client: expose ceph.quota.max_bytes xattr within snapshots

For directories within snapshots, expose the ceph.quota.max_bytes
extended attribute information. This enables fetching quota
information when the snapshot was taken and is particularly useful
when cloning subvolume snapshots, to enforce the quota on the
clone subvolume as well.

Fixes: https://tracker.ceph.com/issues/46278
Signed-off-by: Shyamsundar Ranganathan <srangana@redhat.com>

Merge pull request #36323 from tchaikov/wip-crimson-msgr-v1-v2

crimson: picking peer addr of the compatible type

Reviewed-by: Neha Ojha <nojha@redhat.com>
Reviewed-by: Yingxin Cheng <yingxin.cheng@intel.com>

Merge pull request #35728 from jan--f/c-v-add-subcommand-parse-drive-groups

ceph-volume: add drive-group subcommand

Merge pull request #36342 from tchaikov/wip-crimson-heartbeat-erase

crimson/osd: erase an element by iterator instead

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

Merge pull request #36217 from sebastian-philipp/cephadm-common-mypy-ini

cephadm: use src/mypy.ini instead

Reviewed-by: Joshua Schmid <jschmid@suse.de>
Reviewed-by: Michael Fritch <mfritch@suse.com>

Merge pull request #36235 from matthewoliver/cephadm_iscsi_tcmu_runner

cephadm: Add tcmu-runner container when deploying ceph-iscsi

Reviewed-by: Kefu Chai <kchai@redhat.com>
Reviewed-by: Sebastian Wagner <sebastian.wagner@suse.com>

crimson/osd: implement cls_get_pool_stripe_width

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

crimson/osd: erase an element by iterator instead

we should not remove an element while iterating it in a map, as erasing
the element invalidates the iterator, which causes segmfault when we are
advancing it after erasing the dereferenced element.

in this change, an iterator is used for walking through the map, in
comparision with creating a to-be-removed list, this one is more
efficient and more idiomatic.

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

Merge pull request #36341 from tchaikov/wip-crimson-cls

crimson/osd: correct the function name of cls_cxx_map_get_vals_by_keys()

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

crimson/osd: correct the function name of cls_cxx_map_get_vals_by_keys()

it was an oversight in 7a4c6359e483f8c71276ece5cde16eb0771ac5d2

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

Merge pull request #36079 from winndows/superfluous_break6

msg: Remove superfluous breaks

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

Merge pull request #36297 from dvanders/dvanders_46443

osd: fix crash in _committed_osd_maps if incremental osdmap crc fails

Reviewed-by: Xiaoxi Chen <xiaoxchen@ebay.com>
Reviewed-by: David Zafman <dzafman@redhat.com>
Reviewed-by: Neha Ojha <nojha@redhat.com>

qa/suites/rados/thrash/crc-failures: randomly inject bad incremental osdmap crc

Signed-off-by: Neha Ojha <nojha@redhat.com>

osd: don't write transaction when inc crc failed

80da5f9a987c6a48b93f25228fdac85890013520 exposed a flaw in how
handle_osd_map falls back to a full osdmap if the crc of an incremental
failed.

If the first message in a map message had a crc error, then the
loop would exit with last < start, which would then cause a null
dereference in _committed_osd_maps.

Fixes: https://tracker.ceph.com/issues/46443
Signed-off-by: Dan van der Ster <daniel.vanderster@cern.ch>

qa/standalone/osd: add bad-inc-map.sh

Test that the osd doesn't crash when it gets a bad incremental osdmap.

Related-to: https://tracker.ceph.com/issues/46443
Signed-off-by: Dan van der Ster <daniel.vanderster@cern.ch>

Merge pull request #36287 from dillaman/wip-librbd-close

librbd: use task finisher thread for image open/close callbacks

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

Merge pull request #36253 from changchengx/exclusive

doc: specify RBD_LOCK_MODE_EXCLUSIVE for exclusive-lock

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

Merge pull request #36328 from tchaikov/wip-crimson-cls_cxx_map_get_vals

crimson/osd: implement cls_cxx_map_get_vals()

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

crimson/osd: implement cls_cxx_map_get_vals()

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

doc: specify RBD_LOCK_MODE_EXCLUSIVE for exclusive-lock

The exclusive-lock could be transited transparently between clients
after finishing write operation. To disable "transparent" transition,
it needs to acquire the lock with RBD_LOCK_MODE_EXCLUSIVE.

Signed-off-by: Changcheng Liu <changcheng.liu@aliyun.com>

Merge pull request #34928 from p-se/wip-pse-revise-monitoring-doc

mgr/dashboard: revise monitoring documentation

Reviewed-by: Lenz Grimmer <lgrimmer@suse.com>
Reviewed-by: Stephan Müller <smueller@suse.com>

crimson: use pick_addr() for picking peer addr

in teuthology tests, there is good chance that we have ceph.conf
containing:

mon host = 172.21.15.122

which is translated to two monitors

- a: 172.21.15.122:3300
- a-legacy: 172.21.15.122:6789

both has protocol type of "any". so, to enable crimson to use settings
like this, we should let crimson to accept them, and drop the connection
if the peer claim to be using an incompatible protocol, when they are
exchanging banners.

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

crimson/mon: use mon with only v2 address

crimson msgr supports v2 protocol now, so we can connect to monitor
which only provides v2 addresses.

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

msg/msg_types.h: add pick_addr()

for picking an addr from an entity_addrvec_t by given protocol type.
so:
  - v2 => v2, any
  - v1 => v1, any
  - any => any, v1, v2

and add a helper of `addr_of_type()` to avoid repeatings.

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

Merge pull request #36285 from sebastian-philipp/orch-completion-generic

mgr/orch: Add some more type annotations

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

Merge pull request #36258 from rhcs-dashboard/fix-cpu-stats

mgr/dashboard: cpu stats incorrectly displayed

Reviewed-by: Patrick Seidensal <pseidensal@suse.com>
Reviewed-by: Alfonso Martínez <almartin@redhat.com>
Reviewed-by: Volker Theile <vtheile@suse.com>

mgr/orch: Add some more type annotations

Made `orch.Completion` a generic type

Signed-off-by: Sebastian Wagner <sebastian.wagner@suse.com>

Merge pull request #36012 from adk3798/cephadm_44886

mgr/cephadm: allow use of authenticated registry

Merge pull request #36262 from sebastian-philipp/orch-readd-apply_dg

mgr/cephadm: re-add `apply_drivegroups()`

Reviewed-by: Joshua Schmid <jschmid@suse.de>
Reviewed-by: Kiefer Chang <kiefer.chang@suse.com>

Merge pull request #36306 from smithfarm/wip-add-octopus-to-release-table

doc/releases: add "octopus" column to Release Timeline

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

Merge pull request #36301 from sebastian-philipp/doc-cephadm-status-no-progress

doc/cephadm: `status` doesn't show a progress

Reviewed-by: Michael Fritch <mfritch@suse.com>
Reviewed-by: Zac Dover <zac.dover@gmail.com>

cephadm: Add tcmu-runner container when deploying ceph-iscsi

Currently when we deploy ceph-iscsi via cephadm it doesn't include a
running tcmu-runner. Which means initiators will be able to login but
you wont see the LUNS on the initiator.

This patch deploys an additional tcmu-runner container along side the
ceph-iscsi container that just runs the tcmu-runner service.

Fixes: https://tracker.ceph.com/issues/46540
Signed-off-by: Matthew Oliver <moliver@suse.com>

Merge pull request #36090 from inspur-wyq/wip-37532

mon: fix the 'Error ERANGE' message when conf "osd_objectstore" is filestore

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

Merge pull request #36283 from rzarzynski/wip-bl-raw-privatization

common/bl: don't access raw::data nor raw::len directly. Use getters instead.

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

doc/releases: add "octopus" column to Release Timeline

Octopus has been out for awhile. I suppose this should have been done
earlier, but "better late than never".

Signed-off-by: Nathan Cutler <ncutler@suse.com>

Merge pull request #36245 from smithfarm/wip-mimic-is-eol

doc/releases: Mimic is EOL

Reviewed-by: Abhishek Lekshmanan <abhishek@suse.com>
Reviewed-by: Neha Ojha <nojha@redhat.com>

Merge pull request #36279 from tchaikov/wip-crimson-msgr-v2.1

crimson/net: enable on-wire-encryt and v2.1 support

Reviewed-by: Yingxin Cheng <yingxin.cheng@intel.com>
Reviewed-by: Josh Durgin <jdurgin@redhat.com>

doc/cephadm: `status` doesn't show a progress

Fixes: https://tracker.ceph.com/issues/45858
Signed-off-by: Sebastian Wagner <sebastian.wagner@suse.com>

Merge pull request #35852 from smithfarm/wip-opensuse-os-recommendations

doc/start/os-recommendations: current state of openSUSE

Reviewed-by: Tim Serong <tserong@suse.com>

Merge pull request #36269 from dang/wip-dang-46692

RGW - fix bulkupload, broken by zipper

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

cephadm: use src/mypy.ini instead

Signed-off-by: Sebastian Wagner <sebastian.wagner@suse.com>

librbd: use task finisher thread for image open/close callbacks

There was a potential race condition with utilizing the AsioEngine
to deliver asynchronous image open and close callbacks. This left
the potential for the io_context thread to attempt to destroy itself.

This commit changes the behavior of the image open and close callbacks
to always delete the ImageCtx (now matches the synchronous API behavior)
and it always invokes the callback in Finisher thread whose lifetime is
tied to the CephContext.

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

Merge pull request #36219 from guits/guits-fix_zap_osdid_osdfsid

ceph-volume: filter by osd-id or osd-fsid when zapping

ceph-volume: filter by osd-id or osd-fsid when zapping

2f5c10c12c37e6865ce54bb4940d3779353cba4f introduced a bug:

`ceph-volume lvm zap` command fails under certain conditions.

when passing `--osd-id` or `--osd-fsid` to `ceph-volume lvm zap` command
it tries to zap additionnal devices that have nothing to do with the osd
being zapped.

When calling `api.get_lvs()` in `ensure_associated_lvs()` we have to
pass the osd-id/osd-fsid information so only related devices are
returned by `get_lvs()` method

Closes: https://tracker.ceph.com/issues/46627
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>

messages/MOSDBoot: pass OSDSuperblock by const ref

MOSDBoot's ctor does not change the parameter, so let's pass by const
reference.

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

crimson/os/alienstore: always use fsid in bluestore

alienstore should not be stateful in this perspective, it should proxy
all acccess of fsid to bluestore.

there are couple issues in existing implementation:

* when mkfs, bluestore tries to generate a new osd_fsid if the specified
  one is empty. but we explicitly pass the given uuid down to
  AlienStore::mkfs() so the bluestore can use it. so we should pass it
  down instad of storing it locally.
* when persisting superblock in OSD::mkfs(), superblock.osd_fsid() is
  read from store->get_fsid(), if user specifies an empty uuid, we
  should persist the generated uuid in the superblock.

in this change, all access to fsid is proxied to the underlying
bluestore.

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

stop.sh: stop osd before mon

osd sends a MOSDMarkMeDown message to monitor and waits for its ack
before timeout, so if we can stop osd before stopping mon, stop.sh can
return sooner without waiting until the timeout.

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

crimson/mgr: only pick the addr of the same type

to avoid the attempts to connect an OSD which is bound to a v2
address to a v1 address of a mgr.

in general, osd is bound to both v1 and v2 addresses, but crimson
msgr does not support multiple bound address at the time of writing, so
to avoid the failures when trying to connect to incompatible addresses,
let's filter out them when connecting to monitor. this change
silence warnings like:

peer_addr_for_me v1:172.21.15.106:60008/0 type doesn't match myaddr
v2:0.0.0.0:6802/26710

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

mon/MgrMap: let get_active_addrs() return a const ref

no need to create a temporary instance for referencing those addresses.

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

crimson/mon: only pick the addr of the same type

to avoid the attempts to connect an OSD which is bound to a v2 address
to a v1 addrss of a monitor.

in general, osd is bound to both v1 and v2 addresses, but crimson msgr
does not support multiple bound address at the time of writing, so to
avoid the failures when trying to connect to incompatible addresses,
let's filter out them when connecting to monitor. this change silence
warnings like:

peer_addr_for_me v1:172.21.15.106:60008/0 type doesn't match myaddr
v2:0.0.0.0:6802/26710

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

crimson/osd: print out client caps

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

auth/cephx: implement random()->get_bytes() for crimson

instead of using CryptoRandom use the C++ standard library for
generating secret.

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

crimson/admin: catch thrown exception

if the socket file exists, a std::system_error is thrown. and we should
catch it.

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

crimson/msgr: Revert "don't advertise the on-wire format v2.1."

This reverts commit a74948bc5095b32212189352c163030bfe10db71.

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

crimson/net: enable msgr v2.1 support

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

crimson/net: enable on_wire encryption support

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

crimson/net: set is_rev1 for messenger v2.1 support

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

crimson/net: keep rx_preamble for msgr v2.1 support

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

crimson/net: drop stale TODO

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

crimson/net: use rx_frame_asm for handling data read from wire

by leveraging FrameAssembler, it's much simpler. and it also pave the
road to a better messenger v2.0 and v2.1 protocol support.

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

crimson/net: mark abort_ functions [[noreturn]]

otherwise compiler complains if control reaches end of non-void
function.

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

msg/async/crypto_onwire: drop unused member variable

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

Merge pull request #36259 from majianpeng/bluefs-reduce-ceph_clock_now

os/bluestore/BlueFS: reduce unnecessary ceph_clock_now().

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

Merge pull request #33899 from rs-fabrica/rados_generic_options_usage_message

rados: include generic options in usage message

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

Merge pull request #36115 from BenoitKnecht/diskprediction-local-array-shape

mgr/diskprediction_local: Fix array size error

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

Merge pull request #35306 from changchengx/blk

blk: add option to set device type to select blk driver

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

Merge pull request #36274 from xiexingguo/wip-peer-num-objects

osd/PeeringState: prevent peer's num_objects going negative

Reviewed-by: Neha Ojha <nojha@redhat.com>
Reviewed-by: yanjun <yan.jun8@zte.com.cn>

Merge pull request #36236 from tchaikov/wip-std-bind

test/librados_test_stub: use std::bind

Reviewed-by: Adam C. Emerson <aemerson@redhat.com>

Merge pull request #36071 from rzarzynski/wip-crimson-errorator-assert-cleanup

crimson: improve assertions in errorator

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

osd/PeeringState: prevent peer's num_objects going negative

Saw it in a teuthology run:

 -5645> 2020-07-20 04:34:32.067 7f351e329700  5 osd.5 pg_epoch: 667 ... exit Started/Primary/Active/Backfilling
 -5642> 2020-07-20 04:34:32.067 7f351e329700  5 osd.5 pg_epoch: 667 ... enter Started/Primary/Active/Recovered
 -5633> 2020-07-20 04:34:32.067 7f351e329700 20 osd.5 pg_epoch: 667 ... _update_calc_stats shard 5 primary objects 0 missing 0
 -5632> 2020-07-20 04:34:32.067 7f351e329700 20 osd.5 pg_epoch: 667 ... _update_calc_stats shard 3 objects -1 missing 1
 -5631> 2020-07-20 04:34:32.067 7f351e329700 20 osd.5 pg_epoch: 667 ... _update_calc_stats shard 6 objects 0 missing 0

This will crash the choose_acting() procedure as it will mistakenly
think that peer 3 should continue to perform asynchronous recovery
(e.g., due to num_objects_missing = 1) in contrast to fully
backfill-recovered.

While I did not dig into the real cause, there are a couple of
possible explanations of how num_objects can be off. I think that
if a roll forward or log replay could delete something twice, maybe
there would be an undercount. Or maybe something as simple as a
corruption.

Since _update_calc_stats() is going to fix num_objects_missing
for that peer anyway, let's make sure it always starts with a
clean state.

Fixes: https://tracker.ceph.com/issues/46705
Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>

Merge pull request #36121 from aclamk/wip-bluefs-log-replay-rescue

Rescue procedure for extremely large bluefs log

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

Merge pull request #35909 from dzafman/wip-46275

osd: Cancel in-progress scrubs (not user requested)

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

RGW - fix bulkupload, broken by zipper

Bulkupload depended on the existence of empty bucketinfo.  Fix that, to
avoid a crash.  In additions, the error handler for swift used buckets.

Fixes 46692

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

test: Check for interuption of scrubs with nosrub/nodeep_scrub

Signed-off-by: David Zafman <dzafman@redhat.com>

osd: Cancel in-progress scrubs (not user requested)

This change adds new scrubber.req_scrub to track user
requested scrubs, deep_scrub or repair.

Fixes: https://tracker.ceph.com/issues/46275
Signed-off-by: David Zafman <dzafman@redhat.com>

osd: Arrange code so that it is clearer should not cause any change

Signed-off-by: David Zafman <dzafman@redhat.com>

test: mon-last-epoch-clean.sh fixed to avoid shell globbing

Signed-off-by: David Zafman <dzafman@redhat.com>

osd: Fix dump_scrub_reservation usage

Signed-off-by: David Zafman <dzafman@redhat.com>

Merge pull request #35376 from Devp00l/update-backport-doc

doc: Resolving conflicts with ceph-backport.sh

Reviewed-by: Lenz Grimmer <lgrimmer@suse.com>
Reviewed-by: Nathan Cutler <ncutler@suse.com>
Reviewed-by: Laura Paduano <lpaduano@suse.com>

common/bl: don't access raw::len directly. Use the getter instead.

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

common/bl: don't access raw::data directly. Use the getter instead.

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

Merge pull request #36238 from yuvalif/fix_zippet_notif_merge

rgw/notification: fix merge issues from zipper6

Merge pull request #36268 from tchaikov/wip-crimson-msgr

crimson/net: do not reset need_addr before learning it

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

crimson/net: do not reset need_addr before learning it

because we don't bind both v1 and v2 addresses, when monitor returns a
v1 peer address, as the client side, crimson-osd just drops the
connection. but this failed attempt to learn the myaddr resets
`need_addr`. and this prevents crimson-osd from learning the v2 address
returned by monitor.

in this change, we reset need_addr only after it is learned from the
peer.

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

Merge pull request #36256 from tchaikov/wip-ceph-debug-docker-crimson

ceph-debug-docker: add --flavor option

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

ceph-debug-docker: add --flavor option

* add --flavor option, which is "default" by default, so one can, for
  example, pass "--flavor crimson" to ceph-debug-docker
* extract $repo_url to avoid repeating the shared bits between centos
  and debian derivatives envs.

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

os/bluestore/BlueFS:  avoid useless ceph_clock_now() call.

The overhead of utime_t constructor utime_t()  is less than ceph_clock_now().

Signed-off-by: Jianpeng Ma <jianpeng.ma@intel.com>

Merge PR #36136 into master

* refs/pull/36136/head:
qa/tasks/nfs:Add test for relative and just '/' pseudo path
mgr/nfs: Check if pseudo path is absolute path or just '/'

Reviewed-by: Michael Fritch <mfritch@suse.com>
Reviewed-by: Patrick Donnelly <pdonnell@redhat.com>
Reviewed-by: Nathan Cutler <ncutler@suse.com>

mgr/cephadm: allow use of authenticated registry

Add option to use custom authenticated registry during
bootstrap as well as a registry-login command in order
to let user change authenticated registry login info

Fixes: https://tracker.ceph.com/issues/44886
Signed-off-by: Adam King <adking@redhat.com>

Merge pull request #36239 from sebastian-philipp/cephadm-parallel-hosts

mgr/cephadm: create OSDs in parallel

Reviewed-by: Joshua Schmid <jschmid@suse.de>
Reviewed-by: Kiefer Chang <kiefer.chang@suse.com>

Merge pull request #35839 from mgfritch/cephadm-ignore-mon-mgr-svc-id

python-common: clean-up ServiceSpec.service_id handling

Reviewed-by: Joshua Schmid <jschmid@suse.de>
Reviewed-by: Sebastian Wagner <sebastian.wagner@suse.com>

Merge pull request #36251 from neha-ojha/wip-fix-43888

osd/OSD.cc: remove osd_lock for bench

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

Merge pull request #36261 from tchaikov/wip-unforty-seastar

ceph.spec.in: cull _FORTIFY_SOURCE macro from CXXFLAGS for seastar

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

Merge pull request #36252 from tchaikov/wip-doc-crimson

doc/dev/crimson: remove redundant options

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

Merge pull request #36263 from tchaikov/wip-more-log-for-dashboard-test

mgr/dashboard: print more osd log when backend-api-tests fails

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

ceph-volume: add drive-group subcommand

This new subcommand takes a drive group specification as json and deploys
the OSDs accordingly.

Signed-off-by: Jan Fajerski <jfajerski@suse.com>
Fixes: https://tracker.ceph.com/issues/46689

mgr/dashboard: print more log when backend-api-tests fails

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

Merge pull request #36242 from dillaman/wip-46668

librbd: flush all queued object IO from simple scheduler

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

doc/dev/crimson: add more examples for seastar-addr2line

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

mgr/cephadm: re-add `apply_drivegroups()`

Fixes: d348d7bf8d3663140c089937b62a0b316b69176b
Fixes: https://tracker.ceph.com/issues/46681
Signed-off-by: Sebastian Wagner <sebastian.wagner@suse.com>