 lockup for test_quiesce_authpin_wait

Reviewed-by: Leonid Usov <leonid.usov@ibm.com>

Merge pull request #57148 from zdover23/wip-doc-2024-04-30-cephadm-services-monitoring

doc/cephadm: add default monitor images

Merge pull request #57153 from rkhudov/doc-man-8-radosgw-admin_add-get-lifecycle-config-command

doc/man/8/radosgw-admin: add get lifecycle command

Reviewed-by: Zac Dover <zac.dover@proton.me>

Merge pull request #57154 from cityofships/main

doc/rados/operations: rephrase OSDs peering

Reviewed-by: Zac Dover <zac.dover@proton.me>

Merge pull request #55978 from xxhdx1985126/wip-librbd-unittest-sparsify

test/librbd/test_internal: change to check for 4KB-aligned sparse read

Reviewed-by: Samuel Just <sjust@redhat.com>
Reviewed-by: Ilya Dryomov <idryomov@gmail.com>

doc/rados/operations: rephrase OSDs peering

Signed-off-by: Piotr Parczewski <piotr@stackhpc.com>

Merge pull request #55823 from ceph/wip-lusov-quiescer

qa: cephfs quiesce thrasher

doc/man/8/radosgw-admin: add get lifecycle command

Signed-off-by: rkhudov <rhudov@gmail.com>

Merge pull request #56806 from xxhdx1985126/wip-65399

crimson/osd/pg_recovery: backoff if the recovery/backfill is deferred

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

Merge pull request #56998 from xxhdx1985126/wip-65568

crimson/osd/pg: only kick off snap trimming when PGs are clean

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

Merge pull request #57128 from cyx1231st/wip-fix-ut-seastore-timeout

crimson/os/seastore: misc cleanups and fixes

Reviewed-by: Xuehan Xu <xuxuehan@qianxin.com>

doc/cephadm: add default monitor images

Add a list of default monitor images to the documentation. This commit
is made in response to a request from Eugen Block, and is made using the
information developed by Mr Block here:
https://lists.ceph.io/hyperkitty/list/ceph-users@ceph.io/thread/QGC66QIFBKRTPZAQMQEYFXOGZJ7RLWBN/.

Signed-off-by: Zac Dover <zac.dover@proton.me>

Merge pull request #56892 from xxhdx1985126/wip-crimson-urgent-recovery-version

crimson/osd/osd_operations/client_request: retrieve the correct version for objects to be recovered urgently

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

Merge pull request #56610 from Matan-B/wip-matanb-crimson-replicated-head-obc

crimson/osd/replicated_recovery_backend: Fix recovery obc usage

Reviewed-by: Samuel Just <sjust@redhat.com>
Reviewed-by: Xuehan Xu <xuxuehan@qianxin.com>

Merge pull request #57135 from xxhdx1985126/wip-seastore-checksum-assertion-fix

crimson/os/seastore/transaction_manager: remove incorrect assertions

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

Merge pull request #57129 from xxhdx1985126/wip-65585

crimson/os/seastore/transaction_manager: fix write pipeline phase leak

Reviewed-by: Yingxin Cheng <yingxin.cheng@intel.com>
Reviewed-by: Samuel Just <sjust@redhat.com>

mds: do not try fragmenting or exporting a quiesced directory

And handle inode becoming quiesced after op is created.

Fixes: https://tracker.ceph.com/issues/65603
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>

mds: set/test ALL_LOCKED on fragment_dir request

This is an optimization to obviate repeated calls to acquire_locks.

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

mds: pass bypassfreezing to parent auth pin req

Credit to Leonid for first noticing this.

Fixes: https://tracker.ceph.com/issues/65603
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>

qa: add quiesce tests during fragmentation

Reproduces: https://tracker.ceph.com/issues/65603
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>

qa/suites/fs: add quiescer to the fs suite

Signed-off-by: Leonid Usov <leonid.usov@ibm.com>

qa/tasks: the quiescer task and a waiter task to test it

Fixes: https://tracker.ceph.com/issues/63669
Signed-off-by: Leonid Usov <leonid.usov@ibm.com>

qa: translate empty output from rank_tell to empty dict

Not all commands return JSON, like `dirfrag split`.

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

qa: move reqid_tostr helper

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

pybind/mgr: disable sqlite3/python autocommit

SQLite3 and python's binding were both committing transactions at unintended
points. Turn it off and stop using executescript.

Fixes: https://tracker.ceph.com/issues/65494
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>

qa/tasks/mgr: add tests for sqlite autocommit

That autocommit is properly turned off and that commits via context managers
work as expected.

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

qa/cephadm: ignore stray daemon warning during rados_api_tests

The "stray daemon" that is getting logged about in this test is
from "stray daemon laundry.pid70383 on host smithi027 not managed by cephadm".
It seems the rados_api_tests is creating some additional "laundry" entity
during these tests that gets reported as an actual daemon in the mgr,
but cephadm is unaware of it, resulting in the warning. Originally
we thought to maybe add "laundry" itself to the ignorelist, but
without an additional patch that added extra logging for debug
purposes (which can't be merged) the log statement found in
the logs due to this problem will not say what daemon it found
to be stray. There will just be a generic warning about a stray
daemon. In a real cluster, a user would then check "ceph health detail"
to find out what daemon is stray, but the log scraper can't do this
and just fails the test due to the presence of the warning.

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

mgr/cephadm: set OSD cap for NVMEoF daemon to "profile rbd"

Otherwise it is impossible to make rbd snapshots

Fixes: https://tracker.ceph.com/issues/65691
Signed-off-by: Adam King <adking@redhat.com>

qa/tasks/vstart_runner: run daemons in foreground

This mirrors teuthology and makes it possible to check the exit status of a
daemon.

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

qa/tasks/vstart_runner: add missing poll method

Otherwise you cannot use LocalDaemon.check_status.

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

qa/suites/rados/mgr: add cli/devicehealth tasks

These should have been part of the commit adding the tests.

Fixes: 9ebcbdbed0fbaf37d60fbd8db321742ee5721d8c
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>

qa: reorganize mgr unit tests

Refactor common tasks and allow loading mgrmodules before unittests start.

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

qa: use position-independent link

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

qa: add missing terminating newline

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

pybind/mgr: add killpoint for sqlite3 database setup

To catch unintentional autocommit by sqlite3.

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

mgr: allow specifying module option level

Some are for development purposes and should be filtered out by the dashboard.

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

mon/MgrMonitor: promote standby when unsetting down flag

In the same epoch, to simplify testing.

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

mon/MgrMonitor: only drop active if exists

This avoids an assertion in ::drop_active.

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

Merge pull request #57136 from zdover23/wip-doc-2024-04-29-radosgw-bucketpolicy

doc/radosgw: disambiguate version-added remarks

doc/radosgw: disambiguate version-added remarks

Rewrite two sentences so that they explain precisely what was added to
Ceph and when it was added to Ceph.

This commit is made in response to a request made by Dr Per Mildner on
https://pad.ceph.com/p/Report_Documentation_Bugs.

Signed-off-by: Zac Dover <zac.dover@proton.me>

crimson/osd/osd_operation: yield the cpu before another round the
"with_throttle"

This is to avoid infinite recursive calling to "with_throttle_while"

Signed-off-by: Xuehan Xu <xxhdx1985126@gmail.com>

crimson/osd/pg_recovery: backoff if the recovery/backfill is deferred

Fixes: https://tracker.ceph.com/issues/65399
Signed-off-by: Xuehan Xu <xuxuehan@qianxin.com>

crimson/os/seastore: drop the impossible error eagain

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

crimson/os/seastore/extent_placement_manager: make reserve and block atomic

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

crimson/os/seastore: improve seastore_t logs

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

crimson/os/seastore/transaction_manager: remove incorrect assertions

Indirect mappings always hold the checksums of the corresponding intermediate
mappings, which may have non-zero

Signed-off-by: Xuehan Xu <xuxuehan@qianxin.com>

crimson/os/seastore: add logs for background

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

test/crimson/gtest_seastar: add logs

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

test/crimson/gtest_seastar: cleanups

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

Merge pull request #55449 from xxhdx1985126/wip-seastore-extent-chksum

crimson/os/seastore: support extent checksum verification

Reviewed-by: Yingxin Cheng <yingxin.cheng@intel.com>
Reviewed-by: Chunmei Liu <chunmei.liu@intel.com>

crimson/os/seastore/transaction_manager: fix write pipeline phase leak

At present, if a transaction gets interrupted right after it enters
WritePipeline::ReserveProjectedUsage and before any later continuations
get executed, WritePipeline::ReserveProjectedUsage will be locked
forever.

Signed-off-by: Xuehan Xu <xuxuehan@qianxin.com>

Merge pull request #57088 from xxhdx1985126/wip-seastore-update-onode-size

crimson/os/seastore: only update onode sizes when necessary

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

Merge pull request #57118 from xxhdx1985126/wip-65673

test/ceph_crypto: define __has_feature if the compiler doesn't have it

Reviewed-by: Kefu Chai <kefu.chai@scylladb.com>

Merge pull request #57119 from zdover23/wip-doc-2024-04-28-cephfs-client-auth

doc/cephfs: edit front matter in client-auth.rst

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

doc/cephfs: edit front matter in client-auth.rst

Clarify the front matter in doc/cephfs/client-auth.rst.

Co-authored-by: Anthony D'Atri <anthony.datri@gmail.com>
Signed-off-by: Zac Dover <zac.dover@proton.me>

Merge pull request #57112 from tchaikov/erasure-code-free

erasure-code/shec: use free() to release alloc()'ed memory chunk

Reviewed-by: Rongqi Sun <sunrongqi@huawei.com>

Merge pull request #57113 from tchaikov/erasure-code-free-crush-wrapper

test//erasure-code/TestErasureCodeLrc: free allocated CrushWrapper

Reviewed-by: Rongqi Sun <sunrongqi@huawei.com>

Merge pull request #57111 from tchaikov/erasure-code-nullptr

erasure-code/shec: replace 0 with nullptr when appropriate

Reviewed-by: Rongqi Sun <sunrongqi@huawei.com>

crimson/osd/osdop_params:Unify OpsExecuter::user_modify and osd_op_params_t::user_modify

Before this change OpsExecuter::user_modify was maintained in OpsExecuter::do_write_op.
However, osd_op_params->user_modify was not updated when used in OpsExecuter::prepare_transaction

Fixes: https://tracker.ceph.com/issues/65672
Signed-off-by: Xuehan Xu <xxhdx1985126@gmail.com>

crimson/osd/pg: SnapTrimEvent to support interrupts

SnapTrimEvent operations are scheduled from `PG::on_active_actmap()`
using a `seastar::do_until` loop. This commit replaces the loop type
into an `interruptor::repeat` and SnapTrimEvent are now scheduled by
`start_operation_may_interrupt`.

Previously, `SnapTrimEvent::start` handled interruptions by returning
a `crimson::ct_error::eagain::make();`. Now, the errorator is directly
returned via the `snap_trim_event_ret_t` and interrupts the loop
described above.

As a result, interruptions originated by interval changes are now
supported by SnapTrimEvent.

Signed-off-by: Xuehan Xu <xuxuehan@qianxin.com>

test/ceph_crypto: define __has_feature if the compiler doesn't have it

Refer to https://gcc.gnu.org/onlinedocs/cpp/_005f_005fhas_005ffeature.html
and https://clang.llvm.org/docs/LanguageExtensions.html#has-feature-and-has-extension
for further information

Fixes: https://tracker.ceph.com/issues/65673
Signed-off-by: Xuehan Xu <xuxuehan@qianxin.com>

test//erasure-code/TestErasureCodeLrc: free allocated CrushWrapper

in TestErasureCodeLrc.cc, we create a `CrushWrapper` without
destroying and freeing it. and ASan points this out:
```
Indirect leak of 72000 byte(s) in 1000 object(s) allocated from:
    #0 0x559e00acc2bd in operator new(unsigned long) (/home/jenkins-build/build/workspace/ceph-pull-requests/build/bin/unittest_erasure_code_lrc+0x1ed2bd) (BuildId: 1cd93d0f231006242239eb0d81a8d677e36aeba2)
    #1 0x559e00b2bb76 in __gnu_cxx::new_allocator<std::_Rb_tree_node<std::pair<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int> > >::allocate(unsigned long, void const*) /usr/bin/../lib/gcc/x86_64-linux-gnu/11/../../../../include/c++/11/ext/new_allocator.h:127:27
    #2 0x559e00b2bb00 in std::allocator<std::_Rb_tree_node<std::pair<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int> > >::allocate(unsigned long) /usr/bin/../lib/gcc/x86_64-linux-gnu/11/../../../../include/c++/11/bits/allocator.h:185:32
    #3 0x559e00b2bb00 in std::allocator_traits<std::allocator<std::_Rb_tree_node<std::pair<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int> > > >::allocate(std::allocator<std::_Rb_tree_node<std::pair<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int> > >&, unsigned long) /usr/bin/../lib/gcc/x86_64-linux-gnu/11/../../../../include/c++/11/bits/alloc_traits.h:464:20
    #4 0x559e00b2b9d1 in std::_Rb_tree<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::pair<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int>, std::_Select1st<std::pair<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int> >, std::less<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > >, std::allocator<std::pair<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int> > >::_M_get_node() /usr/bin/../lib/gcc/x86_64-linux-gnu/11/../../../../include/c++/11/bits/stl_tree.h:561:16
    #5 0x559e00b2aae4 in std::_Rb_tree_node<std::pair<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int> >* std::_Rb_tree<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::pair<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int>, std::_Select1st<std::pair<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int> >, std::less<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > >, std::allocator<std::pair<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int> > >::_M_create_node<std::piecewise_construct_t const&, std::tuple<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&>, std::tuple<> >(std::piecewise_construct_t const&, std::tuple<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&>&&, std::tuple<>&&) /usr/bin/../lib/gcc/x86_64-linux-gnu/11/../../../../include/c++/11/bits/stl_tree.h:611:23
    #6 0x559e00b29520 in std::_Rb_tree_iterator<std::pair<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int> > std::_Rb_tree<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::pair<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int>, std::_Select1st<std::pair<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int> >, std::less<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > >, std::allocator<std::pair<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int> > >::_M_emplace_hint_unique<std::piecewise_construct_t const&, std::tuple<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&>, std::tuple<> >(std::_Rb_tree_const_iterator<std::pair<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int> >, std::piecewise_construct_t const&, std::tuple<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&>&&, std::tuple<>&&) /usr/bin/../lib/gcc/x86_64-linux-gnu/11/../../../../include/c++/11/bits/stl_tree.h:2431:19
    #7 0x559e00b255c2 in std::map<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >, int, std::less<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > >, std::allocator<std::pair<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int> > >::operator[](std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&) /usr/bin/../lib/gcc/x86_64-linux-gnu/11/../../../../include/c++/11/bits/stl_map.h:501:15
    #8 0x559e00b188da in CrushWrapper::set_item_name(int, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&) /home/jenkins-build/build/workspace/ceph-pull-requests/src/crush/CrushWrapper.h:466:7
    #9 0x7feb54a4dcc2 in CrushWrapper::insert_item(ceph::common::CephContext*, int, float, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::map<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::less<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > >, std::allocator<std::pair<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > > > > const&, bool) /home/jenkins-build/build/workspace/ceph-pull-requests/src/crush/CrushWrapper.cc:1119:5
    #10 0x559e00ada922 in ErasureCodeTest_create_rule_Test::TestBody() /home/jenkins-build/build/workspace/ceph-pull-requests/src/test/erasure-code/TestErasureCodeLrc.cc:127:5
    #11 0x559e00c0e0a6 in void testing::internal::HandleSehExceptionsInMethodIfSupported<testing::Test, void>(testing::Test*, void (testing::Test::*)(), char const*) /home/jenkins-build/build/workspace/ceph-pull-requests/src/googletest/googletest/src/gtest.cc:2605:10
```

so, in this change, let's manage the lifecycle of the `CrushWrapper`
instance with a smart pointer, so that it is destroyed and free'd
properly, and this should silence the ASan warning.

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