qa/tasks/rados.py: Allow rados task to override config

Problem:

Currently, no option override the config in rados task.

Solution:

Enable override of the config file in rados task.

Signed-off-by: Kamoltat <ksirivad@redhat.com>
(cherry picked from commit 92bf1a8aa8d0d208577c4076d4a86644c01548d5)

qa/tasks/ceph_manager.py: init test_min_size_duration

Added comment about test_min_size_duration
in qa/tasks/thrashosds.

But also use the variable in ceph_manager.py

Signed-off-by: Kamoltat <ksirivad@redhat.com>
(cherry picked from commit 9f19dffc93463513e03908f3506c62e65364c0cd)

qa/suites/rados: Added wait_for_all_active_clean_pgs flag

Added flag to not allow rados suite to delete
the pool unless all pgs are active+clean
and all OSDs are up in the thrashosds side
of the test.

Fixes: https://tracker.ceph.com/issues/59172
Signed-off-by: Kamoltat <ksirivad@redhat.com>
(cherry picked from commit 3ccd10f266cfd7ec6dd1ad930598bfe4ca422a90)

Merge pull request #59251 from zdover23/wip-doc-2024-08-16-backport-59167-to-reef

reef: doc/cephfs: improve "layout fields" text

doc/cephfs: improve "layout fields" text

Improve "layout fields" text in doc/cephfs/file-layouts.rst, as suggesed
by Anthony D'Atri in these comments:

https://github.com/ceph/ceph/pull/59021#discussion_r1704108581
https://github.com/ceph/ceph/pull/59021#discussion_r1704112320

Co-authored-by: Anthony D'Atri <anthony.datri@gmail.com>
Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit 0949b410433837f0509fb73169fa7f22c8f6c256)

Merge pull request #59022 from zdover23/wip-doc-2024-08-05-backport-58891-to-reef

reef: doc/cephfs: edit "Layout Fields" text

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

Merge pull request #59234 from zdover23/wip-doc-2024-08-15-backport-59219-to-reef

reef: doc/rgw/notification: persistent notification queue full behavior

Merge pull request #59231 from idryomov/wip-67353-reef

reef: qa: adjust expected io_opt in krbd_discard_granularity.t

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

Merge pull request #59151 from idryomov/wip-53674-reef

reef: librbd/crypto: fix issue when live-migrating from encrypted export

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

doc/rgw/notification: persistent notification queue full behavior

Fixes: https://tracker.ceph.com/issues/50610
Signed-off-by: Yuval Lifshitz <ylifshit@ibm.com>
(cherry picked from commit d12ba11741dc749bcce315bf467078595fa95b24)

qa: cover a custom object size in krbd_discard_granularity.t

Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
(cherry picked from commit e8126bf2f6411069da5074ac3a5a2ea16c0bba0c)

qa: adjust expected io_opt in krbd_discard_granularity.t

With linux.git commit a00d4bfce7c6 ("rbd: increase io_opt again"),
io_opt is set to object set size.

Fixes: https://tracker.ceph.com/issues/67353
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
(cherry picked from commit 144270eb737850159614abd63c261baaa83a8afc)

Merge pull request #59215 from zdover23/wip-doc-2024-08-14-backport-59168-to-reef

reef: doc/cephfs: improve cache-configuration.rst

doc/cephfs: improve cache-configuration.rst

Improve the text in the section about dealing with cache-pressure alerts
that was added in https://github.com/ceph/ceph/pull/59077. The changes
in this commit were suggested by Anthony D'Atri.

Co-authored-by: Patrick Donnelly <pdonnelly@redhat.com>
Co-authored-by: Anthony D'Atri <anthony.datri@gmail.com>
Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit aa3bdae2314fef2fca8fc12dca006af657235e17)

librbd/migration: make ImageDispatch handle encryption for non-native formats

With NativeFormat now being handled via dispatch, handling encryption
for non-native formats (i.e. mapping to raw image extents and performing
decryption/mapping back on completion) in the migration layer is really
straightforward.

Note that alignment doesn't need to be performed in the migration layer
because it happens on the destination image -- the "align and resubmit"
logic in C_UnalignedObjectReadRequest should kick in before the call to
read_parent().

Fixes: https://tracker.ceph.com/issues/53674
Co-authored-by: Or Ozeri <oro@il.ibm.com>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
(cherry picked from commit 0000c3447407772039121bb4499f243df1c889da)

Conflicts:
src/librbd/migration/ImageDispatch.cc [ commit 20aee5bbbcb5
  ("neorados: Make IOContext getters/setters less weird") not
  in reef ]

librbd: don't make an extra copy of image_extents in C_ImageReadRequest ctor

Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
(cherry picked from commit b20a897061feabc4e22c339c4e7a8aa5155151e8)

qa/workunits/rbd: perform cleanup in test_clone_encryption()

... so that RAW_DEV can be unmapped and future tests can reuse testimg
and other image names without bumping into watchers and older snapshots.

Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
(cherry picked from commit 82d26909cb12b62d88f69f90eb8af692e497bddb)

qa/workunits/rbd: no need to chmod in luks-encryption.sh

Most workunits expect the user to be a member of "disk" group, so we
can pretty much rely on that being the case at this point.

Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
(cherry picked from commit 830cbee7a5f875af04f335266b02ad96e4cd71c4)

librbd/migration: make FormatInterface::read() void again

Now that NativeFormat is handled via dispatch, FormatInterface::read()
can be void again for consistency with FormatInterface::list_snaps().

Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
(cherry picked from commit b6c7f69b8044f9206aa982c2aba6466c49fd2bea)

Conflicts:
src/librbd/migration/ImageDispatch.cc [ commit 20aee5bbbcb5
  ("neorados: Make IOContext getters/setters less weird") not
  in reef ]

librbd/migration: close source image in OpenSourceImageRequest

Currently, on errors in FormatInterface::open(), RawFormat disposes
of src_image_ctx, but QCOWFormat doesn't, which is a leak.  Rather than
having each format do it internally, do it in OpenSourceImageRequest.

Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
(cherry picked from commit 63159d6b431470f5edc4b110cebf46865c550689)

librbd/migration: don't instantiate NativeFormat, handle it via dispatch

Trying to shoehorn NativeFormat under FormatInterface doesn't really
work.  It fundamentally doesn't fit in:

- Unlike for RawFormat and QCOWFormat, src_image_ctx for NativeFormat
  is not dummy -- it's an ImageCtx for a real RBD image.  Pre-creating
  it in OpenSourceImageRequest with the expectation that placeholder
  values would be overridden later forces NativeFormat to reach into
  ImageCtx guts, duplicating the logic in the constructor.  This also
  necessitates calling snap_set() in a separate step, since snap_id
  isn't known at the time ImageCtx is created.

- Unlike for RawFormat and QCOWFormat, get_image_size() and
  get_snapshots() implementations for NativeFormat are dummy.

- read() and list_snaps() implementations for NativeFormat are
  inconsistent: read() passes through io::ImageDispatch layer, but
  list_snaps() doesn't.  Both can be passing through, meaning that in
  essence these are also dummy.

All of this is with today's code.  Additional complications arise with
planned support for migrating from external clusters where src_image_ctx
would require more invasive patching to "move" to an IoCtx belonging to
an external cluster's CephContext and also with other work.

With the above in mind, NativeFormat actually consists of:

1. Code that parses the "type: native" source spec
2. Code that patches ImageCtx, working around the fact that it's
   pre-created in OpenSourceImageRequest
3. A bunch of dummy implementations for FormatInterface

With this change, (1) is wrapped into a static method that also creates
ImageCtx after all required parameters are known and (2) and (3) go away
entirely.  NativeFormat no longer implements FormatInterface and doesn't
get instantiated at all.

Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
(cherry picked from commit cacf7ca941876f64f9a04867ffc6cdcb484d89b9)

librbd/migration/NativeFormat: refactor source spec parsing

In preparation for not instantiating NativeFormat and losing a copy of
the source spec JSON object in m_json_object, refactor the parsing code
to use only const methods (which std::map's operator[] isn't) and local
variables where possible.

Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
(cherry picked from commit 3bbf1f5ddbaa4a8c252d70a384e23852f0c537c1)

librbd/migration/NativeFormat: do pool lookup instead of creating io_ctx

A Rados instance is sufficient to map the pool name to the pool ID,
no need to involve an IoCtx instance as well.  While at it, report
distinctive errors for a non-existing pool and an invalid JSON value
for pool_name key cases.

Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
(cherry picked from commit 1ba9a32598f50073b574b4649736d76b678a1c58)

librbd/migration: make SourceSpecBuilder::parse_source_spec() static

In preparation for divorcing NativeFormat from FormatInterface and
changing when/how src_image_ctx is created, make parse_source_spec()
independent of src_image_ctx.  The "invalid source-spec JSON" error is
duplicated by the "failed to parse migration source-spec" error, so
just get rid of the former to spare having to pass CephContext to
parse_source_spec().

Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
(cherry picked from commit f172fb97be9a6129be7cdbaa87346dc6c8e8ccb1)

librbd/migration/OpenSourceImageRequest: rename io_ctx -> dst_io_ctx

For now, this is just slightly clearer.  The distinction would become
important with planned support for migrating from external clusters.

Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
(cherry picked from commit c14356b1f9eea0988e071f40dc0df005f70edd4d)

librbd/migration: massage some error messages

Add missing spaces, don't use the word stream when reporting errors
on POSIX file operations (open() and lseek64()) and fix a cut-and-paste
typo in RawSnapshot.

Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
(cherry picked from commit 789df2ce38f35ffb3e86974e02868e5fff71e72c)

librbd/api: clean up leftovers in Migration::prepare_import()

Dead code after return and an unused variable.

Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
(cherry picked from commit b92ad49a35536fff87d840ffbf171aee61b05424)

Merge pull request #59149 from zdover23/wip-doc-2024-08-11-backport-59077-to-reef

reef: doc/cephfs: add cache pressure information

Merge pull request #58853 from idryomov/wip-67051-reef

reef: qa/workunits/rbd: avoid caching effects in luks-encryption.sh

Reviewed-by: Ramana Raja <rraja@redhat.com>
Reviewed-by: Mykola Golub <mgolub@suse.com>

Merge pull request #58540 from idryomov/wip-66886-reef

reef: qa: account for rbd_trash object in krbd_data_pool.sh + related ceph{,adm} task fixes

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

doc/cephfs: add cache pressure information

Add information to doc/cephfs/cache-configuration.rst about how to deal
with a message that reads "clients failing to respond to cache
pressure". This procedure explains how to slow the growth of the
recall_caps value so that it does not exceed the
mds_recall_warning_threshold.

The information in this commit was developed by Eugen Block. See
https://lists.ceph.io/hyperkitty/list/ceph-users@ceph.io/thread/5ROH5CWKKOEIQMVXOVRT5OO7CWK2HPM3/#J65DFUPP4BY57MICPANXKI7KAXSZ5Z5P
and https://www.spinics.net/lists/ceph-users/msg73188.html.

Fixes: https://tracker.ceph.com/issues/57115
Co-authored-by: Eugen Block <eblock@nde.ag>
Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit bf26274ae4737417193f8c2b56bea20eb2a358aa)

Merge pull request #57901 from adamemerson/wip-62292-reef

reef: rgw: modify string match_wildcards with fnmatch

Merge pull request #57301 from cbodley/wip-65822-reef

reef: rgw: fix CompleteMultipart error handling regression

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

Merge pull request #52611 from cbodley/wip-62142-reef

reef: valgrind: update suppression for SyscallParam under call_init

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

Merge pull request #59100 from zdover23/wip-doc-2024-08-09-backport-59807-to-reef

reef: docs/rados/operations/stretch-mode: warn device class is not supported

docs/rados/operations/stretch-mode: warn device class is not supported

Signed-off-by: Kamoltat Sirivadhna <ksirivad@redhat.com>
(cherry picked from commit aa1d8cf4fa321e24e850bd5f687a6ddad3ce05e3)

rgw: move publish_complete() back to RGWCompleteMultipart::execute()

move publish_complete() and meta_obj->delete_object() back to execute()
so they only run on success. this allows several member variables to
move back to execute()'s stack as well

Fixes: https://tracker.ceph.com/issues/65746
Signed-off-by: Casey Bodley <cbodley@redhat.com>
(cherry picked from commit ebb37c7bb3aee4663220054c6516164bf046fa8c)

rgw: CompleteMultipart uses s->object for Notification

get_notification() should be associated with the target object
s->object. the meta_obj has the wrong object name, so required passing
s->object->get_name() as an extra argument

importantly, Notification no longer depends on the lifetime of meta_obj
to avoid a dangling pointer, while the lifetime of s->object is guaranteed

Signed-off-by: Casey Bodley <cbodley@redhat.com>
(cherry picked from commit 91dc7f3be945dccd8f59e070e9bc43a2a5df12db)

rgw: CompleteMultipart uses s->object instead of target_obj

most requests operate directly on s->object. there's no reason to
allocate a separate target_obj for the same purpose

Signed-off-by: Casey Bodley <cbodley@redhat.com>
(cherry picked from commit d09b8ab2e077ceb6a0c6dfb99ce1b45d63a28be4)

Merge pull request #58793 from ivancich/wip-67155-reef

reef: test/rgw: address potential race condition in reshard testing

Merge pull request #58435 from cbodley/wip-64465-reef

reef: rgw: cumulatively fix 6 AWS SigV4 request failure cases

Merge pull request #58168 from cbodley/wip-66580-reef

reef: rgw: optimize gc chain size calculation

Merge pull request #57425 from mkogan1/wip-65886-reef

reef: rgw/beast: fix crash observed in SSL stream.async_shutdown()

Merge pull request #57127 from jzhu116-bloomberg/wip-64325-reef

reef: rgw/multisite: avoid writing multipart parts to the bucket index log

Merge pull request #56615 from prazumovsky/wip-63621

reef: rgw/swift: preserve dashes/underscores in swift user metadata names

Merge pull request #56554 from petrutlucian94/wip-64326-reef

reef: RGW: fix cloud-sync not being able to sync folders

Merge pull request #55200 from ifed01/wip-ifed-refactor-spillover-tests-reef

reef: test/store_test: refactor spillover tests

Reviewed-by: Adam Kupczyk <akupczyk@redhat.com>

Merge pull request #55136 from ifed01/wip-ifed-refresh-perf-counters-ref

reef: os: introduce ObjectStore::refresh_perf_counters() method

Reviewed-by: Adam Kupczyk <akupczyk@redhat.com>

Merge pull request #55135 from ifed01/wip-ifed-fast-shutdown-assert-ref

reef: osd: do not assert on fast shutdown timeout

Reviewed-by: Nitzan Mordechai <nmordech@redhat.com>

Merge pull request #55015 from Matan-B/wip-63880-reef

reef: tools/ceph_objectstore_tool: Support get/set/superblock

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

Merge pull request #54533 from pdvian/wip-63365-reef

reef: mgr: remove out&down osd from mgr daemons

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

Merge pull request #59079 from zdover23/wip-doc-2024-08-07-backport-59040-to-reef

reef: doc: documenting the feature that scrub clear the entries from damage…

doc: documenting the feature that scrub clear the entries from damage table on repair

Fixes: https://tracker.ceph.com/issues/67337
Signed-off-by: Neeraj Pratap Singh <neesingh@redhat.com>
(cherry picked from commit 140e77638dc95fefcf62f01fe7e45bc0729439be)

Merge pull request #58998 from guits/wip-64642-reef

reef: src/ceph-volume/ceph_volume/devices/lvm/listing.py : lvm list filters with vg name

src/ceph-volume/ceph_volume/devices/lvm/listing.py : lvm list filters also on vg name

This commit fix the listing of LVs with the same name on multiple VG

Fixes: https://tracker.ceph.com/issues/62320
Signed-off-by: Pierre Lemay <pierre.lemay@gmail.com>
(cherry picked from commit b2c4e62afac32edda142a51eb601420a1a79bb2f)

Merge pull request #58997 from guits/wip-66465-reef

reef: ceph-volume: fix set_dmcrypt_no_workqueue()

Merge pull request #58996 from guits/wip-66446-reef

reef: ceph-volume: do source devices zapping if they're detached.

Merge pull request #58483 from guits/wip-66027-reef

reef: node-proxy: make the daemon discover endpoints

Merge pull request #57475 from guits/wip-66022-reef

reef: cephadm: change loki/promtail default image tags

Merge pull request #59063 from zdover23/wip-doc-2024-08-07-backport-59060-to-reef

reef: doc/dev/peering: Change acting set num

doc/dev/peering: Change acting set num

This commit updates the peering doc page with the changed [2,1,2] to [3,1,2]

Signed-off-by: qn2060 <qn2060@users.noreply.github.com>
(cherry picked from commit 103d65ea6787ab16d1824f71578264e0abb9820c)

Merge pull request #59032 from zdover23/wip-doc-2024-08-05-backport-57905-to-reef

reef: doc/mgr/dashboard: fix TLS typo

doc/mgr/dashboard: fix TLS typo

Change "SSL/TSL" to "SSL/TLS"

Signed-off-by: Mindy Preston <mcpreston@icecube.wisc.edu>
(cherry picked from commit e63b967aeb00a667ab5c91b9b526c7a19a9539ec)

doc/cephfs: edit "Layout Fields" text

Improve the English in the "Layout Fields" section of
doc/cephfs/file-layouts.rst.

Co-authored-by: Anthony D'Atri <anthony.datri@gmail.com>
Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit 9607b66e97f73b5b7c53c2610b928cd6a420c9f2)

Merge pull request #59016 from Svelar/wip-revert-rgw

Revert "reef: rgw/amqp: lock erase and create connection before emplace"

Revert "reef: rgw/amqp: lock erase and create connection before emplace"

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

Merge pull request #58715 from Svelar/wip-67072-reef

reef: rgw/amqp: lock erase and create connection before emplace

Merge pull request #59008 from zdover23/wip-doc-2024-08-04-backport-59006-to-reef

reef: doc/glossary: add "ceph-ansible"

doc/glossary: add "ceph-ansible"

Add a "ceph-ansible" entry to the glossary.

Co-authored-by: Anthony D'Atri <anthony.datri@gmail.com>
Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit 484e9d2b9d81b6de498ac08125c9a458730dd444)

ceph-volume: fix set_dmcrypt_no_workqueue()

`set_dmcrypt_no_workqueue()` from `ceph_volume.util.encryption`

The function `set_dmcrypt_no_workqueue` in `encryption.py` now
dynamically retrieves the installed cryptsetup version using `cryptsetup
--version` command. It then parses the version string using a regular
expression to accommodate varying digit counts. If the retrieved version
is greater than or equal to the specified target version,
`conf.dmcrypt_no_workqueue` is set to True, allowing for flexible version
handling.

Fixes: https://tracker.ceph.com/issues/66393
Signed-off-by: Guillaume Abrioux <gabrioux@ibm.com>
(cherry picked from commit dc28b77a6ea50b3390663ac02eeb80367650b7ed)

ceph-volume: do source devices zapping if they're detached.

One needs to zap source device(s) after DB/WAL migration.
Original imlementation removes LVM tags only which leaves device(s) in a
state where "ceph-volume raw activate" still reconginizes them as
attached to OSD due to information preserved in bdev label.
Hence the need to do more zapping.
Fixes: https://tracker.ceph.com/issues/66315
Signed-off-by: Igor Fedotov <igor.fedotov@croit.io>
(cherry picked from commit ae5ef432845dcf9b061258357ffd97f4eae59a63)

Merge pull request #58978 from zdover23/wip-doc-2024-08-01-backport-58899-to-reef

reef: doc/glossary: add "Prometheus"

Merge pull request #58974 from zdover23/wip-doc-2024-08-01-backport-58972-to-reef

reef: doc: Improve doc/radosgw/placement.rst

doc/glossary: add "Prometheus"

Add a glossary entry for "Prometheus".

Co-authored-by: Anthony D'Atri <anthony.datri@gmail.com>
Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit 1a76464999cb2d2cc748d69e1bb0567544a0bb8b)

doc: Improve doc/radosgw/placement.rst

Signed-off-by: Anthony D'Atri <anthonyeleven@users.noreply.github.com>
(cherry picked from commit 5890a156524b8d1fe3a51ea1c76d800e5903a01c)

Merge pull request #58948 from zdover23/wip-doc-2024-07-31-backport-58946-to-reef

reef: doc/rados/operations: remove vanity cluster name reference from crush…

doc/rados/operations: remove vanity cluster name reference from crush-map.rst

Signed-off-by: Anthony D'Atri <anthonyeleven@users.noreply.github.com>
(cherry picked from commit 62562ec65e3c1cc040a57ab77fc219e0aad25651)

Merge pull request #58941 from zdover23/wip-doc-2024-07-31-backport-58922-to-reef

reef: doc/cephadm: edit "Using Custom Images"

Merge pull request #58944 from zdover23/wip-doc-2024-07-31-backport-58908-to-reef

reef: doc/glossary: add "ceph-fuse" entry

Merge pull request #58938 from zdover23/wip-doc-2024-07-31-backport-58918-to-reef

reef: doc/dev: improve basic-workflow.rst

doc/glossary: add "ceph-fuse" entry

Add a "ceph-fuse" entry to the glossary.

Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit d1c049d24befedc396221dafcc66db8c40acf5ab)

doc/cephadm: edit "Using Custom Images"

Edit the section "Using Custom Images" in
doc/cephadm/services/monitoring.rst.

Co-authored-by: Anthony D'Atri <anthony.datri@gmail.com>
Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit e99fc5c82fd698c222a489a1ebc26c8fcbca321d)

doc/dev: improve basic-workflow.rst

Make various improvements to doc/dev/developer_guide/basic-workflow.rst.

Co-authored-by: Anthony D'Atri <anthony.datri@gmail.com>
Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit b81d6af1b488cdb87143e13da519d70c3a090679)

Merge pull request #58910 from zdover23/wip-doc-2024-07-30-backport-58904-to-reef

reef: doc/cephfs: edit "Dynamic Subtree Partitioning"

doc/cephfs: edit "Dynamic Subtree Partitioning"

Edit the section "Dynamic Subtree Partitioning" in
doc/cephfs/multimds.rst.

Co-authored-by: Anthony D'Atri <anthony.datri@gmail.com>
Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit d14119e37e01bd06eed668d9e3606a66ef205b30)

Merge pull request #58902 from zdover23/wip-doc-2024-07-29-backport-58554-to-reef

reef: docs: removed centos 8 and added squid to the build matrix

docs: removed centos 8 and added squid to the build matrix

Signed-off-by: Yuri Weinstein <yweinste@redhat.com>
(cherry picked from commit 8761bbe16c2a6d19e136254c26d1d67ebe839e3b)

Merge pull request #58893 from zdover23/wip-doc-2024-07-28-backport-58884-to-reef

doc/cephfs: fix "OSD capabilities" link

doc/cephfs: fix "OSD capabilities" link

Fix a broken link from the "OSD capabilites" reference on
doc/cephfs/client-auth.rst.

Fixes: https://tracker.ceph.com/issues/67218
Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit 74b03a8cb2f196171b10d8cbde27324240649e70)

Merge pull request #58876 from zdover23/wip-doc-2024-07-27-backport-58867-to-reef

reef: doc/cephfs: fix "file layouts" link

Merge pull request #58856 from idryomov/wip-67097-reef

reef: qa/suites/krbd: stress test for recovering from watch errors for -o exclusive

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

doc/cephfs: fix "file layouts" link

Fix a broken link to the "File Layouts" page on
doc/cephfs/client-auth.rst.

Fixes: https://tracker.ceph.com/issues/67190
Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit 6bb5ce94c6a2cbfdf978a66d00f428ad4cd4c312)

Merge pull request #58864 from zdover23/wip-doc-2024-07-26-backport-58838-to-reef

reef: doc/rbd/rbd-exclusive-locks: mention incompatibility with advisory locks

doc/rbd/rbd-exclusive-locks: mention incompatibility with advisory locks

This is due to both using the same name for the lock at the RADOS level
("rbd_lock").

Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
(cherry picked from commit 2e83b77c9f610f42765266ff4d4d7bc2a635eac8)

qa/suites/krbd: override extra_system_packages directly on install task

Follow commit c61cb16223fa ("qa/suites/rbd: override
extra_system_packages directly on install task").

Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
(cherry picked from commit 2a9f92f23472dfcb1cb0fabc8572a28e6e5dd91a)

qa/suites/krbd: stress test for recovering from watch errors for -o exclusive

This is based on a test added in commit 237aa221ebad ("qa/suites/krbd:
stress test for recovering from watch errors") for regular mappings.

Fixes: https://tracker.ceph.com/issues/67097
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
(cherry picked from commit 8fee41da8b8cd250bbbd8490604193c0864c1295)

qa/workunits/rbd: avoid caching effects in luks-encryption.sh

Commit 40f6f5224bce ("qa/workunits/rbd: fix issues in
luks-encryption.sh") did the right thing for reads, which solved
most of the issue.  However, it actually made a step in the opposite
direction for writes -- depending on the RBD cache settings, rbd-nbd
virtual devices can behave as physical devices with a volatile write
cache, so fsync is required.

While at it, involving O_DIRECT for reads isn't needed outside of
test_encryption_format().

Fixes: https://tracker.ceph.com/issues/67051
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
(cherry picked from commit 9e1c0342e9931e898d795d47f923edd3b4324f4a)

Merge pull request #57873 from ceph/reef-release

v18.2.4

Reviewed-by: Anthony D'Atri <anthony.datri@gmail.com>
Reviewed-by: Guillaume Abrioux <gabrioux@ibm.com>
Reviewed-by: Ilya Dryomov <idryomov@redhat.com>
Reviewed-by: Laura Flores <lflores@redhat.com>

Merge pull request #58775 from zdover23/wip-doc-2024-07-24-backport-58524-to-reef

reef: doc/cephfs: streamline a paragraph

Merge branch 'reef' into reef-release

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