mds: use 128 bits for waiters on MDSCacheObject

Adding a new inode lock will overflow inode wait bits into the MDSCacheObject
wait bits. Make space for the quiescelock.

This includes a minor refactor to no longer attempt scoping the set of masks we
test in MDSCacheObject::waiting when calling MDSCacheObject::is_waiter_for.
This optimization wasn't worth the overhead and would be awkard to keep as
std::bitset cannot be used as a key for a std::multimap (easily).  Instead, we
use the sequence number as a key which helps us to avoid allocating another map
whenever we call MDSCacheObject::take_waiting.

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit 84b33ea9fe6268adfaca1e17b282e68788e38bf4)

mds: provide mechanism to authpin while freezing

When a subtree is freezing, it's no longer possible to acquire new authpins.
This is a problem when a compound request like quiescing a subtree is trying to
acquire authpins for each sub-op. This creates a situation where some quiesce
sub-ops complete with authpins (thereby preventing the tree from becoming
"frozen") and new sub-ops cannot acquire authpins (because the tree is
"freezing"). To circumvent this, allow some authpin requests to proceed if
FLAG_BYPASSFREEZING is set.

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit a9964a7ccc4394f923fb0f1c76eb8fa03fe8733d)

mds: add command to get specific op

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit 3197a975a521654500173175c6e856c54df497ef)

mds: finish request before completing internal req

Importantly, this marks the request dead so it cannot be killed.

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit c944eadaf6987a812e738459214872af1ff7678a)

mds: complete internal op if killed

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit ba3dac94e6a90ed76970fef89b1a82dcedb4ec2b)

mds: avoid killing dead requests

This is important when you have:

* sub-op A finishes and resumes parent
* parent kills all sub-ops (including sub-op A)

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit b72e153d2b36956e97b8686aeb6aaacffe928750)

mds: add command to kill request

Fixes: https://tracker.ceph.com/issues/63374
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit 9956f9ad78a3b6ebe520dcd10e33a59c361c4d91)

mds: add path argument to `ops` and `dump tree` to stream result to local file

This file can be collected during testing to avoid sending a large JSON result
over the `ceph tell`/asok interface.

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit a1303a52af7504c3748a75ade9619c7a6c454831)

mds: print internal_request filepaths if present

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit eaea93e05018294e893e0e5241b3293e10971f5e)

mds: add more information to debug message

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit f8a26a61305ba8b9fbb9ce47b5221e3adb754c9b)

mds: remove redundant parenthesis

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit 1513327c1a55614ac7e2f1a6da0de06e5fd2214d)

mds: implement Mutation::dump method

So we can dump it from asok commands.

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit ba7d1f5f2a2e4355155e00173b0d2ade562a8945)

mds: make LockType fields const

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit 9858e01abd57458b04f87fcaeb432ab7edb6ac07)

mds: annotate mdr with try_rdlock_snap_layout failure

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit 6fa38e4e4ff230ca2a1c6d2de63d561d7cfee826)

mds: refactor if into switch

To make it easier to read and match other constructs in this code.

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit d74d3aadd08ad37a74bf3a9ade18d0b6670c92a0)

mds: call Locker method using this

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit 401e6f49fc2e1c725c8e89e23cdc8afcf5f92876)

mds: simplify assert

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit 180979d7014a6c3799bb7282d90d6e6af860978a)

mds: dump locks passed to Locker::acquire_locks

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit 3057ec9572dd2c1805db09ffd5eaac8ab40a45b2)

mds: add LockOp::print method for debugging

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit 8ddf631886ce4ac34f74e9e62eacc2c33c6e445f)

mds: use new insert template via print

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit d1b066cdbd1ec65b620996cb7f4da7006dfcb7d0)

mds: add request result to mutation for analysis by tests

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit d1765e621611fef5517c2f2227d7786ae41e2001)

mds: add comment on locking order rules

From a past commit.

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit 32a544f9a71aa68547db5bb0fdd517cafbf7d928)

mds: allow specifying rdlock position

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit 944d00ca1fcafb39e45cb431c63f00e71a7976df)

mds: remove dead method

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit ed5987883a411d090d1820ef00453e8147bc1753)

common: provide a template for object dumps

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit 92df0fcc649345f438d0d7cc6fab658c02060f07)

common: support long running ops without slow warnings

A quiesce op in the MDS may run for an indeterminate amount of time. Warnings
should not be generated.

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit c036b151465a2ecb8bf24943e1030988f7e6c6ce)

common: simplify loop

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit cdef9b6dd0a4d0fe39843c5a8198295d14b1c2e7)

common: add JSONFormatterFile class

To stream JSON to a file for collection (for testing).  The idea here that
trying to send a very large JSON result over `ceph tell` is best to be avoided.

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit 8928a1266fc6bc492dac34253edb36c045e491e5)

common: use more efficient vector for stack

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit 770ab517c2b05045c09af3d0a56b6446bd20bb58)

include: use larger int for large gathers

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit 58923668320ba373312c68a8869a1089da41d804)

Merge PR #56202 into squid

* refs/pull/56202/head:
mds/quiesce-db: keep the db thread alive until shutdown
mds/quiesce-db: incorporate review comments
mds/quiesce: declare QuiesceDbPeerListing and QuiesceDbPeerAck
mds/quiesce: resolve the quiesce cluster at the mds monitor
include/types: add an I/O helper for std::unordered_map
messages: avoid using mutable members in MMDSQuiesce*
mds/quiesce-db: incorporate review comments
doc/cephfs/fs-volumes: doc fixes and updates
pybind/mgr: correct type hints for `get_quiesce_leader_info`
mds/quiesce: only use ACTIVE daemons for the quiesce cluster
mds,messages: quiesce db inter-rank messaging
mds/quiesce: MDSRankQuiesce - integration of the quiesce db manager
doc/cephfs/fs-volumes: Add info about the quiesce command
doc: fixes for local dev builds
mgr/volumes: support for `fs subvolume quiesce`
mgr/volumes: use `volume_exception_to_retval` as a decorator
pybind/mgr: add a `one-shot` parameter to send_command
mds/quiesce: QuiesceAgent implementation and unit tests
mds/quiesce: QuiesceDb.h and QuiesceDbManager with tests
common/Timer.cc: improve debug messages from the timer_thread
mds: MDSRank.cc: return status from `send_message_mds`
encoding: add emplace variants for map dencoders
common/Cond: make C_SaferCond private members protected to facilitate inheritance
qa/tasks/cephfs: give the tests more time to run heavy fs workloads

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

Merge pull request #56365 from rhcs-dashboard/wip-65037-squid

squid: mgr/dashboard: fix M retention frequency display

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

Merge pull request #56371 from rhcs-dashboard/wip-65047-squid

squid: mgr/dashboard: fix retention add for subvolume

Reviewed-by: afreen23 <NOT@FOUND>

mgr/dashboard: fix retention add for subvolume

- Added parameters for subvolume and subvolume group when adding a new
snap schedule.
- Added call to remove retention policies when removing a snap schedule
  in case it is the last one with same path

Fixes: https://tracker.ceph.com/issues/64524
Signed-off-by: Ivo Almeida <ialmeida@redhat.com>
(cherry picked from commit 80e1207f4b536fe6edbc81e61cbf951e135eba54)

mgr/dashboard: fix M retention frequency display

Fix snap schedule list monthly retention frequency display

Fixes: https://tracker.ceph.com/issues/64982
Signed-off-by: Ivo Almeida <ialmeida@redhat.com>
(cherry picked from commit 0bd8d3b6bc8ecbe99e22545549a63de3444e5f75)

Merge pull request #56339 from zdover23/wip-doc-2024-03-20-backport-56333-to-squid

squid: doc/cephfs: fix architecture link to correct relative path

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

Merge pull request #56267 from zdover23/wip-doc-2024-08-19-backport-56240-to-squid

squid: doc/rados: update common.rst

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

Merge pull request #56210 from joscollin/wip-64918-squid

squid: qa: enhance labeled perf counters test for cephfs-mirror

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

Merge pull request #56051 from lxbsz/wip-64585

squid: mds: allow lock state to be LOCK_MIX_SYNC in replica for filelock

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

Merge pull request #56030 from vshankar/wip-64617-squid

squid: mds: check file layout in mknod

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

Merge pull request #56017 from vshankar/wip-64758-squid

squid: mds: relax certain asserts in mdlog replay thread

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

Merge pull request #55997 from dparmar18/wip-64741-squid

squid: client: do not proceed with I/O if filehandle is invalid

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

Merge pull request #55989 from joscollin/wip-64744-squid

squid: mds: fix `dump dir` indicate dir is not cached

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

Merge pull request #55940 from neesingh-rh/wip-64701-squid

squid: mgr/volumes: support to reject CephFS clones if cloner threads are not available

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

Merge pull request #55830 from kotreshhr/wip-64583-squid

squid: qa: Fix fs/full suite

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

Merge pull request #55805 from anoopcs9/wip-64566-squid

squid: client: Fix return in removexattr for xattrs from `system.` namespace

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

doc/cephfs: fix architecture link to correct relative path

Signed-off-by: molpako <35787116+molpako@users.noreply.github.com>
(cherry picked from commit ceeb1ca4b38dc307e493b04a59d5d07fed23971f)

Merge pull request #56253 from rhcs-dashboard/wip-64929-squid

squid: mgr/dashboard: ceph authenticate user from fs

Reviewed-by: afreen23 <NOT@FOUND>
Reviewed-by: Nizamudeen A <nia@redhat.com>

Merge pull request #56229 from adk3798/squid-docs-typo-replace-osd

squid: doc: cephadm/services/osd: fix typo

Reviewed-by: Anthony D'Atri <anthonyeleven@users.noreply.github.com>
Reviewed-by: Zac Dover <zac.dover@proton.me>

Merge pull request #55927 from ajarr/wip-64669-squid

squid: qa: add diff-continuous and compare-mirror-image tests to rbd and krbd suites respectively

Reviewed-by: Ilya Dryomov <idryomov@gmail.com>

Merge pull request #56194 from vshankar/wip-64924-squid

squid: mds: disable `defer_client_eviction_on_laggy_osds' by default

Reviewed-by: Dhairya Parmar <dparmar@redhat.com>

qa/suites: add diff-continuous and compare-mirror-image tests

... to rbd and krbd suites respectively.

This allows the compare-mirror-image tests introduced in ea3a567
to be run against various kernel branches, e.g., testing branch.
And allows diff_continuous test in rbd_suite to run against distro
kernel.

Fixes: https://tracker.ceph.com/issues/64574
Signed-off-by: Ramana Raja <rraja@redhat.com>
(cherry picked from commit 92b254138dd43755490392ffcf5ebca04872e116)

qa/suites/rbd: rename nbd folder to device folder

Signed-off-by: Ramana Raja <rraja@redhat.com>
(cherry picked from commit af43f6162464a20625319a2783a104f7032ba21e)

doc/rados: update common.rst

Clarify some overwritten paragraphs in
doc/rados/configuration/common.rst. Cf.
https://github.com/ceph/ceph/pull/50795/files#r1153897231.

Reviewed-by: Anthony D'Atri <anthony.datri@gmail.com>
Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit 233b670704d15c3a69b8d91885b60f3fb101ec58)

Merge pull request #56261 from zdover23/wip-doc-2024-03-19-backport-56247-to-squid

squid: docs/rbd: fix typo in arg name

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

docs/rbd: fix typo in arg name

Replace "{image-}" with "{image-id}" in the "rbd trash rm"
command description.

Signed-off-by: N Balachandran <nibalach@redhat.com>
(cherry picked from commit f3eb489520fd4fae057e61275d16c6c8fd596f3f)

docs/rbd: replace introspect with inspect

Replace "introspect" with "inspect" in the rbd basic commands
description.

Signed-off-by: N Balachandran <nibalach@redhat.com>
(cherry picked from commit ebf2f60f784728c04d8ec59015d666bafcef8218)

docs/rbd: typo in "retrieving image information"

Replace "for the image" with "of the image".

Signed-off-by: N Balachandran <nibalach@redhat.com>
(cherry picked from commit 4fd5c134536d652ae1f9e05ecf52cb81adb3b850)

Merge pull request #56255 from zdover23/wip-doc-2024-03-18-backport-56248-to-squid

squid: doc/rbd: minor changes to the rbd man page

doc/rbd: minor changes to the rbd man page

Fixes typos and grammar for some commands. Adds
additional details for some commandds.

Signed-off-by: N Balachandran <nibalach@redhat.com>
(cherry picked from commit 5dcff6a4b8d835fc55e454af977dc5ebad99d37f)

Merge pull request #56251 from guits/wip-64931-squid

squid: node-proxy: fix RedFishClient.logout() method

mgr/dashboard: ceph authenticate user from fs

Fixes: https://tracker.ceph.com/issues/64660
Signed-off-by: Pedro Gonzalez Gomez <pegonzal@redhat.com>
(cherry picked from commit 19a9f79b5f218d8cb757579bf8c0f5d610dde3a8)

node-proxy: support more Location value formats

After some tests, it turns out that depending on the hardware,
the header 'Location' which is returned by the server after logged can be different.
I could notice the following:

either:

Location: scheme://address:port/redfish/v1/SessionService/Session

or

Location: /redfish/v1/SessionService/Session

a previous tracker [1] was opened because I thought only the first one existed, which is wrong.

Fixes: https://tracker.ceph.com/issues/64951
[1] https://tracker.ceph.com/issues/64894

Signed-off-by: Guillaume Abrioux <gabrioux@ibm.com>
(cherry picked from commit d7ccf26983c41344a12f33b2a30fc79b65cc548f)

node-proxy: fix RedFishClient.logout() method

the endpoint passed down to util.query() is wrong:
is passes the full url (scheme://addr:port/path) where it should only
pass the path. The cause is that RedFishClient.login() basically stores
the value of the Location header in `self.location`.

The consequence of this is that it makes the client unable to properly logout.

Fixes: https://tracker.ceph.com/issues/64894
Signed-off-by: Guillaume Abrioux <gabrioux@ibm.com>
(cherry picked from commit b1d828d1d2f31c02f225bb375d915353582d158a)

Merge pull request #56125 from cbodley/wip-64834-squid

squid: RGW - Swift retarget needs bucket set on object

Reviewed-by: J. Eric Ivancich <ivancich@redhat.com>

Merge pull request #56144 from cbodley/wip-64876-squid

squid: rgw_lc:  replace strftime w/fmt and chrono:calendar

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

Merge pull request #56156 from yuvalif/wip-64887-squid

squid: rgw/kafka: set message timeout to 5 seconds

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

Merge pull request #56181 from ivancich/wip-64909-squid

squid: rgw: rgw-restore-bucket-index -- sort uses specified temp dir

Reviewed-by: Michael Kidd <linuxkidd@gmail.com>

Merge pull request #56216 from idryomov/wip-63770-squid

squid: librbd: account for discards that truncate in ObjectListSnapsRequest

Reviewed-by: Ramana Raja <rraja@redhat.com>

Merge pull request #55761 from ajarr/wip-64553-squid

squid: qa: Add tests to validate synced images on rbd-mirror

Reviewed-by: Ilya Dryomov <idryomov@gmail.com>

Merge pull request #56234 from zdover23/wip-doc-2024-03-16-backport-56182-to-squid

squid: doc/glossary: add "librados" entry

doc/glossary: add "librados" entry

Add a "librados" entry to the glossary.

Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit 2a93a8e837a42559f8a81c6fd9274b24f4fdf7f6)

doc: cephadm/services/osd: fix typo

Signed-off-by: Lorenz Bausch <info@lorenzbausch.de>
(cherry picked from commit d513f1ca69d38f0d7eb40a2d68275caed322cc68)

Merge pull request #55715 from cbodley/wip-64541-squid

squid: rgw: RGWSI_SysObj_Cache::remove() invalidates after successful delete

Reviewed-by: J. Eric Ivancich <ivancich@redhat.com>

rgwlc: implement Filter size limit checks

i.e., this implements ObjectSizeGreaterThan and ObjectSizeLessThan
for supported expiration and transition actions.

Fixes: https://tracker.ceph.com/issues/63304
Signed-off-by: Matt Benjamin <mbenjamin@redhat.com>
(cherry picked from commit c9b19be15362c63526e92f4d968f942a40cab150)

rgwlc: implement NewerNoncurrentVersions

Per AWS doc, this value controls "how many noncurrent versions
Amazon S3 will retain." [1]  We understand this to mean, retain
NewerNoncurrentVersions of any object, regardless of expiration.

Removed unused RGWLifecycleConfiguration::has_same_action() (cleanup)

[1] https://docs.aws.amazon.com/AmazonS3/latest/API/API_NoncurrentVersionExpiration.html.

Signed-off-by: Matt Benjamin <mbenjamin@redhat.com>
(cherry picked from commit a4313e77a31c5165254068b53dfac3d277b0a6a8)

Merge pull request #56207 from zdover23/wip-doc-2024-03-15-backport-56188-to-squid

squid: doc/rbd: add clone mapping command

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

test/librbd: add DiffIterateTruncate and DiffIterateWriteAndTruncate

Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
(cherry picked from commit d84485ec534ccf42327242a9141906f4881c030e)

librbd: don't clip expanded diff on truncate in ObjectListSnapsRequest

If the diff was expanded due to LIST_SNAPS_FLAG_WHOLE_OBJECT, clipping
it when handling a truncate is wrong -- when subtracting that interval,
we either split the expanded extent into two or chop off a piece of it.
However the point of LIST_SNAPS_FLAG_WHOLE_OBJECT is to report a single
extent covering the entire object.

Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
(cherry picked from commit 0ebca5fe22d976f1a17f8078f20a47c49a057edb)

librados/snap_set_diff: ignore truncates above size at start

Because currently calc_snap_set_diff() only ever appends to the running
diff, an excessive (either too large or completely bogus) zero extent
is reported in cases where an object is first expanded (with a snapshot
taken at that point) and then truncated but still above the size of the
object as of the starting snapshot.

Fixes: https://tracker.ceph.com/issues/63770
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
(cherry picked from commit 5b11fb314abefea390c5663c33e9e1ace6179974)

Merge pull request #56155 from rhcs-dashboard/wip-64884-squid

squid: mgr/dashboard: fix snap schedule time format

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

mgr/vol: don't import IndexException since it's unused

Introduced-by: 079f722c37ef6cc8bd3cc26c49ae119dd83431f9
Signed-off-by: Rishabh Dave <ridave@redhat.com>
(cherry picked from commit b4c1d552aa4066efdda3493192927cb780693469)

doc/rbd: add clone mapping command

Add a command that explains how to map a formatted clone when the parent
image and the formatted clone have different encryption types.

Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit d34c1879c8886ec7f19c7a93490c4736ae9a6d20)

Merge pull request #56185 from zdover23/wip-doc-2024-03-14-backport-56160-to-squid

squid: doc/rbd: add map information for clone images to rbd-encryption.rst

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

mds/quiesce-db: keep the db thread alive until shutdown

With the change we can now avoid having to join it during the membership update, preventing potential deadlocks

Signed-off-by: Leonid Usov <leonid.usov@ibm.com>
(cherry picked from commit 8b896a9e145796119f0451201dc4d53ddac97db0)

mds/quiesce-db: incorporate review comments

Signed-off-by: Leonid Usov <leonid.usov@ibm.com>
(cherry picked from commit 3e012f7ba5b8871d1bcf74d402c58553e18835dd)

mds/quiesce: declare QuiesceDbPeerListing and QuiesceDbPeerAck

With these dedicated structs we can fully defer to QuiesceDbEncoding
when encoding/decoding quiesce db messages

Signed-off-by: Leonid Usov <leonid.usov@ibm.com>
(cherry picked from commit 205fd3388ef0ed3011bb21384e79b34b6a6611ec)

mds/quiesce: resolve the quiesce cluster at the mds monitor

Signed-off-by: Leonid Usov <leonid.usov@ibm.com>
(cherry picked from commit 759925748e8be355a743f6edd106052d5a026dfd)

include/types: add an I/O helper for std::unordered_map

Signed-off-by: Leonid Usov <leonid.usov@ibm.com>
(cherry picked from commit 42a5fb3b7562dac09dbf4575558820297317965e)

messages: avoid using mutable members in MMDSQuiesce*

Signed-off-by: Leonid Usov <leonid.usov@ibm.com>
(cherry picked from commit 2fbe40e72e8352a3bf47190d7bc8c80bb60eb7bd)

mds/quiesce-db: incorporate review comments

Signed-off-by: Leonid Usov <leonid.usov@ibm.com>
(cherry picked from commit 9846d35a2ca0fc68c0464657616d259b19273b79)

doc/cephfs/fs-volumes: doc fixes and updates

Signed-off-by: Leonid Usov <leonid.usov@ibm.com>
(cherry picked from commit 629ffe1d317558a85d59d09e59f0ff117d71338e)

pybind/mgr: correct type hints for `get_quiesce_leader_info`

Signed-off-by: Leonid Usov <leonid.usov@ibm.com>
(cherry picked from commit 12d687bd4447687f34a43f637f16af72131189da)

mds/quiesce: only use ACTIVE daemons for the quiesce cluster

Signed-off-by: Leonid Usov <leonid.usov@ibm.com>
(cherry picked from commit 7e42824478ea24f588dabfaa8e317a99a0cee8cc)

mds,messages: quiesce db inter-rank messaging

Fixes: https://tracker.ceph.com/issues/63708
Signed-off-by: Leonid Usov <leonid.usov@ibm.com>
(cherry picked from commit 82f3dbc3ef7b577219f8e1cfaeae4009a809029e)

mds/quiesce: MDSRankQuiesce - integration of the quiesce db manager

* create an instance of the QuiesceDbManager in the rank
* update membership with a new mdsmap
* add an admin socket command for sending requests to the manager

Signed-off-by: Leonid Usov <leonid.usov@ibm.com>
(cherry picked from commit edf4bce948477ebd57eaeb9eecdd4beae6a2c6a7)

doc/cephfs/fs-volumes: Add info about the quiesce command

Signed-off-by: Leonid Usov <leonid.usov@ibm.com>
(cherry picked from commit 09e08ac6a43005b09598d936b1faccd15be85067)

doc: fixes for local dev builds

Signed-off-by: Leonid Usov <leonid.usov@ibm.com>
(cherry picked from commit 88fb668938f051b723c518a7faa6aa759bc39829)

mgr/volumes: support for `fs subvolume quiesce`

Signed-off-by: Leonid Usov <leonid.usov@ibm.com>
(cherry picked from commit d151876d5bf34bbd7aff2cd8ce6dd2e1dffbc414)

mgr/volumes: use `volume_exception_to_retval` as a decorator

When used as a decorator, it saves one indented try-catch block inside the decorated method.
This can be applied to most of the methods in the file, subject to a separate refactoring commit

Signed-off-by: Leonid Usov <leonid.usov@ibm.com>
(cherry picked from commit 78afc6136118e14e2cf6a5651bfe11777c9881a4)

pybind/mgr: add a `one-shot` parameter to send_command

with the parameter set, the message won't be held on to when the remote end resets
or fails to reconnect.

Signed-off-by: Leonid Usov <leonid.usov@ibm.com>
(cherry picked from commit 9907efd0132197feed4b7cc4e7c17a7ec9104a47)

mds/quiesce: QuiesceAgent implementation and unit tests

QuiesceAgent is the layer that converts updates from the QuiesceDb
into calls to the QuiesceProtocol APIs, and then sends async acks
back to the db manager following the quiesce protocol events.

Signed-off-by: Leonid Usov <leonid.usov@ibm.com>
(cherry picked from commit 3de0882ad36d4f08f2f171b1ffd263da5a78f00f)