crimson/os/seastore/transaction_manager: complete indirect mappings
before reading

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

crimson/os/seastore/lba_mapping: treat indirect mappings as valid
non-clone ones

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

crimson/os/seastore/lba_mapping: treat indirect mappings as valid
non-zero-reserved mappings

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

crimson/os/seastore/lba_mapping: add interfaces to determine the pending
state of mappings' child extents

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

crimson/os/seastore: laddr_offset_t can align to mulitple UNIT_SIZE now

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

crimson/os/seastore/lba_mapping: drop the deprecated "duplicate" method

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

crimson/os/seastore/cached_extent: more correctness asserts

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

crimson/os/seastore/transaction_manager: hide
TransactionManager::remap_pin()

It doesn't have public use cases any more.

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

crimson/os/seastore/object_data_handler: LBACursor based clones

Avoiding unnecessary lba tree searches

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

crimson/os/seastore/object_data_handler: cleanup deprecated functions

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

crimson/os/seastore/object_data_handler: LBACursor based overwrite

This should avoid unnecessary lba tree searches in the old
implementation of ObjectDataHandler::overwrite()

Overwrites of ObjectDataBlocks are dealt with by first punching holes
in the lba tree and then inserting new extents in the holes.

Specifically, overwrites are classified into two categories:
1. the range of the overwrite falls in a single lba mapping;
2. the range of the overwrite crosses multiple lba mappings.

For the first category, ObjectDataHandler processes the overwrites in
the following way:
1. if the mapping is a pending one (corresponds to a pending extent),
   merge the overwrite with the data of the pending extent;
2. otherwise, if the overwrite can, apply it with delta based
   overwrites;
3. otherwise, punch a hole in the mapping, insert a new extent with the
   data of the overwrite.

For the second category, the overwrite is processed as follows:
1. if the left boundary of the overwrite is inside an existing mapping,
   deal with the mapping in a way similar to the single-mapping
   overwrites;
2. remove all lba mappings that are strictly within the range of the
   overwrite;
3. deal with the right boundary of the overwrite in the same way as the
   left boundary.

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

crimson/os/seastore/lba_mapping: make LBAMapping duplicate
light-weighted and safe

This commit removes LBAMapping::child_pos, forces TransactioManager
methods to link children directly through child_pos_t, so that
LBAMapping::duplicate() can be a shallow one.

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

common: remove seastore_obj_data_write_amplification

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

crimson/os/seastore/btree: allow BtreeCursor to represent the end of the
tree

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

crimson/os/seastore/lba_manager: add LBACursor related interfaces

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

crimson/os/seastore/lba: remove refresh_lba_mapping

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

crimson/os/seastore/lba_mapping: add next() to get the next adjacent
mapping

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

crimson/os/seastore/lba_manager: replace refresh_lba_cursor() with
cursor.refresh()

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

crimson/os/seastore/lba_mapping.h: allow null direct_cursor in indirect
mappings

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

Merge pull request #64657 from ceph/wip-main-rocky9

build: Rocky 9 support

Merge pull request #64757 from VinayBhaskar-V/wip-unittest-72325

test/rbd-mirror: eliminate a race in ResyncRequestedRemoteNotPrimary

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

Merge pull request #64750 from bluikko/doc-bucketops-fixes-plus-radosgw

doc/radosgw: Simple fixes and improvements, links improvements

Merge pull request #64565 from synarete/ss-samba-metrics-image-devbuilds-centos-amd64

mgr/cephadm: use same tag for all samba images

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

Merge pull request #64737 from zdover23/wip-doc-2025-07-29-cephfs-ceph-dokan

doc/cephfs: edit ceph-dokan.rst (2 of x)

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

test/rbd-mirror: eliminate a race in ResyncRequestedRemoteNotPrimary

Adjust the wait_for_notification call in TestMockImageReplayerSnapshotReplayer.ResyncRequestedRemoteNotPrimary
to expect 2 notifications instead of 1. This allows the test to correctly wait for both expected events
i.e for finish_sync() and handle_replay_complete(locker, -EREMOTEIO, "remote image demoted"), ensuring the
replayer transitions to STATE_COMPLETE and is_replaying() returns false as intended.

Fixes: https://tracker.ceph.com/issues/72325
Signed-off-by: VinayBhaskar-V <vvarada@redhat.com>

Merge pull request #64351 from ifed01/wip-ifed-discard-threads-better-lifecycle

blk/kernel: improve DiscardThread life cycle.

Reviewed-by: Yite Gu <guyite@bytedance.com>

Merge pull request #64060 from zhscn/wip-seastore-2q

crimson/os/seastore: introduce 2Q cache replacement algorithm

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

Merge pull request #64452 from xxhdx1985126/wip-72087

crimson/os/seastore/epm: make max_generation and gc_formula configurable

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

Merge pull request #64617 from xxhdx1985126/wip-seastore-invalidate-logical

crimson/os/seastore/logical_child_node: LogicalChildNode should respond to the invalidation event

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

Merge pull request #63604 from guojidan/kv-tool

crimson: Add objectstore tool with partial functionality

Reviewed-by: Aishwarya Mathuria <amathuri@redhat.com>

doc/radosgw: Simple fixes and improvements, links improvements

Fix table with a column separator problem in s3/bucketops.rst.

Remove whitespaces at end of lines in s3/bucketops.rst.

Linkify mention of multizone into multisite.rst in bucket_logging.rst.

Separate units from numbers with a space in bucket_logging.rst

Consistency in capitalization and full stop usage in table data in
s3-notification-compatibility.rst s3/bucketops.rst.

Use ref for intra-docs link instead of "external links" feature in
s3/bucketops.rst notifications.rst s3.rst, add a label in start of
s3-notification-compatibility.rst for it. Follow label format that seems
to be in the majority.
Use auto-generated link text that ref provides.
Reflow the text in the cell. Extend table syntax width to accommodate
longer text in cell.

Use ref similarly on links to s3/bucketops.rst. Add a label in it and
use it from bucket_logging.rst and notifications.rst.

Delete unused external link definition in s3/bucketops.rst.

Remove multiple whitespace at the end of lines in notifications.rst
s3-notification-compatibility.rst bucketops.rst.

Change tab characters to spaces in indentation in bucketops.rst

Signed-off-by: Ville Ojamo <14869000+bluikko@users.noreply.github.com>

Merge pull request #64070 from samarahu/samarahu-get-storage-admin

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

Merge pull request #64733 from stackhpc/doc-fix-mclock-links

doc: Fix links to mClock config reference

Merge pull request #64466 from chanyoung/fix-list-complete

crimson/.../omap_btree_node_impl: fix complete flag for depth 2 omap scan

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

qa/suites/crimson-rados: Introduce objectstore_tool testing

Signed-off-by: Matan Breizman <mbreizma@redhat.com>

qa/tasks/ceph_objectstore_tool: Support Crimson tool

fixes: https://tracker.ceph.com/issues/71504

Signed-off-by: junxiang Mu <1948535941@qq.com>
Signed-off-by: Matan Breizman <mbreizma@redhat.com>

Merge pull request #64323 from guits/cv-seastore

ceph-volume: add seastore OSDs support

Reviewed-by: Adam King <adking@redhat.com>
Reviewed-by: Matan Breizman <mbreizma@redhat.com>
Reviewed-by: Yingxin Cheng <yingxin.cheng@intel.com>

doc/cephfs: edit ceph-dokan.rst (2 of x)

Edit the last half of doc/cephfs/ceph-dokan.rst. This is part of a
larger project whose goal is the improvement of the CephFS
documentation.

Follows https://github.com/ceph/ceph/pull/64690.

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

doc: Fix links to mClock config reference

Signed-off-by: Pierre Riteau <pierre@stackhpc.com>

Merge pull request #63555 from Matan-B/wip-matanb-crimson-asock-race-coro

crimson/osd: committed/handle_osd_maps into coroutines

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

Merge pull request #63270 from VinayBhaskar-V/wip-resync-changes

rbd-mirror: prevent image deletion if remote image is not primary

Reviewed-by: Ramana Raja <rraja@redhat.com>
Reviewed-by: Ilya Dryomov <idryomov@gmail.com>

Merge pull request #64601 from tchaikov/wip-dencoder-without-hitset

tools/ceph-dencoder: do not test HitSet types

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

Merge pull request #64705 from tchaikov/wip-pybind-setup-license-classifier

pybind: replace deprecated license classifiers with SPDX expression

Reviewed-by: Ilya Dryomov <idryomov@gmail.com>
Reviewed-by: John Mulligan <jmulligan@redhat.com>

Merge pull request #64711 from idryomov/wip-72134-cont

qa/suites/krbd: request msgr1 explicitly in unmap subsuite

Reviewed-by: Miki Patel <miki.patel132@gmail.com>

Merge pull request #64583 from ceph/wip-yuriw-release-19.2.3-main

doc: squid 19.2.3 release notes

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

Merge pull request #63348 from AliMasarweh/wip-alimasa-68183

RGW | fix conditional Delete, MultiDelete and Put

Reviewed-by: Casey Bodley <cbodley@redhat.com>, Matt Benjamin <mbenjamin@redhat.com>

Merge pull request #64503 from Matan-B/wip-matanb-crimson-misdirected-fix

crimson/osd/../client_request: fix is_misdirected

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

qa/suites/krbd: request msgr1 explicitly in unmap subsuite

Since commit 5011cc926cd4 ("qa/suites/krbd: run unmap subsuite with
msgr1 only"), unmap.t is run only against msgr1.  pre-single-major.yaml
kernel has actually been gone for some time now, but there is still
value in maintaining a msgr1-only test.  With the default switched to
msgr2 in commit a577f6fa405c ("krbd: "rbd device map" command should
use msgr2 by default"), msgr1 needs to be requested explicitly.

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

doc/man/rbd: update ms_mode default (legacy -> prefer-crc)

The default was switched to msgr2 in prefer-crc mode in commit
a577f6fa405c ("krbd: "rbd device map" command should use msgr2 by
default").

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

Merge pull request #63945 from ArbitCode/wip-raja-get-account-summary

rgw/iam: getAccountSummary API

Merge pull request #64592 from ArbitCode/wip-raja-get-caller-identity-docs

doc/radosgw: get caller identity docs

Merge pull request #64289 from sseshasa/wip-mclock-new-benchmark-mechanism

src/ceph_osd, osd: Implement running benchmark during OSD creation - Phase 1

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

crimson/osd/../client_request: drop rwoedered with balanced/localize reads

This is the counterpart of https://github.com/ceph/ceph/pull/62806
Which fixes classical: https://tracker.ceph.com/issues/70715

Fixes: https://tracker.ceph.com/issues/71007
Signed-off-by: Matan Breizman <mbreizma@redhat.com>

crimson/osd/../client_request: rename is_misdirected and add logs

This function mainly verifies replicated read related info.

Signed-off-by: Matan Breizman <mbreizma@redhat.com>

crimson/osd: handle_osd_map to coroutines

Signed-off-by: Matan Breizman <mbreizma@redhat.com>

crimson/osd: committed_osd_maps into coroutines

Signed-off-by: Matan Breizman <mbreizma@redhat.com>

crimson/osd: committed_osd_maps fix paramter types

Signed-off-by: Matan Breizman <mbreizma@redhat.com>

Merge pull request #64593 from chanyoung/fix-omap-iterate

crimson/.../omap_btree_node_impl: fix inner node iterate

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

Merge pull request #64585 from myoungwon/wip-parallel-write-ool

crimson/os/seastore/rbm: improve large sequential write by issuing writes in parallel

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

Merge pull request #64498 from xxhdx1985126/wip-seastore-io-block-time

crimson/os/seastore/async_cleaner: measure io block time

Reviewed-by: Matan Breizman <mbreizma@redhat.com>
Reviewed-by: Yingxin Cheng <yingxin.cheng@intel.com>

pybind: replace deprecated license classifiers with SPDX expression

Replace the deprecated "License :: OSI Approved :: GNU Lesser General
Public License v2 or later (LGPLv2+)" classifier with the SPDX license
expression "LGPL-2.0-or-later" in the license field.

Recent setuptools versions warn that license classifiers are deprecated
in favor of SPDX expressions in the license field. This change eliminates
the deprecation warning while maintaining the same licensing terms.

References:
- https://setuptools.pypa.io/en/stable/history.html#v77-0-0
- https://packaging.python.org/en/latest/guides/writing-pyproject-toml/#license
- https://spdx.org/licenses/LGPL-2.0-or-later.html

The warning looked like:
```
[1/1] Generating ../../../lib/cython_modules/lib.3/rados.cpython-313-x86_64-linux-gnu.so
/usr/lib/python3.13/site-packages/setuptools/dist.py:759: SetuptoolsDeprecationWarning: License classifiers are deprecated.
!!

        ********************************************************************************
        Please consider removing the following classifiers in favor of a SPDX license expression:

        License :: OSI Approved :: GNU Lesser General Public License v2 or later (LGPLv2+)

        See https://packaging.python.org/en/latest/guides/writing-pyproject-toml/#license for details.
        ********************************************************************************

!!
  self._finalize_license_expression()
```

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

Merge pull request #64418 from ifed01/wip-ifed-discard-mkfs

os/bluestore: make full device discard on mkfs a default behavior

Merge pull request #64142 from phlogistonjohn/jjm-smb-ip-bind

smb: support custom ip address binds

Reviewed-by: Adam King <adking@redhat.com>
Reviewed-by: Anoop C S <anoopcs@cryptolab.net>

Merge pull request #64690 from zdover23/wip-doc-2025-07-25-cephfs-ceph-dokan

doc/cephfs: edit ceph-dokan.rst (1 of x)

doc/cephfs: edit ceph-dokan.rst (1 of x)

Edit the first 62 lines of doc/cephfs/ceph-dokan. This is part of a
larger project whose goal is the improvement of the CephFS
documentation.

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

Merge PR #64523 into main

* refs/pull/64523/head:

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

Merge pull request #64409 from bolzo/mds_disallow_root_fragmentation

MDS: disallow root directory fragmentation

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

Merge PR #64308 into main

* refs/pull/64308/head:

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

Merge PR #64297 into main

* refs/pull/64297/head:

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

Merge pull request #63524 from batrick/i71462

 mds: wrong snap check for directory with parent snaps

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

Merge PR #62554 into main

* refs/pull/62554/head:

Reviewed-by: Venky Shankar <vshankar@redhat.com>
Reviewed-by: Patrick Donnelly <pdonnell@ibm.com>

Merge pull request #64465 from tchaikov/wip-crimson-formatter-sans-fmt-local

crimson/common: replace deprecated fmt::localtime() with localtime_r()

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

Merge pull request #63408 from aainscow/ec_recovery_zero_detect

OSD: EC recovery zero detect

Merge pull request #64545 from ajarr/cleanup-librbd-mirror-enable-req

librbd/mirror: cleanup EnableRequest::handle_get_mirror_image()

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

Merge pull request #64193 from tchaikov/wip-bluestore-test-shutdown

os/bluestore: fix memory leak in HybridAllocator destructor

Merge pull request #64661 from afreen23/revert-stylus

Revert "mgr/dashboard: Fix stylus issue"

Reviewed-by: Naman Munet <nmunet@redhat.com>

Merge pull request #64352 from aclamk/aclamk-bs-remove-cache_blob

os/bluestore: Get rid of unused CACHE_BLOB_BL

Merge pull request #64219 from kamoltat/wip-ksirivad-fix-qa-rados-test

qa/workunits/rados/test.sh: add timeout mechanism and more info to workloads with parallel tests

Merge pull request #63642 from ifed01/wip-ifed-dynamic-vsel-params

os/bluestore: make vselector reserved* parameters applicable in run-time

doc/releases/squid: More notable changes in RGW

Signed-off-by: Adam C. Emerson <aemerson@redhat.com>
Signed-off-by: Yuri Weinstein <yweinste@redhat.com>

Merge pull request #64667 from afreen23/fix-localization-load

mgr/dashboard: Fix issue with loading localization module

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

Merge pull request #62790 from ceph/rm-ceph-common-postun

ceph.spec.in: don't rm ceph conf and logs in ceph-common postun

Merge pull request #64535 from yuvalif/wip-yuval-71979

rgw/test/notification: add more info when retry test fail

RGW | Added debugs in cases where precondition check fails

Signed-off-by: Ali Masarwa <amasarwa@redhat.com>

librbd/mirror: cleanup EnableRequest::handle_get_mirror_image()

In the EnableRequest state machine, clean up the handling of the async
request to fetch the mirror image, particularly when a non-primary image
is being created by the rbd-mirror daemon.

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

Merge pull request #64402 from sam0044/sam0044-bug_72034

mgr/cephadm: updating maintenance health status in the serve loop

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

Merge pull request #63825 from leomylonas/main

mgr/cephadm: handle possibly undefined template variable in haproxy.cfg.j2

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

Merge pull request #63391 from Kushal-deb/fix-rgw-bootstrap

mgr/rgw: Improve error handling and add --resume flag to 'ceph rgw realm bootstrap'

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

Merge pull request #64185 from kshtsk/wip-add-cephadm-file-path

qa/tasks/cephadm: allow to select from 'cephadm' and 'cephadm.py'

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

Merge pull request #64635 from abitdrag/tracker_72134

krbd: "rbd device map" command should use msgr2 by default

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

Merge pull request #62886 from Kushal-deb/fix-issue-2302464-ceph_orch_commands_to_report_the_exit_code

mgr/cephadm: Provide appropriate exit codes for orch operations

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

Merge pull request #64454 from anoopcs9/fix-var-lib-samba-perms

cephadm: Bind mount /var/lib/samba with 0755

Reviewed-by: John Mulligan <jmulligan@redhat.com>
Reviewed-by: Sachin Prabhu <sp@spui.uk>
Reviewed-by: Shwetha K Acharya <Shwetha.K.Acharya@ibm.com>

mgr/dashboard: Fix issue with loading localization module

- this module is only required to be imported from polyfill
-  nx now loading this module as well breaking CI tests and dashboard `@angular/localize`

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

Merge pull request #64616 from afreen23/fix-smb

mgr/dashboard: Fix redirection of SMB enable module

Reviewed-by: Naman Munet <nmunet@redhat.com>

Revert "mgr/dashboard: Fix stylus issue"

This reverts commit c7053dff52bc1a93af45b8017a6e13a578a2f71e.

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

krbd: "rbd device map" command should use msgr2 by default

Making msgr2 and ms_mode=prefer-crc as default option for "rbd device
map" command

Fixes: https://tracker.ceph.com/issues/72134
Signed-off-by: Miki Patel <miki.patel132@gmail.com>

Merge PR #62682 into main

* refs/pull/62682/head:

Reviewed-by: Patrick Donnelly <pdonnell@ibm.com>
Reviewed-by: Venky Shankar <vshankar@redhat.com>

doc/mgr: add documentation for bind_addrs option

Add some documentation for the bind_addrs option including a warning
about how combining public_addrs and bind_addrs gets quirky.

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

mgr/smb: add bind_addrs field to cluster resource type

Add bind_addrs, which is largely a wrapper around the service spec's
bind_addrs to the smb cluster resource.

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

cephadm: add support for specific network binds to smb service

Add a bunch of code to support specific IP address (and/or interface -
see below) binds for the smb service. When the smb service is not
clustered it is using container networking - in this case we use
publish options for the container manager to only listen on the supplied
addresses.

When the smb service is clustered we need to jump through a bunch of
hoops to configure each service individually. Many are easy with just
a short set of CLI options. CTDB only listens on the (first) node
address that it can bind to and only that. smbd has complex interactions
based on the `interfaces` and `bind interfaces only` config parameters.
Because these parameters may be unique to a node (addresses certainly
will be - and interfaces names could be) we can not store this in
the registry based conf. Instead, we take the slightly hacky approach
of generating a stub conf file with just the interfaces related params
in them and telling sambacc to generate a config that includes this
stub config.

IMPORTANT: When using ctdb with public addresses smbd doesn't know what
additional IPs it may need to listen to, so instead of binding to
a fixed IP we configure it to use an interface. This does have a
downside of possibly listening to another address on the same interface
we don't want it to. Additionally, I have observed that as addresses
are added or removed from the interface by ctdb, smbd doesn't
consistently start listening to those addresses.

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