ceph-fuse: perform cleanup if test_dentry_handling failed

If remount failed due to some reason then ceph_abort() is
getting called which causes child process termination
without cleanup.
To fix this issue, ceph_abort() call moved after
performing cleanup.

Fixes: https://tracker.ceph.com/issues/54049
Signed-off-by: Nikhilkumar Shelke <nshelke@redhat.com>
(cherry picked from commit 8c778e79840f1aa9b9731e2ef20881da0d122fda)

Merge pull request #45328 from kamoltat/wip-ksirivad-backport-pacific-fix-autoscale-doc

 pacfic: doc/rados/operations/placement-groups: fix --bulk docs
Reviewed-by: Neha Ojha <nojha@redhat.com>

Merge pull request #45301 from rhcs-dashboard/wip-53700-pacific

pacific: mgr/dashboard: fix white screen on Safari

Reviewed-by: Sarthak0702 <NOT@FOUND>
Reviewed-by: Aashish Sharma <aasharma@redhat.com>
Reviewed-by: Ernesto Puerta <epuertat@redhat.com>
Reviewed-by: Nizamudeen A <nia@redhat.com>

Merge pull request #45326 from ceph/wip-yuri-pacific-clients

qa/tests: added upgrade-clients/client-upgrade-pacific-quincy test

Reviewed-by: Josh Durgin <jdurgin@redhat.com>
Reviewed-by: Ilya Dryomov <idryomov@gmail.com>

doc/rados/operations/placement-groups: typo two 'the'

typo, should not have two 'the' next to each
other.

Signed-off-by: Kamoltat <ksirivad@redhat.com>
(cherry picked from commit 7f1c7637e229c6a1bd69a4b204a56ae49c7fec44)

doc/rados/operations/placement-groups: fix --bulk commands

Some parts of the documents regarding
the bulk flag  have typos.

Command for creating a pool

was: `ceph osd create test_pool --bulk`

should be: `ceph osd pool create test_pool --bulk`

Command for setting bulk value in a pool

was: `ceph osd pool set test_pool bulk=<true/false/1/0>`

should be: `ceph osd pool set test_pool bulk <true/false/1/0>`

Also removed a bit of trailing white spaces.

Changed `complements` to `complement`.

https://tracker.ceph.com/issues/54485

Signed-off-by: Kamoltat <ksirivad@redhat.com>
(cherry picked from commit 4a01fc77985e5cf919b99eca86c4c7e8aae242f0)

qa/tests: added upgrade-clients/client-upgrade-pacific-quincy test

Signed-off-by: Yuri Weinstein <yweinste@redhat.com>

Merge pull request #45311 from idryomov/wip-test-notify-rebuild-object-map-pacific

pacific: test/librbd/test_notify.py: effect post object map rebuild assert

Reviewed-by: Deepika Upadhyay <dupadhya@redhat.com>

test/librbd/test_notify.py: effect post object map rebuild assert

Instead of just optionally skipping update_features test, commit
9c0b239d70cd ("qa/upgrade: conditionally disable update_features
tests") moved it after rebuild_object_map test.  This isn't right
because update_features test invalidates the object map as a side
effect and rebuild_object_map test is what makes it valid again:

  assert((image.flags() & RBD_FLAG_OBJECT_MAP_INVALID) == 0)

Let's make this assert effective, at least when update_features
test isn't skipped.

Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
(cherry picked from commit 5b94f080a359a70883ba34e5bd03e0e4601a6c6a)

Merge pull request #45228 from rkachach/backport_of_45054_and_44779

pacific: mgr/cephadm: Delete ceph.target if last cluster

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

Merge pull request #45226 from rkachach/backport_of_pr_45023

pacific: mgr/cephadm: Show an error when invalid format

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

Merge pull request #45298 from zdover23/wip-doc-pr45139-backport-2nd-attempt-pacific

doc/start: include A. D'Atri's hardware-recs recs

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

mgr/dashboard: fix white screen on Safari

Safari do not support lookbehind in regular expression.

Fixes: https://tracker.ceph.com/issues/53665
Signed-off-by: 胡玮文 <huww98@outlook.com>
(cherry picked from commit d4b239c63595e1dc9772a135bf6b9c0608f852ce)

doc/start: include A. D'Atri's hardware-recs recs

This PR restores material about partition alignment
and material about separating OS and OSD data that
was removed in an earlier rewrite. The restoration
of this information was requested by Anthony D'Atri in
https://github.com/ceph/ceph/pull/45123/

This PR also includes several refinements to the language
that could not be made to this text until now, owing to my
(Zac's) ignorance and illiteracy.

I call upon Mark Nelson (and anyone else with sufficient
command of the current state of storage technology) to advise
me on whether the Ceph Foundation feels comfortable in the year
2022 referring to QLC as an emerging technology.

Signed-off-by: Zac Dover <zac.dover@gmail.com>
(squash) more notes and revisions

Signed-off-by: Zac Dover <zac.dover@gmail.com>
(cherry picked from commit 07a1f7fe0463c5ef4424646e811626ff4bbf729c)
Signed-off-by: Zac Dover <zac.dover@gmail.com>

Merge pull request #45181 from ideepika/wip-54324-pacific

pacific: tools/rbd: expand where option rbd_default_map_options can be set

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

Merge pull request #45179 from ideepika/wip-54379-pacific

pacific: common: replace BitVector::NoInitAllocator with wrapper struct

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

Merge pull request #45170 from pponnuvel/wip-54381-pacific

pacific: rbd-mirror: make mirror properly detect pool replayer needs restart

Reviewed-by: Ilya Dryomov <idryomov@redhat.com>
Reviewed-by: Deepika Upadhyay <dupadhya@redhat.com>
Reviewed-by: Sunny Kumar <sunkumar@redhat.com>

qa/suites/krbd: add rbd_default_map_options override coverage

Add coverage to test precedence, override, and option merge on rbd map.

Signed-off-by: Christopher Hoffman <choffman@redhat.com>
(cherry picked from commit 19d46b918169601afe9eb834a2361fb015048c67)

qa/suites/krbd: rename rxbounce subsuite

A new job that doesn't want ms_mode to be set underneath it is about to
be added.  Rename rxbounce to ms_modeless to make this purpose obvious.

Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
(cherry picked from commit 7f391c5688105e55f7799a9d45721ec49531747d)

rbd: support pool and image level overrides for rbd_default_map_options

Fixes: https://tracker.ceph.com/issues/52850
Signed-off-by: Christopher Hoffman <choffman@redhat.com>
(cherry picked from commit 9afc9712824a92fd6bdb2574c5880ab835236ed1)

Merge pull request #45234 from adk3798/pacific-extra-all

pacific: mgr/cephadm: extend extra_container_args to other service types

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

Merge pull request #45229 from adk3798/pacific-admin-drain

pacific: mgr/cephadm: block draining last _admin host

Reviewed-by: Michael Fritch <mfritch@suse.com>
Reviewed-by: Redouane Kachach <rkachach@redhat.com>

Merge pull request #45231 from adk3798/pacific-admin-label-rm

pacific: mgr/cephadm: block removing last instance of _admin label

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

Merge pull request #45262 from idryomov/wip-cmake-disable-dpdk-warnings-pacific

pacific: cmake: pass RTE_DEVEL_BUILD=n when building dpdk

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

cmake: pass RTE_DEVEL_BUILD=n when building dpdk

ceph is still using the Makefile based building system for building
DPDK. and DPDK enables -Werror if RTE_DEVEL_BUILD is 'y' which is
enabled by default when the dpdk is built from a git repo.

but newer GCC is more picky than the older versions, to prevent
the possible FTBFS when we switch to newer GCC for building old
branches whose dpdk submodule might be include the changes addressing
those warnings. let's just disable this option.

the only effect of this option is to add -Werror to CFLAGS. but
the building warnings from DPDK is not our focus when developing
Ceph in the most cases. so it should be fine.

see also
https://github.com/ceph/dpdk/blob/eac901ce29be559b1bb5c5da33fe2bf5c0b4bfd6/doc/build-sdk-quick.txt#L18

Signed-off-by: Kefu Chai <tchaikov@gmail.com>
(cherry picked from commit 91a616b26e830e85962200d0bac86bee7e022892)

Conflicts:
cmake/modules/BuildDPDK.cmake [ commit d3c315703ae6 ("cmake:
  pass -Wunused-but-set-variable when building dpdk") not in
  pacific ]

Merge pull request #45151 from yaarith/wip-45121-pacific

pacific: mgr/devicehealth: skip null pages when extracting wear level

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

Merge pull request #45111 from ljflores/wip-54352-pacific

pacific: mgr/dashboard: dashboard turns telemetry off when configuring report

Reviewed-by: Yaarit Hatuka <yaarit@redhat.com>
Reviewed-by: Nizamudeen A <nia@redhat.com>
Reviewed-by: Sarthak0702 <sarthak.0702@gmail.com>

Merge pull request #44959 from BenoitKnecht/wip-54232-pacific

pacific: mon: Abort device health when device not found

Reviewed-by: Yaarit Hatuka <yaarit@redhat.com>

Merge pull request #45146 from guits/wip-54402-pacific

pacific: ceph-volume: abort when passed devices have partitions

mgr/cephadm: extend extra_container_args to other service types

Otherwise, without this change, this can only be used for mgr,
mon and crash (daemons without their own service spec class)

Fixes: https://tracker.ceph.com/issues/54390
Signed-off-by: Adam King <adking@redhat.com>
(cherry picked from commit d3c14a17dc5cafef199f4fc3ce657bab54d89b4a)

Conflicts:
src/python-common/ceph/deployment/drive_group.py
src/python-common/ceph/deployment/service_spec.py

mgr/cephadm: block removing last instance of _admin label

Fixes: https://tracker.ceph.com/issues/54425
Signed-off-by: Adam King <adking@redhat.com>
(cherry picked from commit fbe0c3fd23f9005986959bade149093c340f6238)

Conflicts:
src/pybind/mgr/rook/module.py

mgr/cephadm: block draining last _admin host

Fixes: https://tracker.ceph.com/issues/54413
Signed-off-by: Adam King <adking@redhat.com>
(cherry picked from commit a0d21c7108e8f95b541bdb7653d2595f68e42520)

mgr/cephadm: Delete ceph.target if last cluster
Fixes: https://tracker.ceph.com/issues/46655
Signed-off-by: Redouane Kachach <rkachach@redhat.com>

mgr/cephadm: Adding logic to cleanup several dirs after an rm-cluster
Fixes: https://tracker.ceph.com/issues/53010
       https://tracker.ceph.com/issues/53815
Signed-off-by: Redouane Kachach <rkachach@redhat.com>

mgr/cephadm: Show an error when invalid format
Fixes: https://tracker.ceph.com/issues/54198
Signed-off-by: Redouane Kachach <rkachach@redhat.com>

Merge pull request #45002 from idryomov/wip-krbd-rxbounce-option-pacific

pacific: rbd: recognize rxbounce map option

Reviewed-by: Deepika Upadhyay <dupadhya@redhat.com>

Merge pull request #45003 from idryomov/wip-54128-pacific

pacific: krbd: return error when no initial monitor address found

Reviewed-by: Deepika Upadhyay <dupadhya@redhat.com>

common: replace BitVector::NoInitAllocator with wrapper struct

in c++20, the deprecated `struct std::allocator<T>::rebind` template was
removed, so `BitVector` no longer compiles. without a `rebind` to
inherit, `std::allocator_traits<NoInitAllocator>::rebind_alloc<U>` was
looking for `NoInitAllocator<U>`, but it isn't a template class

further investigation found that in c++17, `vector<__u32, NoInitAllocator>`
was rebinding this `NoInitAllocator` to `std::allocator<__u32>` and
preventing the no-init optimization from taking effect

instead of messing with the allocator to avoid zero-initialization, wrap
each __u32 in a struct whose constructor does not initialize the value

Fixes: https://tracker.ceph.com/issues/54279
Signed-off-by: Casey Bodley <cbodley@redhat.com>
(cherry picked from commit 4f0ad8aab6b21a1fd57a7c1630d298e31b5d9bb6)

Merge pull request #44897 from aclamk/wip-53480-pacific

pacific: crush: Fix segfault in update_from_hook

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

Merge pull request #44847 from kamoltat/wip-ksirivad-pacific-backport-44241

pacific: mon: osd pool create <pool-name> with --bulk flag

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

Merge pull request #44613 from ifed01/wip-ifed-fix-ram-gridy-fsck-pac

pacific: os/bluestore: make shared blob fsck much less RAM-greedy.

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

Merge pull request #43313 from idryomov/wip-keyring-resolve-error-pacific

pacific: auth,mon: don't log "unable to find a keyring" error when key is given

Reviewed-by: Deepika Upadhyay <dupadhya@redhat.com>

rbd-mirror: make mirror properly detect pool replayer needs restart

When a PoolReplayer detects remote pool metadata change it
sets "stopping" flag expecting the Mirror will restart it.

Although setting "stopping" flag makes the PoolReplayer::run
thread to terminate, the thread's is_started function will still
return true until join is called (and reset the thread id).

This made impossible for the Mirror to detect (by calling
PoolReplayer::is_running) that the PoolReplayer needed restart.

Fixes: https://tracker.ceph.com/issues/54258
Signed-off-by: Mykola Golub <mgolub@suse.com>
(cherry picked from commit ad4a2990b87834fe4ae8c9111547d071aa6e75e5)

Merge pull request #44975 from vshankar/wip-54241

pacific: mds: ignore unknown client op when tracking op latency

Reviewed-by: Xiubo Li <xiubli@redhat.com>
Reviewed-by: Jeff Layton <jlayton@redhat.com>

Merge pull request #45100 from vshankar/wip-54217

pacific: mds: kill session when mds do ms_handle_remote_reset

Reviewed-by: Xiubo Li <xiubli@redhat.com>

Merge pull request #45099 from vshankar/wip-54220

pacific: mds: fix seg fault in expire_recursive

Reviewed-by: Nikhilkumar Shelke <nshelke@redhat.com>

Merge pull request #45016 from Vicente-Cheng/wip-54194-pacific

pacific: mds: mds_oft_prefetch_dirfrags default to false

Reviewed-by: Venky Shankar vshankar@redhat.com

Merge pull request #44910 from batrick/i54161

pacific: mon/MDSMonitor: sanity assert when inline data turned on in MDSMap from v16.2.4 -> v16.2.[567]

Reviewed-by: Venky Shankar vshankar@redhat.com

Merge pull request #44504 from MrFreezeex/wip-53761-pacific

pacific: qa/suites/fs: add prefetch_dirfrags false to thrasher suite

Reviewed-by: Venky Shankar vshankar@redhat.com

Merge pull request #42932 from kotreshhr/wip-52384-pacific

pacific: mgr/volumes: Subvolume removal and clone failure fixes

Reviewed-by: Venky Shankar vshankar@redhat.com

Merge pull request #44869 from neha-ojha/wip-mgr-fixes-pacific

pacific: mgr: various fixes for mgr scalability

Reviewed-by: Pere Diaz Bou <pdiazbou@redhat.com>
Reviewed-by: Adam King adking@redhat.com
Reviewed-by: Nizamudeen A <nia@redhat.com>

mgr/devicehealth: skip null pages when extracting wear level

Some devices have null pages in their ata_device_statistics struct; skip
those pages in order to avoid an AttributeError when extracting device's
wear level.

Fixes: https://tracker.ceph.com/issues/51554
Signed-off-by: Yaarit Hatuka <yaarit@redhat.com>
(cherry picked from commit 2864ac30d4170ba7b5f60ae01ecfdeee707e026a)

ceph-volume: abort when passed devices have partitions

ceph-volume doesn't prevent from using db and/or wal devices
with existing partitions on them.
This can lead to a data loss situation.

Fixes: https://tracker.ceph.com/issues/54376
Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
(cherry picked from commit 75c91a8c6f37a38d69d5da8b1e7d49d9c636230b)

Merge pull request #45123 from zdover23/wip-hardware-recs-backport-2022-02-23-pacific

doc/start: remove journal info from hardware recs

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

Merge pull request #45075 from chrisphoffman/wip-54298-pacific

pacific: cls/rbd: GroupSnapshotNamespace comparator violates ordering rules

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

Merge pull request #45072 from idryomov/wip-qemu-task-rbd-package-pacific

pacific: qa/tasks/qemu: make sure block-rbd.so is installed

Reviewed-by: Deepika Upadhyay <dupadhya@redhat.com>

Merge pull request #44629 from sebastian-philipp/pacific-backport-44367

pacific: mgr/cephadm: Fix count for OSDs with OSD specs

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

Merge pull request #44632 from sebastian-philipp/pacific-backport-44306

pacific: cephadm: deal with ambiguity within normalize_image_digest

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

doc/start: remove journal info from hardware recs

This PR removes mentions of journaling from the hardware
recommendations.

Journaling was a FileStore-related practice. BlueStore is
the default backend for Ceph OSDs and has been since
Luminous. The documentation should reflect that.

Signed-off-by: Zac Dover <zac.dover@gmail.com>
(cherry picked from commit 276bbd8f2b516993813fedfdfd1371e87bb0d44a)

mgr/cephadm: Fix count for OSDs with OSD specs

osd counting is special

Signed-off-by: Sebastian Wagner <sewagner@redhat.com>
Conflicts:
src/pybind/mgr/cephadm/module.py
src/pybind/mgr/cephadm/tests/test_cephadm.py

mgr/dashboard: Contact Info should be visible only when Ident channel is checked

Fixes:https://tracker.ceph.com/issues/54133
Signed-off-by: Sarthak0702 <sarthak.0702@gmail.com>
(cherry picked from commit 15211a6378a6fee9316f79ba0b27821891527c38)

 Conflicts:
src/pybind/mgr/dashboard/frontend/src/app/ceph/cluster/telemetry/telemetry.component.spec.ts
- removed test for perf channel, which doesn't exist in Pacific

mgr/dashboard: telemetry activate: show ident fields when checked

Signed-off-by: Aaryan Porwal <aaryanporwal2233@gmail.com>
(cherry picked from commit ad5b3f200529fc0bc511ce99eed338afcaef6a62)

mgr/dashboard: dashboard turns telemetry off when configuring report

Signed-off-by: Sarthak0702 <sarthak.0702@gmail.com>
(cherry picked from commit 97c57adf8565756dbf24f3c46ed3916303903fb7)

Conflicts:
src/pybind/mgr/dashboard/frontend/src/app/ceph/cluster/telemetry/telemetry.component.spec.ts
- removed tests for perf channel, which doesn't exist in Pacific

Merge pull request #44996 from pcfe/patch-1

pacific: doc/cephadm: Add CentOS Stream install instructions

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

cephadm: count number of docker.io occurrences.

We have to be super careful adding any new hardcoded
regstistry names to the code base. Let's make this an informed
decision next time.

Signed-off-by: Sebastian Wagner <sewagner@redhat.com>
(cherry picked from commit 0fe2e54db774271e4fc18b45aba36b66cbc71779)

mgr/cephadm: improve doc strings for ambigous config values

Signed-off-by: Sebastian Wagner <sewagner@redhat.com>
(cherry picked from commit 391ddf15530d4063d68b03dce1da9999490930d6)

cephadm: sync normalize_image_digest with mgr/cephadm

Signed-off-by: Sebastian Wagner <sewagner@redhat.com>
(cherry picked from commit 84bef14999e0422cf3a2c456c8972211303f91f1)

mgr/cephadm: normalize_image_digest: reworked

Reworked to only care about known short names. I coudn't
find any algorithm, as this problem is ambiguous.

Fixes: https://tracker.ceph.com/issues/53594
Signed-off-by: Sebastian Wagner <sewagner@redhat.com>
(cherry picked from commit a48b8344bd9ebb47c23e589f2c7ecdd18a6f3085)

mgr/cephadm: normalize_image_digest: comment to tests

`pytest --doctest` ftw. Be aware those tests actually fail, as the
implementation contradicts the function documentation.

Signed-off-by: Sebastian Wagner <sewagner@redhat.com>
(cherry picked from commit dc19e3a867b015d8f4c6f5128603d71be3ce5b0d)

Merge pull request #44882 from adk3798/pacific-validate-yumdnf-baseurl

pacific: cephadm: validate that the constructed YumDnf baseurl is usable

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

Merge pull request #44880 from adk3798/pacific-shared-ceph-ceph-volume

pacific: cephadm: add shared_ceph_folder opt to ceph-volume subcommand

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

Merge pull request #44877 from adk3798/pacific-parse-ipv6-route

pacific: cephadm: _parse_ipv6_route: Fix parsing ifs w/o route

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

Merge pull request #44661 from sebastian-philipp/pacific-backport-44600

pacific: mgr/cephadm: auto-enable mirroring module when deploying service

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

Merge pull request #44658 from sebastian-philipp/pacific-backport-43889

pacific: mgr/cephadm: store contianer registry credentials in config-key

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

Merge pull request #45046 from mgfritch/backport-44905-pacific

pacific: cephadm: chown the prometheus data dir during redeploy

Reviewed-by: Adam King <adking@redhat.com>
Reviewed-by: Patrick Seidensal <pseidensal@suse.com>

Merge pull request #44742 from idryomov/wip-rbd-switch-arguments-fix-pacific

pacific: rbd: add missing switch arguments for recognition by get_command_spec()

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

Merge pull request #44690 from MrFreezeex/wip-53938-pacific

pacific: cls/journal: skip disconnected clients when calculating min_commit_position

Reviewed-by: Deepika Upadhyay <dupadhya@redhat.com>
Reviewed-by: Mykola Golub <mgolub@suse.com>

Merge pull request #44094 from MrFreezeex/wip-53386-pacific

pacific: rbd-mirror: make RemoveImmediateUpdate test synchronous

Reviewed-by: Mykola Golub <mgolub@suse.com>
Reviewed-by: Deepika Upadhyay <dupadhya@redhat.com>

Merge pull request #44754 from idryomov/wip-rbd-mirror-delprop-races-pacific

pacific: rbd-mirror: fix races in snapshot-based mirroring deletion propagation

Reviewed-by: Mykola Golub <mgolub@suse.com>
Reviewed-by: Arthur Outhenin-Chalandre <arthur.outhenin-chalandre@cern.ch>

Merge pull request #44998 from idryomov/wip-53419-pacific

pacific: librbd: fix various memory leaks

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

Merge pull request #44999 from idryomov/wip-52522-pacific

pacific: librbd: fix use-after-free on ictx in list_descendants()

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

Merge pull request #45005 from sunnyku/wip-54168-pacific

pacific: mgr/rbd_support: fix schedule remove

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

mds: kill session when mds do ms_handle_remote_reset

if the mds decide to reuse the old connection it will
do reset_session and should  also kill the session
which are open state in MDSDaemon::ms_handle_remote_reset
to prevent the situation client session is stuck in
opening state and never has chance to becaome open.

the root cause is client missed the request_open
reply but the mds session has become open already.
so we should kill the session in mds side and let
mds recreate the session when received the connect
request from client.

Fixes: http://tracker.ceph.com/issues/53911
Signed-off-by: YunfeiGuan <yunfeiguan@xtaotech.com>
(cherry picked from commit 3651deb4e0b0c102adcaddce79ee4e053f033418)

Merge pull request #45008 from idryomov/wip-rbd-help-positional-optional-pacific

pacific: rbd: mark optional positional arguments as such in help output

Reviewed-by: Sunny Kumar <sunkumar@redhat.com>

mds: fix seg fault in expire_recursive

Range-based for should not be used when we are altering the container.
Use iterator explicitly instead.

Fixes: https://tracker.ceph.com/issues/53805
Signed-off-by: 胡玮文 <huww98@outlook.com>
(cherry picked from commit d48a2cf7e2481cf9758f2934464ec6d9c35d898b)

Merge pull request #45044 from votdev/wip-54284-pacific

cls/rbd: GroupSnapshotNamespace comparator violates ordering rules

For

  GroupSnapshotNamespace a(1, "group-1", "snap-2");
  GroupSnapshotNamespace b(1, "group-2", "snap-1");

both a < b and b < a evaluate to true.  This violates STL strict weak
ordering requirements which is a problem because GroupSnapshotNamespace
is used as a key in std::map (ictx->snap_ids at least), etc.

Fixes: https://tracker.ceph.com/issues/49792
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
(cherry picked from commit 830e72ab9d66c8f5703ea27da5249b02dd16ccd0)

qa/suites/rbd: make sure block-rbd.so is installed

Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
(cherry picked from commit 8f0fd0af3da8581c47dc916303615264714a0489)

qa/tasks/qemu: make sure block-rbd.so is installed

Fixes: https://tracker.ceph.com/issues/54286
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
(cherry picked from commit 525ff61cfc8516b4d7bed6f819b00a0b6cb7be0a)

mgr/dashboard: "Please expand your cluster first" shouldn't be shown if cluster is already meaningfully running

This PR will assume that a cluster is already up and fully running. If this should not be the expected behaviour, deployment tools have to set 'INSTALLED' explicitly. Without this assumption it might happen that upgraded and fully running clusters, e.g. Octopus -> Pacific, will show the 'Expand Cluster' on first log in.

cephadm will take care that the bootstrap phase will write the necessary key to show the 'Expand cluster' page.

Fixes: https://tracker.ceph.com/issues/54215
Signed-off-by: Volker Theile <vtheile@suse.com>
(cherry picked from commit 48fff60b63785ec07f71d3e59394b0c08357247c)

Conflicts:
src/cephadm/cephadm

Signed-off-by: Volker Theile <vtheile@suse.com>

mgr/cephadm: add a test for enabling cephfs mirroring module

Add a test that checks that when cephfs mirror service is enabled
the mirroring mgr module gets enabled.

Actually-written-by: Sebastian Wagner <sewagner@redhat.com>
Signed-off-by: John Mulligan <jmulligan@redhat.com>
(cherry picked from commit bcb4fa70f9f739dce3e1c111db0a322804350f9d)

Conflicts:
src/pybind/mgr/cephadm/tests/test_services.py

mgr/cephadm: auto-enable mirroring module when deploying service

Automatically enable the mgr's mirroring module when creating
cephfs-mirror services. This will trigger a mgr respawn.

Fixes: https://tracker.ceph.com/issues/50593
Based roughly on 50dc1d0decb2fbf7b9129bddba940969410be5cd

Signed-off-by: John Mulligan <jmulligan@redhat.com>
(cherry picked from commit e030130fd1d680ebf9d2c6c9f95f44109fbf3036)

mgr/cephadm: store contianer registry credentials in config-key

Signed-off-by: Daniel Pivonka <dpivonka@redhat.com>
(cherry picked from commit 2c29286b370b1079a35e0e0d47c33a7dabab3468)

Conflicts:
src/pybind/mgr/cephadm/module.py
src/pybind/mgr/cephadm/serve.py

Merge pull request #44725 from cfsnyder/wip-53703-pacific

pacific: osdc: add set_error in BufferHead, when split set_error to right

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

Merge pull request #44865 from cbodley/wip-54090

pacific: qa: remove centos8 from supported distros

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

Merge pull request #44545 from cfsnyder/wip-53718-pacific

pacific: osd/OSDMapMapping: fix spurious threadpool timeout errors

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

Merge pull request #44879 from adk3798/pacific-co-locate-doc

pacific: doc/cephadm: Co-location of daemons

Reviewed-by: Michael Fritch mfritch@suse.com

Merge pull request #44146 from k0ste/wip-53435-pacific

pacific: mgr/prometheus: Fix the per method stats exported

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

Merge pull request #44830 from adk3798/pacific-rm-iscsi

pacific: mgr/cephadm/iscsi: use `mon_command` in `post_remove` instead of `check_mon_command`

Reviewed-by: Michael Fritch mfritch@suse.com

Merge pull request #44829 from adk3798/pacific-cpu-limit

pacific: mgr/cephadm: allow miscellaneous container args at service level

Reviewed-by: Michael Fritch mfritch@suse.com