os/bluestore: Fix collect_alerts

In some cases bdev->collect_alerts() was called twice on the same device.
It happened when DB and SLOW are the same device.
Now only called once.

It made no difference when we just set alert string,
but with EXTBLKDEV we append warning message.

Signed-off-by: Adam Kupczyk <akupczyk@ibm.com>
(cherry picked from commit 08510edf9f2ce3d184b123bf31c0b10e72f629fd)

os/bluestore: extblkdev related health warnings

Replaced start error with health warnings:
EXTBLKDEV: plugin xxx not loaded
EXTBLKDEV: plugin xxx used on mkfs, but now uses plugin yyy

Signed-off-by: Adam Kupczyk <akupczyk@ibm.com>
(cherry picked from commit 950cd2c746e4002fb8ab15ef3565fc646e07f285)

extblkdev/fcm: Replace errors with health warning

Now plugin does not assert or fail to load,
but instead raises following health warnings:
EXTBLKDEV: multivolume fcm will not work properly
EXTBLKDEV: failed accessing FCM utilization log
EXTBLKDEV: bdev_enable_discard not enabled - free space will leak

Signed-off-by: Adam Kupczyk <akupczyk@ibm.com>
(cherry picked from commit 0e9478d4a22374e617a93f635fb1fb639918e147)

extblkdev/fcm: Cleanup

Add `override` to virtual interface functions.

Signed-off-by: Adam Kupczyk <akupczyk@ibm.com>
(cherry picked from commit d7086d0d6da94b36d6161884592f186de6c1e8d9)

extblkdev: Fix preload

One execution path did not save preload_result value.
Fixed now.

Signed-off-by: Adam Kupczyk <akupczyk@ibm.com>
(cherry picked from commit bc59cd5acc014b9b301e2cda5231f843197f9ef5)

blk/extblkdev: Health warns from ExtBlkDev plugins

Expanded interface by collect_alerts.
Created category "EBD_PROBLEMS" for ExtBlkDev problems.

Signed-off-by: Adam Kupczyk <akupczyk@ibm.com>
(cherry picked from commit f4beeb2038691e289c853a973ee9dfdff18ace4d)

blk: Expand collect_alerts to allow specialization

Previously we had BlockDevice::collect_alerts that had fixed
implementation.
Expanded BlockDevice::collect_alerts into virtual, so KernelDevice can
override it.

Signed-off-by: Adam Kupczyk <akupczyk@ibm.com>
(cherry picked from commit e952b0c9fd240ad0318c744091bf32201e4ee76c)

blk,os/bluestore: do not call extblkdev::detect_device on every
KernelDevice::open()

Instead we can do that once during mkfs and once again during Bluestore
mounting if relevant label is attached.

Signed-off-by: Igor Fedotov <igor.fedotov@croit.io>
(cherry picked from commit cf0a466c226ae6fef41bbbaf7ad7d3db52263b1b)

extblkdev/fcm: Add plugin identification for fcm

Signed-off-by: Adam Kupczyk <akupczyk@ibm.com>
(cherry picked from commit b157db9cabccaef6ba4393e90515f594b2e76a35)

extblkdev/fcm: Do not enable set_keepcaps by default

When one runs FCM enabled hardware config.set_keepcaps must be set.
However, having it as default is inadvised.

Signed-off-by: Adam Kupczyk <akupczyk@ibm.com>
(cherry picked from commit dfb43c0d4e9fa4013b2a1e0367bee88a706497e5)

extblkdev/fcm: Change douts/derrs

Remove printing fcm values to dout(1).
Upgraded inability to access FCM log from dout to derr.

Signed-off-by: Adam Kupczyk <akupczyk@ibm.com>
(cherry picked from commit 1054d63e52cf5bc9367eb83c5b6becbfe0b38397)

extblkdev/fcm: Create perf counters with fcm stats

Created "extblkdev" section.
Each time plugin is asked to `get_stats()` also write to perf counters.

Signed-off-by: Adam Kupczyk <akupczyk@ibm.com>
(cherry picked from commit 52424420db560750fd3bbb6e1dac550beab0848d)

fixup warning of fcm plugin core

Signed-off-by: Adam Kupczyk <akupczyk@ibm.com>
(cherry picked from commit 2a0e3418d8da4406a13a16dcd39090a21c150582)

extblkdev/fcm: derr when FCM and no block device discard

Add derr message when FCM is in use but discard is not enabled.

Signed-off-by: Adam Kupczyk <akupczyk@ibm.com>
(cherry picked from commit 9c0c40cfa697a905151305b2f79bfe65ca9366c7)

extblkdev/fcm: Refuse to operate on multimedia lvm block devices

BlueStore is selecting were data is put to the device.
Merging 2 FCM devices together means that BlueStore will see free space
on one of the devices, but not know the other is full and asking to put
data there. It will cause -ENOSPC while free space is reported.

Signed-off-by: Adam Kupczyk <akupczyk@ibm.com>
(cherry picked from commit 4f85256062af71bc33554ed5664bb4d6bfd5dfa7)

extblkdev/fcm: Adapt FCM plugin to RHCS9

Device /sys content is different from what original plugin expected.
Adapted to it.

Signed-off-by: Adam Kupczyk <akupczyk@ibm.com>
(cherry picked from commit b703ac411809a58ef8666e2c8f95d8af83f4ebcf)

added FCM plugin

Signed-off-by: Martin Ohmacht <mohmacht@us.ibm.com>
(cherry picked from commit b0174c013dcfb7c076231da4cd9d8e6659beddea)

os/bluestore: Make extblkdev missing a warning

When mounting BlueStore for OSD problem with extblkdev meta are fatal error.
When mounting BlueStore in other cases extblkdev are just derr messages.

Signed-off-by: Adam Kupczyk <akupczyk@ibm.com>
(cherry picked from commit f38ef0e652c9c00f23091ddb80c623d258104bef)

extblkdev: Make preload one-time function

Subsequent calls to extblkdev::preload() do not do anything.
It is assumed that first call already loads all plugins that could be
loaded.

Signed-off-by: Adam Kupczyk <akupczyk@ibm.com>
(cherry picked from commit 358047f3cc3b4382899c636b8cdc67dcd303124b)

os/bluestore: Add config bluestore_use_ebd

When EBD(extblkdev) plugin is in use usually it needs to present all the time.
For bluestore deployed with EBD plugin it makes it an error if bluestore tries
to mount and EBD plugin is not present.

Preload of extblkdev plugins was misplaced.
Moved loading plugins into BlueStore.
This way both OSD and tools can load plugins.

Plugins are now loaded only:
- before mkfs
- when extblkdev plugin is signalled in label meta

Signed-off-by: Adam Kupczyk <akupczyk@ibm.com>
(cherry picked from commit addb8388637737c79099bf96e0fe770d6f0dbbf8)

extblkdev: Expand interface, add plugin id retrieval

Plugin identification is a tool to implement two features:
1) Check the plugin is initialized
2) Verify that the plugin initialzied is the proper one

Signed-off-by: Adam Kupczyk <akupczyk@ibm.com>
(cherry picked from commit f4a002a06492e98c1d4468e7d202b249c5d7ea1a)

Merge PR #68347 into tentacle

* refs/pull/68347/head:
nvmeofgw: propagate quorum feature to the NVMeofMonClient,
fix upgrade
code review changes
nvmeofgw: disaster set/clear command, introduced disaster-locations map,
 nvmeofgw: added support to nvmeof stretched cluster:
nvmeofgw: prevent map corruption while processing beacons from deleted gws
mon: add NVMEOF_BEACON_DIFF to mon_feature_t and mon CompatSet
nvmeofgw: beacon diff implementation in the monitor and in the MonClient.

Reviewed-by: Alexander Indenbaum <aindenba@redhat.com>
Reviewed-by: Patrick Donnelly <pdonnell@ibm.com>
Reviewed-by: Radoslaw Zarzynski <rzarzyns@redhat.com>

Merge PR #67513 into tentacle

* refs/pull/67513/head:
neorados: Fix Neorados CephContext leak

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

Merge PR #67508 into tentacle

* refs/pull/67508/head:
qa: suppress MismatchedFree operator delete RocksDB

Reviewed-by: Yaarit Hatuka <yaarithatuka@gmail.com>

Merge PR #67321 into tentacle

* refs/pull/67321/head:
qa: set column for insertion
qa: bail sqlite3 on any error
qa: use actual sqlite3 blob instead of string
test: use json_extract instead of awkward json_tree

Reviewed-by: Yaarit Hatuka <yaarithatuka@gmail.com>

Merge PR #68651 into tentacle

* refs/pull/68651/head:
rgw: read_obj_policy() consults s3:prefix when deciding between 403/404

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

Merge PR #68524 into tentacle

* refs/pull/68524/head:
qa: Leak_StillReachable RocksDB error_handler

Reviewed-by: Yaarit Hatuka <yaarithatuka@gmail.com>

Merge PR #68518 into tentacle

* refs/pull/68518/head:
pybind/mgr: call new _ceph_exit for killpoints
mgr: add ceph binding for exit

Reviewed-by: Yaarit Hatuka <yaarithatuka@gmail.com>

Merge PR #68506 into tentacle

* refs/pull/68506/head:
rgw/beast: use strand executor for timeout timer to prevent concurrent socket access

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

Merge PR #68463 into tentacle

* refs/pull/68463/head:
librbd: tweak ReadResult's handler for SparseBufferlist type
librbd: avoid losing sparseness in read_parent()

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

Merge PR #68371 into tentacle

* refs/pull/68371/head:
qa/tasks/pykmip: archive pykmip log after server down
qa/tasks/pykmip: use OpenSSL names instead IANA
qa/tasks/pykmip: drop py2 deps
Revert "qa/rgw/crypt: disable failing kmip testing"

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

Merge PR #66358 into tentacle

* refs/pull/66358/head:
rgw/auth: a forwarded CreateBucket request in case of multisite has an empty
rgw/s3: Always include x-amz-content-sha256 header in AWS v4 signatures for S3 compatibility

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

Merge PR #68583 into tentacle

* refs/pull/68583/head:
mgr/dashboard: Difference in "path" value observed when rgw user level export created via dashboard vs cli

Reviewed-by: Afreen Misbah <afreen@ibm.com>

Merge PR #68475 into tentacle

* refs/pull/68475/head:
mgr/dashboard : Restrict create storage class with existing name

Reviewed-by: Afreen Misbah <afreen@ibm.com>

Merge PR #68369 into tentacle

* refs/pull/68369/head:
found duplicate series for the match group {fs_id="-1"}

Reviewed-by: Ankush Behl <cloudbehl@gmail.com>

Merge PR #67949 into tentacle

* refs/pull/67949/head:
mgr/dashboard: [storage-class]: Deleting local storage class from UI does not remove its entry from zone

Reviewed-by: Afreen Misbah <afreen@ibm.com>

Merge PR #67912 into tentacle

* refs/pull/67912/head:
mgr/dashboard: Add restore events in notification screen

Reviewed-by: Pedro Gonzalez Gomez <pegonzal@redhat.com>

Merge PR #67871 into tentacle

* refs/pull/67871/head:
tentacle: mgr/dashboard: add helper text to bucket form > policy and other spacing fixes

Reviewed-by: Pedro Gonzalez Gomez <pegonzal@redhat.com>

Merge PR #68604 into tentacle

* refs/pull/68604/head:
mr/dashboard: remove rgw_servers filter from radosgw-sync-overview grafana dashboard

Reviewed-by: Afreen Misbah <afreen@ibm.com>

Merge PR #68546 into tentacle

* refs/pull/68546/head:
mgr/dashboard: mgr/dashboard: Carbonize Realm Name and Token block in Multi-site Replication Wizard

Reviewed-by: Afreen Misbah <afreen@ibm.com>

Merge PR #68542 into tentacle

* refs/pull/68542/head:
mgr/dashboard: [snap-visibility]Edit Client config option remains stuck in loading when nfs user is configured

Reviewed-by: Afreen Misbah <afreen@ibm.com>

Merge PR #68520 into tentacle

* refs/pull/68520/head:
mgr/dashboard : fix-non-versioning-bucket-issue

Reviewed-by: Afreen Misbah <afreen@ibm.com>

Merge PR #68513 into tentacle

* refs/pull/68513/head:
mgr/dashboard: Option to select archive option while Import Multi-site Token

Reviewed-by: Afreen Misbah <afreen@ibm.com>

Merge PR #68512 into tentacle

* refs/pull/68512/head:
mgr/dashboard: sync policy created for a bucket in Object >> Multi-site >> Sync-policy, is not reflecting under bucket's replication

Reviewed-by: Afreen Misbah <afreen@ibm.com>

Merge PR #67840 into tentacle

* refs/pull/67840/head:
mgr/dashboard: Fix make check failures
mgr/dashboard: Round off y-axis value of area chart
mgr/dashboard: Fix padding of overview page
mgr/dashboard: Add capacity thresholds
mgr/dashboard: Fix loading states in storage overview card
mgr/dashboard: Add tootltip to storage overview
mgr/dashboard: Fixing message when prometheus is disabled in performance charts
mgr/dashboard: show miscellaneous data used
mgr/dashboard: fix consumption chart units
mgr/dashboard: rename expand-cluster to add-storage
mgr/dashboard: update onboarding screen as per design
mgr/dashboard: Fix scrubbing state
mgr/dashboard: Fix snapshot Api firing twice
mgr/dashboard: Add data resileincy panel
mgr/dashboard: Add data resileincy card
mgr/dashboard: add storage consumption card
mgr/dashboard : update telemetry notification for simple mode
mgr/dashboard:revamp on-borading screen
mgr/dashboard: Generic Performace Chart - Carbon
mgr/dashboard: Add filtering of alerts via route
mgr/dashboard: Add skeleton states for alerts card
mgr/dashboard: Fix css in alerts card
mgr/dashboard: Fix breaking layout in overview page
mgr/dashboard: Add hardware tab to health card
mgr/dashboard: Added variations of alerts card sub total layout
mgr/dashboard: Css fixes for health card and alerts card
fix for quorum in API
mgr/dashboard: Add systems tab to health card
mgr/dashboard: Add alerts card
mgr/dashboard: Add health check panel
mgr/dashboard: Add health card
mgr/dashboard: side-panel enhancements
mgr/dashboard: introduce side panel as a reusable component
mgr/dashboard: Removed Raw capacity toggle
mgr/dashboard: Added unit tests
Added qurey data
mgr/dashboard: Added tool definition tip
Added query tital and used capacity data
mgr/dashboard: Add storage card to overview page

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

nvmeofgw: propagate quorum feature to the NVMeofMonClient,
          reverted feature bit NVMEOF_BEACON_DIFF:

   -NVMeofGwMon adds a quorum_features indication to the MonClient map.
   -MonClient initially sends beacons without applying the BEACON_DIFF logic.
   -MonClient begins applying the BEACON_DIFF logic only when the BEACON_DIFF bit
    is set in the quorum_features field of the NVMeoF monitor map.
   -added mon commands:
     nvme-gw set beacon-diff disable
     nvme-gw set beacon-diff enable
   -performed changes in encode/decode of the BEACON_DIFF feature
   -reverted NVMEOF_BEACON_DIFF bit

Signed-off-by: Leonid Chernin <leonidc@il.ibm.com>
(cherry picked from commit f3f8bde10c86e0c5d4f5286f8e249fc26fa3605d)

Conflicts:
 ../src/include/ceph_features.h
 conflict because UMBRELLA feature  was in main and should not be in tentacle

fix upgrade
 remove invalid code from NVMeofGwSerialize.h
 put encode/decode at the file NVmeofGwMap.h

Signed-off-by: Leonid Chernin <leonidc@il.ibm.com>
(cherry picked from commit c319e6a2770f06165a18ffc9170de80f1c6d1967)

code review changes

Signed-off-by: Leonid Chernin <leonidc@il.ibm.com>
(cherry picked from commit 380ebf0087aa90f1417b5f99384e296135a84fd1)

nvmeofgw: disaster set/clear command, introduced disaster-locations map,
          changes in  behavior of failover/failback, additional
          validations added

Signed-off-by: Leonid Chernin <leonidc@il.ibm.com>
(cherry picked from commit 67f3a7f594dd048622bca223970e87ddae7ddf2d)

 nvmeofgw: added support to nvmeof stretched cluster:
  GW commands added : set location and set admin state enable/disable
  added start-failback <location> command.

  failover logic is impacted by GW location
  implement GW admin commands  enable/disable
  added map  for location-failback-in-progress
  failback between locations happens only by monitor command
  implemented new ana-group relocation process used
  when inter-location failback command sent
  added upgrade rules

fixes: https://tracker.ceph.com/issues/74210

Signed-off-by: Leonid Chernin <leonidc@il.ibm.com>
(cherry picked from commit 315da540478d201b59f35060604e5ab6f9e6f7bb)

nvmeofgw: prevent map corruption while processing beacons from deleted gws

Fix race issue of map corruption when deleted gw sends beacons
but this gw data was removed from pending map and still exists in map.
Process beacons only if GW's data exists in both maps:
main-map and pending-map, otherwise just ignore beacons.

fixes: https://tracker.ceph.com/issues/74160

Signed-off-by: Leonid Chernin <leonidc@il.ibm.com>
(cherry picked from commit 3636dc3598b7022bff623d686cbee83ed288d966)

mon: add NVMEOF_BEACON_DIFF to mon_feature_t and mon CompatSet

NOPE NOPE
In order for the client to safely send BEACON_DIFF messages, it
needs to be the case that the leader at the time of receipt will
support BEACON_DIFF.

Simply using the connection features for the MonClient's target mon is
insufficient, because it might be a peon.  If the peon supports
BEACON_DIFF and the leader does not the leader will either crash or
interpret it as a full BEACON.  Neither outcome is acceptable.

Instead, we need to wire up a feature bit to the MonMap mon_feature_t
members and the CompatSet.

Adding FEATURE_BEACON_DIFF to ceph::features::mon get_supported()
and get_persistent() ensures that once all monitors in the quorum
support it, MonMap::get_required_features() will include it.
See Elector::propose_to_peers, Monitor::(win|lose)_election,
MonmapMonitor::apply_mon_features.

Once FEATURE_BEACON_DIFF is present in MonMap::get_required_features():
- Monitor::apply_monmap_to_compatset_features() will prevent
  downgrades of the monitors by updating the CompatSet to include
  CEPH_MON_FEATURE_INCOMPAT_NVMEOF_BEACON_DIFF
- Monitor::calc_quorum_requirements() will set
  Monitor::required_features to require the NVMEOF_BEACON_DIFF
  for any monitor peers.
- MonClient::get_monmap_required_features() will eventually include
  ceph::features::mon::FEATURE_NVMEOF_BEACON_DIFF.

Signed-off-by: Samuel Just <sjust@redhat.com>
(cherry picked from commit e55ad7bce2fb85096cd31ff9846403f9dbd01e85)

nvmeofgw: beacon diff implementation in the monitor and in the MonClient.

     -monclient encodes subsystems by beacon-diff rules if BEACON_DIFF
      bit is enabled by quorum
     -monitor processes beacons by beacon-diff new schema
     -monitor detects sequence out of order(ooo) condition and handles it
     -in case ooo detected monitor send ack to the gw with the expected correct sequence
     -monitor skips failovers for some interval when ooo detected
     -monitor ignores all becons with incorrect sequences until gw sends expected one
     -coding upgrade rules

Signed-off-by: Leonid Chernin <leonidc@il.ibm.com>
Fixes: https://tracker.ceph.com/issues/72394
(cherry picked from commit 3555a28e45c5b44289f12abe2fc843e21c7ebf87)

Conflicts:
  src/mon/NVMeofGwMon.h
  check_beacon_timeout() was in tentacle, and in main, but wasn't a part of this commit.
  Conflict resolution - this should be there because it is required for fast failover

  src/test/CMakeLists.txt
  the problem was that there is an empty line at the end of the file in main, but not in tentacle

Merge PR #68387 into tentacle

* refs/pull/68387/head:
mgr/cephadm: fix mgmt-gateway startup on IPv6 VIP

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

Merge PR #68357 into tentacle

* refs/pull/68357/head:
qa/suites/upgrade/telemetry-upgrade: ignore expected health warning
qa/suites/orch/cephadm: replace "reef" with "v18.2.8"
qa/suites/fs/upgrade/mds_upgrade_sequence: replace "reef" with "v18.2.8"
qa/suites/upgrade: use tagged versions of reef

Reviewed-by: Patrick Donnelly <pdonnell@ibm.com>
Reviewed-by: Redouane Kachach <rkachach@redhat.com>
Reviewed-by: Nitzan Mordechai <nmordech@redhat.com>

Merge PR #68149 into tentacle

* refs/pull/68149/head:
qa: Add nvmeof upgrade from v20.2.0

Reviewed-by: Aviv Caro <Aviv.Caro@ibm.com>

Merge PR #68238 into tentacle

* refs/pull/68238/head:
mgr/dashboard: Rename Alert breadcrumb to Alert Rules

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

Merge PR #68086 into tentacle

* refs/pull/68086/head:
mgr/cephadm: Add KMIP server support for NVMeoF gateway

Reviewed-by: Redouane Kachach <rkachach@redhat.com>
Reviewed-by: Tomer Haskalovitch <tomer.haska@ibm.com>

Merge PR #67799 into tentacle

* refs/pull/67799/head:
mgr/dashboard: Add secure and verify-host-name to "listener add" on NVMeoF CLI.

Reviewed-by: Aviv Caro <Aviv.Caro@ibm.com>
Reviewed-by: Tomer Haskalovitch <tomer.haska@ibm.com>

Merge PR #67637 into tentacle

* refs/pull/67637/head:
mgr/dashboard: Misleading warning when no eligible devices are available for OSD creation

Reviewed-by: Afreen Misbah <afreen@ibm.com>

mgr/dashboard: Fix make check failures

- styleints, yaml was not rebased, and overview test failures

Signed-off-by: Afreen Misbah <afreen@ibm.com>

rgw: read_obj_policy() consults s3:prefix when deciding between 403/404

when read_obj_policy() gets ENOENT, it only returns 404 NoSuchKey if the
requester has s3:ListBucket permission. however, policy that allows
s3:ListBucket may be conditional on the s3:prefix to restrict listings
to certain paths/object names. add the requested object name to the iam
environment as s3:prefix to match aws behavior here

Fixes: https://tracker.ceph.com/issues/74398
Signed-off-by: Casey Bodley <cbodley@redhat.com>
(cherry picked from commit 363a81ca8d0bf2f1e84b2d48aa02be40d5398147)

Merge PR #68345 into tentacle

* refs/pull/68345/head:
mgr/dashboard: Add location to gateway info command in NVMeoF CLI

Reviewed-by: Tomer Haskalovitch <tomer.haska@ibm.com>

Merge PR #68148 into tentacle

* refs/pull/68148/head:
qa: ignore NVMEOF_GATEWAY_DOWN in nvmeof_scalability.yaml
qa/tasks/nvmeof.py: retry do_check if gw in CREATED
qa/tasks/nvmeof.py: Fix tharsher daemon_rm revival

Reviewed-by: Aviv Caro <Aviv.Caro@ibm.com>

Merge PR #68038 into tentacle

* refs/pull/68038/head:
librbd: store CRC32C with initial value -1 to match msgr2 validation
librbd: add rbd_aio_write_with_crc32c API for precomputed checksums

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

Merge PR #67600 into tentacle

* refs/pull/67600/head:
mgr/dashboard: Add DHCHAP controller key to NVME host commands

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

Merge PR #67526 into tentacle

* refs/pull/67526/head:
qa: Fix coredumps caused by udisks

Reviewed-by: Patrick Donnelly <pdonnell@ibm.com>
Reviewed-by: Aviv Caro <Aviv.Caro@ibm.com>

Merge PR #67470 into tentacle

* refs/pull/67470/head:
mgr/dashboard: Adding rados ns option into add_ns_req

Reviewed-by: Tomer Haskalovitch <tomer.haska@ibm.com>
Reviewed-by: Aviv Caro <Aviv.Caro@ibm.com>

Merge PR #67326 into tentacle

* refs/pull/67326/head:
mgr/dashboard: bump nvmeof submodule to 1.7.2
mgr/dashboard: bump nvmeof submodule to 1.6.7
mgr/dashboard: bump nvmeof submodule to 1.6.5

Reviewed-by: Afreen Misbah <afreen@ibm.com>
Reviewed-by: Nizamudeen A <nia@redhat.com>

Merge PR #66933 into tentacle

* refs/pull/66933/head:
mgr/cephadm: add rbd_with_crc32c parameter to nvmeof service spec

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

Merge PR #68431 into tentacle

* refs/pull/68431/head:
qa/valgrind: generalize and group MismatchedFree suppressions
qa/valgrind: generalize suppressions for gcc-14 MismatchedFree
qa: Leak_StillReachable RocksDB error_handler
qa: suppress false positive delete map mismatch errors

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

Merge PR #68399 into tentacle

* refs/pull/68399/head:
ceph_mon: Fix shutdown order to destroy Monitor before closing mon store

Reviewed-by: Kamoltat Sirivadhna <ksirivad@redhat.com>

Merge PR #64692 into tentacle

* refs/pull/64692/head:
mgr/cephadm: fixing generate_cert to pick the right root subject
mgr/cephadm: include cluster FSID in root CA Common Name (CN)

Reviewed-by: Guillaume Abrioux <gabrioux@redhat.com>

Merge PR #66027 into tentacle

* refs/pull/66027/head:
Check if `HTTP_X_AMZ_COPY_SOURCE` header is empty

Reviewed-by: Yuval Lifshitz <ylifshit@redhat.com>

Merge PR #67573 into tentacle

* refs/pull/67573/head:
rgw/lc: Do not delete DM if its at end of pagination list.

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

Merge PR #67349 into tentacle

* refs/pull/67349/head:
kv/RocksDB: Add instrumentation to BinnedLRUCache

Reviewed-by: Igor Fedotov <ifedotov@suse.com>
Reviewed-by: Jaya Prakash Madaka <jayaprakash@ibm.com>

mr/dashboard: remove rgw_servers filter from radosgw-sync-overview grafana dashboard

Fixes: https://tracker.ceph.com/issues/76239
Signed-off-by: Aashish Sharma <aasharma@redhat.com>
(cherry picked from commit b0601df2a55e3fc56370687b27693779ad356be5)

mgr/dashboard: Difference in "path" value observed when rgw user level export created via dashboard vs cli

Fixes: https://tracker.ceph.com/issues/75425
Signed-off-by: Dnyaneshwari Talwekar <dtalweka@redhat.com>
(cherry picked from commit 0efa59c305ca4c0f4488c27c0f946c7afbf3d5f6)

Merge PR #68316 into tentacle

* refs/pull/68316/head:
qa: allow multiple mgr sessions during eviction test

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

Merge PR #68315 into tentacle

* refs/pull/68315/head:
qa: resolve py3.12 regression for random.sample

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

Merge PR #68314 into tentacle

* refs/pull/68314/head:
qa/cephfs: do not validate error string in "fs authorize" tests
mon/AuthMonitor: add osd w cap for superuser client

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

Merge PR #68313 into tentacle

* refs/pull/68313/head:
mds: use SimpleLock::WAIT_ALL for wait mask

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

Merge PR #67808 into tentacle

* refs/pull/67808/head:
mds: scrub pins more inodes than the mds_cache_memory_limit

Reviewed-by: Kotresh Hiremath Ravishankar <khiremat@redhat.com>
Reviewed-by: Igor Fedotov <ifedotov@suse.com>

Merge PR #68169 into tentacle

* refs/pull/68169/head:
qa/workunits: Add updated kernel archive URL

Reviewed-by: Kotresh Hiremath Ravishankar <khiremat@redhat.com>
Reviewed-by: Ilya Dryomov <idryomov@redhat.com>

Merge PR #68439 into tentacle

* refs/pull/68439/head:
mds: add ref counting to LogSegment

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

Merge PR #67341 into tentacle

* refs/pull/67341/head:
ceph-volume: skip redundant NVMe mkfs discards

Reviewed-by: Teoman Onay <tonay@ibm.com>

ceph-volume: skip redundant NVMe mkfs discards

- Avoid redundant discard during mkfs when discard is disabled
- Reduces mkfs time on large NVMe devices by skipping long running discard operations

Fixes: https://tracker.ceph.com/issues/74908
Signed-off-by: Ujjawal Anand <ujjawal.anand@ibm.com>
(cherry picked from commit daebcfb8944789a47da20ffec3e03a5b2737f711)

Merge PR #68379 into tentacle

* refs/pull/68379/head:
cephadm: wait for latest osd map after ceph-volume before OSD deploy

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

Merge PR #68286 into tentacle

* refs/pull/68286/head:
ceph-volume: skip mkfs discard for LVM NVMe OSDs

Reviewed-by: Teoman Onay <tonay@ibm.com>

Merge PR #68121 into tentacle

* refs/pull/68121/head:
orch/cephadm: fix osd.default creation

Reviewed-by: Teoman Onay <tonay@ibm.com>

Merge PR #68108 into tentacle

* refs/pull/68108/head:
ceph-volume: skip virtual cdrom devices in inventory

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

Merge PR #67989 into tentacle

* refs/pull/67989/head:
ceph-volume: include LVM mapper devices in get_devices()

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

Merge PR #67916 into tentacle

* refs/pull/67916/head:
src/ceph-volume: fast device unavailable as error

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

Merge PR #67418 into tentacle

* refs/pull/67418/head:
mgr/cephadm: validate hostname in NodeProxyCache
node-proxy: improve HTTP error logging in client
node-proxy: get serial number instead of SKU
node-proxy: allow multiple sources per component
node-proxy: re-auth and retry once on 401
node-proxy: fix flake8 E721 in _dict_diff
node-proxy: make the update loop interval configurable
mgr/node-proxy: fix "ceph orch hardware status --category criticals"
node-proxy: normalize storage data per member
node-proxy: encapsulate send logic in dedicated method
node-proxy: log actual data delta in reporter
node-proxy: add periodic heartbeats in main and reporter loops
node-proxy: adjust log levels
node-proxy: add unit tests
node-proxy: add tox config for mypy, flake8, isort, black
node-proxy: black and isort formatting pass
node-proxy: fix mypy errors
node-proxy: handle nested Redfish paths for components
node-proxy: split out config, bootstrap and redfish logic
node-proxy: refactor config loading
node-proxy: add 'vendor based' redfish system selection
node-proxy: introduce component spec registry and overrides for updates
mgr/cephadm: safe status/health access in node-proxy agent and inventory
node-proxy: narrow build_data exception handling and re-raise
node-proxy: refactor Endpoint/EndpointMgr and fix chassis paths
node-proxy: use safe field access in storage update
node-proxy: reduce log verbosity for missing optional fields

Reviewed-by: Teoman Onay <tonay@ibm.com>

Merge PR #67343 into tentacle

* refs/pull/67343/head:
ceph-volume: fix test_reject_readonly_device unit test
ceph-volume: single lvs call to speed up exclude_lvm_osd_devices
ceph-volume: avoid Device() instantiation in lvm OSD filtering
ceph-volume: avoid RuntimeError on ceph-volume raw list with non-existent loop devices

Reviewed-by: Teoman Onay <tonay@ibm.com>

Merge PR #68344 into tentacle

* refs/pull/68344/head:
test/rgw: remove depracated boto dependency

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

Merge PR #67850 into tentacle

* refs/pull/67850/head:
mgr, qa: clarify module checks in DaemonServer
mgr, qa: add `pending_modules` to asock command
mgr, common, qa, doc: issue health error after max expiration is exceeded
mgr: ensure that all modules have started before advertising active mgr

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

mgr/dashboard: mgr/dashboard: Carbonize Realm Name and Token block in Multi-site Replication Wizard

Fixes: https://tracker.ceph.com/issues/76085
Signed-off-by: Sagar Gopale <sagar.gopale@ibm.com>
(cherry picked from commit 5155b2125db002c21be1b014ccd6389b7c7b386e)

mgr/dashboard: Round off y-axis value of area chart

- by default y-axos set to 1 for all
- the value round off for area chart is seperated from y-axis ticks
- also fixes a bug where all IOPS y-ticks being repeated 1,1,0,0

The following values are set for now:

IOPS: valueDecimals=0, axisDecimals=1
Latency: valueDecimals=2, axisDecimals=1
Throughput: valueDecimals=2, axisDecimals=1
Consumption: valueDecimals=2, axisDecimals=1

Fixes https://tracker.ceph.com/issues/76191

Signed-off-by: Afreen Misbah <afreen@ibm.com>
(cherry picked from commit a6012c3954830f4b60ef49218d40d4e8ed3bd14c)

mgr/dashboard: [snap-visibility]Edit Client config option remains stuck in loading when nfs user is configured

Fixes: https://tracker.ceph.com/issues/76200
Signed-off-by: Dnyaneshwari Talwekar <dtalweka@redhat.com>
(cherry picked from commit 820a57fa3255597552bd952bf59cce863347f856)