doc: users now need to provide scrub_mdsdir and recursive flags
for mdsdir scrub at CephFS root.

Fixes: https://tracker.ceph.com/issues/59350
Signed-off-by: Dhairya Parmar <dparmar@redhat.com>
(cherry picked from commit 93dfc111bf7e11b3122d2a2df0346b341164dd08)

qa: add recursive flag to test_flag_scrub_mdsdir

Code has been changed, in order to scrub ~mdsdir at root,
recursive flag also needs to be provided along with
scrub_mdsdir.

Fixes: https://tracker.ceph.com/issues/59350
Signed-off-by: Dhairya Parmar <dparmar@redhat.com>
(cherry picked from commit e40ca408a1343754b80ba1e88e0b669689ada164)

mds: remove code to bypass dumping empty header scrub info

Previously when ~mdsdir was scrubbed at CephFS root, it's header
was kept empty, thus it became necessary to not dump it's values
for 'scrub status'. Now since both the scrubs(~mdsdir and root)
run under the same header, this code is no more needed.

Fixes: https://tracker.ceph.com/issues/59350
Signed-off-by: Dhairya Parmar <dparmar@redhat.com>
(cherry picked from commit 548af10f1057fe3fb46e23887b012a664b45b7f8)

mds: dump_values no more needed

Previouly, two individual scrubs were initiated to scrub ~mdsdir
at root where the ~mdsdir scrub wasn't provided any tag thus, it
was necessary to not dump it's values for output of 'scrub start'.
Now since mdsdir and root scrub run under single header, there is
no need for this anymore, thus removing this redundant code.

Fixes: https://tracker.ceph.com/issues/59350
Signed-off-by: Dhairya Parmar <dparmar@redhat.com>
(cherry picked from commit 2f06feea33c88798e4e9d654bbff8b05e77f0681)

mds: enqueue ~mdsdir at the time of enqueing root

This would avoid the need to run individual scrubs for
~mdsdir and root, i.e. run both the scrubs under the
same header, this also helps to avoid edge case where
in case ~mdsdir is huge and it's taking time to scrub it,
the scrub status would report something like this until
root inodes kick in:

{
    "status": "scrub active (757 inodes in the stack)",
    "scrubs": {}
}

Fixes: https://tracker.ceph.com/issues/59350
Signed-off-by: Dhairya Parmar <dparmar@redhat.com>
(cherry picked from commit f85b22818b474911cfdc15e7c6be1d2979939888)

mds: adjust MDSRank::command_tag_path invocation of enqueue_scrub()

Signed-off-by: Dhairya Parmar <dparmar@redhat.com>
(cherry picked from commit f654de977b8dddcbd4a39ffb4a338b549ca02bb0)

doc/scrub: documented stray evaluation using recursive scrub

Also documented the use of new flag introduced in this PR

Fixes: https://tracker.ceph.com/issues/53724
Signed-off-by: Dhairya Parmar <dparmar@redhat.com>
(cherry picked from commit d22e60fa6900449e443a23087b16f0c451d79c2b)

qa: added testcases

- test_stray_evaluation_with_scrub
this assures that evaluating strays with scrub works fine and no
crash is detected.

- test_flag_scrub_mdsdir
test the new flag to scrub ~mdsdir at CephFS root

Fixes: https://tracker.ceph.com/issues/51824
Signed-off-by: Dhairya Parmar <dparmar@redhat.com>
(cherry picked from commit 632c8b04cc91097cda3b35f639e623c2b07abb68)

mds: make `scrub status` print flag `scrub_mdsdir`

and make sure it does not dispose JSON for recursive scrub
on ~mdsdir while scrubbing with flag 'scrub_mdsdir'

Signed-off-by: Dhairya Parmar <dparmar@redhat.com>
(cherry picked from commit b32c572192fea394d1f56fb0e3d0f6cf672d5ec3)

mds: add scrub_mdsdir to ScrubHeader

Signed-off-by: Dhairya Parmar <dparmar@redhat.com>
(cherry picked from commit c85347fefb1b5a965a6640da9cb48d867642a295)

mds: do not dump multiple JSON obj

As enqueue_scrub is invoked twice to perform stray evaluation
at root using new op scrub_mdsdir, tests fail as command
scrub start / scrub_mdsdir would return two JSON object which
JSONDecodeError cannot parse, this patch would make sure we do
not dump JSON for recursive scrub on ~mdsdir than happens along with
scrubbing root (only when using scrub_mdsdir op)

Signed-off-by: Dhairya Parmar <dparmar@redhat.com>
(cherry picked from commit 4a976b5b0110c787b9a8e39360fd41164355036d)

mds: evaluate strays while performing scrub on root path

Running scrub at CephFS root doesn't iterate over ~mdsdir,
this feature is needed but the cluster admin should have
the liberty to decide whether to evaluate strays at root
or not. Therefore a new option 'scrub_mdsdir' can be useful.

Fixes: https://tracker.ceph.com/issues/53724
Signed-off-by: Dhairya Parmar <dparmar@redhat.com>
(cherry picked from commit e65eb2bb7895287c068691afe7a7a1710af0d360)

mds: remove inode from scrub_stack if being purged

it an inode is being purged, scrub should not be ran on it.

Fixes: https://tracker.ceph.com/issues/51824
Signed-off-by: Dhairya Parmar <dparmar@redhat.com>
(cherry picked from commit ce12cd71ffa703c8e3f05617cd8261d4ff820e7e)

mds: do not scrub inode if it is purging

Scrub could evaluate strays but there was a race condition where
an inode pinned for scrub would land for purging which would indeed
lead to crashes.

Fixes: https://tracker.ceph.com/issues/51824
Signed-off-by: Dhairya Parmar <dparmar@redhat.com>
(cherry picked from commit a8680d37821854fb6d8685db85bdf2b658f12981)

Merge pull request #50798 from zdover23/wip-doc-2023-03-31-backport-50674-to-pacific

pacific: doc/rados/ops: edit health-checks.rst (1 of x)

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

doc/rados/ops: edit health-checks.rst (1 of x)

Edit docs/rados/operations/health-checks.rst (1 of x).

https://tracker.ceph.com/issues/58485

Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit cb1e20f879edb82ae619693a4f1f5e88fd0593dd)
(cherry picked from commit bfeadab4cacfea984b24be32f74aae29d750c148)

Merge pull request #50771 from zdover23/wip-doc-2023-03-30-backport-50742-to-pacific

pacific: doc/start: update "notify us" section

doc/start: update "notify us" section

Rewrite the "Notify Us" section in doc/start/documenting-ceph.rst so
that it is written in English sentences.

(cherry picked from commit 613ece24777c4b222054fe9a6d82ba51949c45ea)
Signed-off-by: Zac Dover <zac.dover@proton.me>

Merge pull request #50789 from zdover23/wip-doc-2023-03-30-backport-50764-to-pacific

pacific: doc/start: format procedure in documenting-ceph

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

doc/start: format procedure in documenting-ceph

Organize three steps in a procedure into one step with three sub-steps.

Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit 6f3cd78f6116f3e19a8f6c5e0ebda816067a334e)

Merge pull request #50763 from zdover23/wip-doc-2023-03-30-backport-50744-to-pacific

pacific: doc/rados: add link to ops/health-checks.rst

doc/rados: add link to ops/health-checks.rst

Add an internal navigation link to
doc/rados/operations/health-checks.rst.

Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit f2f968f65bd1f1a6975fd340c66703cea335dcf6)
(cherry picked from commit 1ddc2b410258449e5d89a415a0494da2b8a79060)

Merge pull request #50753 from zdover23/wip-doc-2023-03-30-backport-50750-to-pacific

pacific: doc/rados/operations: Fix typo in erasure-code.rst

doc/rados/operations: Fix typo in erasure-code.rst

Signed-off-by: Sainithin Artham <sai.artham.19cse@bmu.edu.in>
(cherry picked from commit 3269fea2a209928137e55db16328c581988c048a)

Merge pull request #50741 from zdover23/wip-doc-2023-03-29-backport-50723-to-pacific

pacific: doc/start: documenting-ceph - add squash procedure

doc/start: documenting-ceph - add squash procedure

Add a procedure to doc/start/documenting-ceph.rst that explains how to
perform an interactive rebase to squash commits.

Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit 71ee225d7b8bf54bd190a8b1f2422c6837601f16)

Merge pull request #49923 from adk3798/wip-58539-pacific

pacific: mgr/cephadm: increase ingress timeout values

Reviewed-by: Redouane Kachach <rkachach@redhat.com>

Merge pull request #49925 from adk3798/wip-58541-pacific

pacific: mgr/cephadm: support for extra entrypoint args

Reviewed-by: Anthony D'Atri <anthonyeleven@users.noreply.github.com>
Reviewed-by: Redouane Kachach <rkachach@redhat.com>

Merge pull request #50709 from zdover23/wip-doc-2023-03-28-backport-49514-to-pacific

pacific: doc/glossary: link to CephX Config ref

Merge pull request #49632 from lxbsz/wip-58343

pacific: mds: switch submit_mutex to fair mutex for MDLog

doc/glossary: link to CephX Config ref

Link to the CephX Configuration Reference in the "CephX" entry in the
glossary.

Signed-off-by: Zac Dover <zac.dover@gmail.com>
(cherry picked from commit 36a8f27e969d7103975ebfe312a4be6ae9f281f5)

Merge pull request #50703 from zdover23/wip-doc-2023-03-28-backport-50697-to-pacific

pacific: doc/glossary: add "Scrubbing"

doc/glossary: add "Scrubbing"

Add "Scrubbing" 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 4a66819da48cee430ca74830e1a6f63685b6b240)

mgr/cephadm: increate ingress timeout values

This commit increases the 'timeout client' and 'timeout server' from 1s
to 30s. S3cmd multi-part uploads were failing with the shorter timeout.

Fixes: https://tracker.ceph.com/issues/58183
Signed-off-by: Frank Ederveen <frank.ederveen@gmail.com>
(cherry picked from commit 07701912cfe5b4e72d0f63ce0a8782c3f8c3da7e)

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

mgr/cephadm: fix extra container/entrypoint args with spaces

Fixes: https://tracker.ceph.com/issues/57338
Prior, doing extra container args like

- "--cpus"
- "2"

would work fine as the two args would be passed separately and
eventually placed in the final podman/docker run command
with a space between them. However, trying to do something like

- "--cpus 2"

instead would fail, as it would be translated to

--extra-container-args=--cpus 2

causing "2" to be considered its own arg, which cephadm
wouldn't know how to handle. Another way this can cause problems
is listed in the linked tracker. Either way, leaving the spaces
in the args was causing problems, and the simplest way to handle
it seems to be to just split on the original arg on the spaces
into multiple args

Signed-off-by: Adam King <adking@redhat.com>
(cherry picked from commit a161381ba816609b5041d2072e7e6cec39eaa753)

doc/cephadm: docs for extra entrypoint args

Signed-off-by: Adam King <adking@redhat.com>
(cherry picked from commit 8ee799e9cc0800fbedc8d9343aa3207bc9fd429e)

Conflicts:
doc/cephadm/services/index.rst

mgr/cephadm: support for extra entrypoint args

Args specified in the service spec to be added as
args for the entrypoint when we deploy the daemon

Fixes: https://tracker.ceph.com/issues/57944
Signed-off-by: Adam King <adking@redhat.com>
(cherry picked from commit db8e4462c968e1d29b0688efafad3442abb8779c)

Conflicts:
src/cephadm/cephadm
src/pybind/mgr/cephadm/module.py
src/pybind/mgr/cephadm/services/cephadmservice.py
src/pybind/mgr/cephadm/tests/test_cephadm.py
src/pybind/mgr/cephadm/tests/test_services.py

python-common: add extra_entrypoint_args to service specs

To allow users to set extra args to be passed to
the entrypoint in the podman/docker run command
built by cephadm

Signed-off-by: Adam King <adking@redhat.com>
(cherry picked from commit 98f6a9379e2dccbaf599923644b7aaa7200ded1f)

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

Merge pull request #49931 from adk3798/wip-58199-pacific

pacific: mgr/prometheus: fix module crash when trying to collect OSDs metrics

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

Merge pull request #50007 from adk3798/57060-pacific-backport

pacific: mgr/cephadm: Fix how we check if a host belongs to public network

Reviewed-by: Redouane Kachach <rkachach@redhat.com>

Merge pull request #50371 from guits/wip-58909-pacific

pacific: drive_group: fix limit filter in drive_selection.selector

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

Merge pull request #50511 from adk3798/wip-58775-pacific

pacific: mgr/cephadm: fix backends service in haproxy config with multiple nfs of same rank

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

Merge pull request #50512 from adk3798/wip-58882-pacific

pacific: cephadm: set pids-limit unlimited for all ceph daemons

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

Merge pull request #50513 from adk3798/wip-58773-pacific

pacific: mgr/cephadm: try to avoid pull when getting container image info

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

Merge pull request #49867 from vshankar/tr-58573

pacific: mds: account for snapshot items when deciding to split or merge a directory

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

Merge pull request #50514 from adk3798/wip-58777-pacific

pacific: cephadm: Adding poststop actions and setting TimeoutStartSec to 200s

Reviewed-by: Redouane Kachach <rkachach@redhat.com>

Merge pull request #50515 from adk3798/wip-58988-pacific

pacific: mgr/cephadm: don't add mgr into iscsi trusted_ip_list if it's already there

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

Merge pull request #50516 from adk3798/wip-58990-pacific

pacific: mgr/cephadm: be aware of host's shortname and FQDN

Reviewed-by: Redouane Kachach <rkachach@redhat.com>

Merge pull request #50679 from zdover23/wip-doc-2023-03-27-backport-50675-to-pacific

pacific: doc/rados: clean up ops/bluestore-migration.rst

doc/rados: clean up ops/bluestore-migration.rst

Clean up internal links, fix the numbering of a procedure, and implement
Anthony D'Atri's suggestions in
https://github.com/ceph/ceph/pull/50487 and
https://github.com/ceph/ceph/pull/50488.

https://tracker.ceph.com/issues/58485

Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit b28be76d0f9a5d78943fe3c583a234d0b50356f1)

Merge pull request #50673 from zdover23/wip-doc-2023-03-26-backport-50654-to-pacific

pacific: doc/glossary: add "User"

Reviewed-by: Cole Mitchell <cole.mitchell.ceph@gmail.com>

doc/glossary: add "User"

Add "User" to glossary.rst.

Co-authored-by: Anthony D'Atri <anthony.datri@gmail.com>
Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit fd6bfaf3fe6337511dd7ab344c2936f22a5d43bf)

Merge pull request #50663 from zdover23/wip-doc-2023-03-25-backport-50660-to-pacific

pacific: doc/operations: fix prompt in bluestore-migration

doc/operations: fix prompt in bluestore-migration

Fix a single prompt in bluestore-migration.rst.

Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit 5e54641aec134188c8001e212ea66c1a15d5945c)

Merge pull request #50642 from zdover23/wip-doc-2023-03-23-backport-50634-to-pacific

pacific: doc/rados: edit user-management.rst (1 of x)

doc/rados: edit user-management.rst (1 of x)

Edit doc/rados/operations/user-management.rst. (1 of x)

Co-authored-by: Anthony D'Atri <anthony.datri@gmail.com>
Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit 3f90958909815a287279eccb22193ba5d67cfc1d)

Merge pull request #50620 from zdover23/wip-doc-2023-03-22-backport-50606-to-pacific

pacific: doc/rados: line-edit erasure-code.rst

doc/rados: line-edit erasure-code.rst

Line-edit doc/rados/operations/erasure-code.rst.

Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit 2ff9965e02756636f8d33fd96b7c909f2fca7414)

Merge pull request #50591 from zdover23/wip-doc-2023-03-20-backport-50488-to-pacific

pacific: doc/rados: edit operations/bs-migration (2 of x)

Merge pull request #50588 from zdover23/wip-doc-2023-03-20-backport-50487-to-pacific

pacific: doc/rados: edit operations/bs-migration (1 of x)

doc/rados: edit operations/bs-migration (2 of x)

Disambiguate and improve the English language in
doc/rados/operations/bluestore-migration.rst up to but not including the
section called "Whole Host Replacement".

Co-authored-by: Cole Mitchell <cole.mitchell.ceph@gmail.com>
Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit ca803a24c64059023733e21d755edb9c6c973ecf)

doc/rados: edit operations/bs-migration (1 of x)

Disambiguate and improve the English language in
doc/rados/operations/bluestore-migration.rst up to but not including the
section called "Whole Host Replacement".

Co-authored-by: Cole Mitchell <cole.mitchell.ceph@gmail.com>
Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit 96166b463b1f8800e59f80faebfba1208248725e)

Merge pull request #50583 from zdover23/wip-doc-2023-03-19-backport-50579-to-pacific

pacific: doc/rados: line edit mon-lookup-dns top matter

doc/rados: line edit mon-lookup-dns top matter

Improve the syntax of the top matter of
doc/rados/operations/mon-lookup-dns.rst.

Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit ed5919964bd914c118f556baf3c517e5e35df277)

Merge pull request #50567 from zdover23/wip-doc-2023-03-17-backport-50563-to-pacific

pacific: doc/rados: line-edit upmap.rst

Reviewed-by: Cole Mitchell <cole.mitchell.ceph@gmail.com>

doc/rados: line-edit upmap.rst

Edit all of doc/rados/upmap.rst.

Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit a4a681846499c8d0041a6632e7e31fa83e22ed0e)

Merge pull request #50527 from epuertat/wip-59052-pacific

pacific: mgr/dashboard: force TLS 1.3

Reviewed-by: Avan Thakkar <athakkar@redhat.com>

Merge pull request #47726 from neesingh-rh/wip-57201-pacific

pacific: mgr/snap_schedule: replace .snap with the client configured snap dir name

Reviewed-by: Venky Shankar <vshankar@redhat.com>
Reviewed-by: Kotresh HR <khiremat@redhat.com>
Reviewed-by: Xiubo Li <xiubli@redhat.com>

Merge pull request #49829 from ivancich/wip-58542-pacific

pacific: rgw: fix the problem of duplicate idx when bi list

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

Merge pull request #49669 from vshankar/wip-58349

pacific: mds: fix scan_stray_dir not reset next.frag on each run of stray inode

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

Merge pull request #49656 from dparmar18/wip-58254-pacific

pacific: mds/PurgeQueue: don't consider filer_max_purge_ops when _calculate_ops

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

Merge pull request #49538 from ifed01/wip-ifed-fix-xlock-pac

pacific: mds: do not acquire xlock in xlockdone state

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

Merge pull request #49458 from mchangir/wip-57761-pacific

pacific: qa: wait for scrub to finish

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

Merge pull request #49440 from neesingh-rh/wip-58250-pacific

pacific: mds: avoid ~mdsdir's scrubbing and reporting damage health status

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

Merge pull request #49303 from neesingh-rh/wip-57971-pacific

pacific: cephfs-top: addition of sort feature and limit option

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

Merge pull request #50050 from mchangir/wip-58598-pacific

pacific: mon: disable snap id allocation for fsmap pools

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

Merge pull request #50029 from joscollin/wip-58668-pacific

pacific: cephfs-top: drop curses.A_ITALIC

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

Merge pull request #49966 from vshankar/tr-57728

pacific: pybind/mgr: check for empty metadata mgr_module:get_metadata()

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

Merge pull request #49956 from dparmar18/wip-58601-pacific

pacific: mds/Server: do not allow -ve reclaim flags to cause client eviction

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

Merge pull request #49944 from lxbsz/wip-58603

pacific: ceph_fuse: retry the test_dentry_handling if fails

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

Merge pull request #49941 from lxbsz/wip-58608

pacific: mds: fix and skip submitting invalid osd request

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

Merge pull request #49671 from vshankar/wip-58346

pacific: mds: md_log_replay thread blocks waiting to be woken up

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

mgr/dashboard: force TLS 1.3

Fixes: https://tracker.ceph.com/issues/58942
Signed-off-by: Ernesto Puerta <epuertat@redhat.com>
(cherry picked from commit cd8946631b222dfff1727083d75dfe2dd825d1a4)

mgr/cephadm: be aware of host's shortname and FQDN

The idea is to gether the shortname and FQDN as part
of gather-facts, and then if we ever try to check if a certain
host is in our internal inventory by hostname, we can check
these other known names. This should avoid issues where
we think a hostname specified by FQDN is not in our
inventory because we know the host by the shortname
or vice versa.

Fixes: https://tracker.ceph.com/issues/58738
Signed-off-by: Adam King <adking@redhat.com>
(cherry picked from commit 6443cf15d54a5e50c245dd08c3db005bb8521b6a)

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

mgr/cephadm: don't add mgr into iscsi trusted_ip_list if it's already there

Fixes: https://tracker.ceph.com/issues/58792
Signed-off-by: Mykola Golub <mykola.golub@clyso.com>
(cherry picked from commit 4a74a020a7754f0dac533a50e085f5690df90798)

cephadm: Adding poststop actions and setting TimeoutStartSec to 200s
Fixes: https://tracker.ceph.com/issues/58241
Fixes: https://tracker.ceph.com/issues/58242
Signed-off-by: Redouane Kachach <rkachach@redhat.com>
(cherry picked from commit 5824b0131f6c1959116b3de6b2f2322d95a86bc6)

Conflicts:
src/cephadm/cephadm

mgr/cephadm: try to avoid pull when getting container image info

only if use_repo_digest is not set.

The commit ac88200 introduced this possibility to skip pulling, but
doing this unconditionally broke a use case when one was able to have
a ceph image on a floating tag, and was able to upgrade to a new image
pushed to that tag. As using a floating tag is possible only when
use_repo_digest is enabled (the default), now skipping the pull
if use_repo_digest is disabled will not break it anymore.

Signed-off-by: Mykola Golub <mykola.golub@clyso.com>
(cherry picked from commit 4d049880a0effa78b69179694430d8a274829342)

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

mgr/cephadm: fix error_ok not being passed in unit tests

Signed-off-by: Adam King <adking@redhat.com>
(cherry picked from commit d9292afeddee11da87f7dce38edc34e9442a4368)

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

mgr/cephadm: try to avoid pull when getting container image info

Fixes: https://tracker.ceph.com/issues/58051
Signed-off-by: Mykola Golub <mykola.golub@clyso.com>
(cherry picked from commit ac882006e8bca484ad5dfe93e7374ffec07ef919)

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

mgr/cephadm: allow _run_cephadm_json to be silent on error

Signed-off-by: Mykola Golub <mykola.golub@clyso.com>
(cherry picked from commit ad3f576dde1fdc64a94d50f648052f42f5338fef)

Conflicts:
src/pybind/mgr/cephadm/serve.py
src/pybind/mgr/cephadm/tests/test_cephadm.py

qa/distros: pass --allowerasing --nobest when installing container-tools

One of the tests in the orch suite is running distro install
commands from multiple distros, causing it to first install
container-tools 3.0 and then later install container-tools,
which fails, causing the test to fail. This is sort of a bandaid
fix to getthe test to work. It will cause whatever the last
version of the package to be installed to end up being installed
(and will do so without error) which is what we want in the tests.

Fixes: https://tracker.ceph.com/issues/57771
Signed-off-by: Adam King <adking@redhat.com>
(cherry picked from commit 3011d954986e38ae8f7c9cd027ef2a88dff9a3d8)

Conflicts:
qa/distros/container-hosts/centos_8.stream_container_tools.yaml
qa/distros/container-hosts/centos_8.stream_container_tools.yaml~3011d954986 (qa_distros: pass --allowerasing --nobest when installing container-tools)
qa/distros/container-hosts/centos_8.stream_container_tools_crun.yaml
qa/distros/container-hosts/rhel_8.6_container_tools_rhel8.yaml
qa/distros/podman/rhel_8.6_container_tools_3.0.yaml

cephadm: set pids-limit unlimited for all ceph daemons

We actually had this setup before, but ran into issues.
Some teuthology test had failed in the fs suite, so it was
modified to only affect iscsi and rgw daemons (https://github.com/ceph/ceph/pull/45798)
and then the changes were reverted entirely (so no pids-limit
modifying code at all) in quincy and pacific because
the LRC ran into issues with the change related to the podman
version (https://github.com/ceph/ceph/pull/45932). This new patch
now addresses the podman versions, specifically that the patch
that makes -1 work for a pids-limit seems to have landed in
podman 3.4.1 based on https://github.com/containers/podman/pull/12040.
We'll need to make sure that this doesn't break anything in the
fs suites again as I don't remember the details of the first
issue, or why having it only set the pids-limit for iscsi and rgw fixes it.
Assuming that isn't a problem we should hopefully be able to unify
at least how reef and quincy handle this now that the podman version
issue is being addressed in this patch.

See the linked tracker issue for a discussion on why we're going at
this again and why I'm trying to do this for all ceph daemon types.

Fixes: https://tracker.ceph.com/issues/58685
Signed-off-by: Adam King <adking@redhat.com>
(cherry picked from commit 84c988dcfd1db20e2eca478e9bf89899b2029c5c)

Conflicts:
src/cephadm/cephadm

cephadm: only apply pids-limit to iscsi and rgw

Signed-off-by: Adam King <adking@redhat.com>
(cherry picked from commit 9c8f3942d5154e927d511bbcc9a0275a3dcecf05)

cephadm: remove containers pids-limit

The default pids-limit (docker 4096/podman 2048) prevent some
customization from working (http threads on RGW) or limits the number
of luns per iscsi target.

Fixes: https://tracker.ceph.com/issues/52898
Signed-off-by: Teoman ONAY <tonay@redhat.com>
(cherry picked from commit de8b3c2676e65eb61df54c65cfd3b3af1e68da56)

mgr/cephadm: fix haproxy nfs backend server ip gathering

Fixes: https://tracker.ceph.com/issues/58465
Previously, if there were 2 nfs daemons of the same
rank, we could not check the rank generation, which
is intended to mark which one is the "real" on of that
rank in cases where we cannot remove the other one due
to its host being offline. The nfs of a given rank with
the highest rank_generation is the one we want haproxy
to use for its backend IP. Since we didn't actually
check this, it was random, depending on what order we
happened to iterate over the nfs daemons of the same
rank, which IP we actually got. If the nfs with the
lower rank_generation on an offline host happened
to come later in the iterations, we'd use that one
for the IP, which is incorrect.

Signed-off-by: Adam King <adking@redhat.com>
(cherry picked from commit e0dd6cac0229e75feb85b10bb31a70e4638733cc)

mgr/cephadm: don't attempt daemon actions for daemons on offline hosts

They'll just fail anyway, and it will waste time waiting
for the connection to timeout. We have other places in
the serve loop that will check if the host is back
online.

Signed-off-by: Adam King <adking@redhat.com>
(cherry picked from commit e1d3e38f87bce4e53eb7e4c29c812a7ed750fa80)

Merge pull request #50243 from cbodley/wip-58767

pacific: cls/rgw: remove index entry after cancelling last racing delete op

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

Merge pull request #50000 from smanjara/wip-57701-pacific

pacific: qa/suites/rgw: fix and update tempest and barbican tests

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

Merge pull request #49841 from k0ste/wip-58553-pacific

pacific: rgw/beast: fix interaction between keepalive and 100-continue

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

Merge pull request #49809 from votdev/issue_58470_empty_tags

pacific: rgw: an empty tagset is allowed by S3

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