admin/doc-requirements: bump sphinx to 4.4.0

bump sphinx to latest stable. to address following build failure

ERROR: sphinx-autodoc-typehints 1.17.0 has requirement Sphinx>=4, but you'll have sphinx 3.5.4 which is incompatible.
ERROR: sphinx-substitution-extensions 2022.2.16 has requirement sphinx>=4.0.0, but you'll have sphinx 3.5.4 which is incompatible.

also bump bump sphinx-rtd-theme, otherwise we'd have following
build failure:

ERROR: sphinx-rtd-theme 0.5.2 has requirement docutils<0.17, but you'll have docutils 0.17.1 which is incompatible.

Signed-off-by: Kefu Chai <tchaikov@gmail.com>
(cherry picked from commit 0a5fab53b3804be5ef1377a2f35006b8df857d39)

mgr/cephadm: set docstring for shim() methods

this allows the "rpc"ized methods of OrchestratorClientMixin to
have the docstring defined by the original methods.

Signed-off-by: Kefu Chai <tchaikov@gmail.com>
(cherry picked from commit d0db2ae4f946e1a985402640ef8f1733b40e91ef)

mgr/cephadm: add empty line after param list in docstring

this helps to silence the warning from sphinx, like

src/pybind/mgr/orchestrator/_interface.py:docstring of orchestrator._interface.Orchestrator.remove_osds:9: WARNING: Field list ends without a blank line; unexpected unindent.

Signed-off-by: Kefu Chai <tchaikov@gmail.com>
(cherry picked from commit d9b8e38e3dfe8e6eec6d56ee934c4632de46fc68)

doc/conf.py: silence warnings from breathe

breathe calls doxygen for extracting/generating docs from code.
while doxygen complains at seeing undocumented fields/func. these
warnings could fail the sphinx-build command, if it takes warnings
as errors.

in this change, these warnings are silenced.

Signed-off-by: Kefu Chai <tchaikov@gmail.com>
(cherry picked from commit 8891d653198c30f9578499126e1ee9ee67eca04a)

mgr/cephadm: document notes using "note::" directive

so it can be rendered by sphinx in a better way.

Signed-off-by: Kefu Chai <tchaikov@gmail.com>
(cherry picked from commit ba3ccee01b31ef9e39a5016a0ffda18628ec3bc2)

mgr/cephadm: improve the formatting of docstring

add an empty line before a doctest block would help
sphinx to tell where the session starts.

see also https://www.sphinx-doc.org/en/master/usage/restructuredtext/basics.html#doctest-blocks

Signed-off-by: Kefu Chai <tchaikov@gmail.com>
(cherry picked from commit 8685fffdf20eeb4e2068c421e351aa02c48ff860)

mgr/cephadm: use block quote for "typical use"

otherwise sphinx takes "Typical use" and the following line as a
field. see also

https://www.sphinx-doc.org/en/master/usage/restructuredtext/basics.html#field-lists

Signed-off-by: Kefu Chai <tchaikov@gmail.com>
(cherry picked from commit 05798f0cae9afda598f5a154c62fdd24bab9ca30)

Merge pull request #45678 from rhcs-dashboard/wip-54586-pacific

pacific: mgr/dashboard: highlight the search text in cluster logs

Reviewed-by: Avan Thakkar <athakkar@redhat.com>
Reviewed-by: Ernesto Puerta <epuertat@redhat.com>
Reviewed-by: Nizamudeen A <nia@redhat.com>

Merge pull request #45716 from adk3798/pacific-backport-march

Cephadm Pacific Batch Backport March

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

Merge pull request #45632 from adk3798/pacific-ssh-offline

pacific: mgr/cephadm: add keep-alive requests to ssh connections

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

Merge pull request #45785 from ronen-fr/wip-rf-45640-pacific

pacific: osd/scrub: restart snap trimming only after scrubbing is done

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

Merge pull request #45773 from ljflores/wip-53605-pacific

pacific: mgr/telemetry: fix waiting for mgr to warm up

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

Merge pull request #45731 from ronen-fr/wip-rf-42951-pacific

pacific: osd/scrub: destruct the scrubber shortly before the PG is destructed

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

Merge pull request #45729 from ronen-fr/wip-rf-42479-pacific

pacific: osd/scrub: remove reliance of Scrubber objects' logging on the PG

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

Merge pull request #45803 from ljflores/wip-telemetry-cephadm-link

pacific: cephadm: fix broken telemetry documentation link

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

cephadm: fix broken telemetry documentation link

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

Merge pull request #45789 from zdover23/wip-doc-2022-04-06-backport-to-pacific-basic-workflow

doc/dev: s/repostory/repository/ (really)

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

doc/dev: s/reposotory/repository/ (really)

This corrects the heinous misspelling described in the
substitution expression in the title. This misspelling is
all the more egregious because it appears in a title, and
therefore would be used to create links if it had not been
caught.

Signed-off-by: Zac Dover <zac.dover@gmail.com>
(cherry picked from commit 298b446c35d19ce43ede513a802d0655bcbdf82f)

qa/suites/fs: stop looping in mds upgrade test if upgrade failed

Signed-off-by: Adam King <adking@redhat.com>
(cherry picked from commit 37019aad80aec15f9a34301c6051f065eb913e29)

mgr/cephadm: fixing prometheus port handling
Fixes: https://tracker.ceph.com/issues/51072
Signed-off-by: Redouane Kachach <rkachach@redhat.com>
(cherry picked from commit 8eb1397d77dace25f387e88137a1807993a0796d)

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

cephadm: respect --skip-firewalld flag

Fixes: https://tracker.ceph.com/issues/54137
Signed-off-by: Adam King <adking@redhat.com>
(cherry picked from commit d97057f8d7263cce8efc0857e3fe4a10faee30c8)

qa/tasks/cephfs: increase timeout in test_nfs.py

Signed-off-by: Matan Breizman <mbreizma@redhat.com>
(cherry picked from commit 44ad552093b4f0dc21563dd9f804974ade239440)

python-common/drive_group: add extra_container_args to supported features

Should have been added when extending extra container args
to all the services but was missed

Signed-off-by: Adam King <adking@redhat.com>
(cherry picked from commit f036bdaf5a1e5f6b18a9591949be878fea8bb70d)

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

doc/cephadm/operations.rst: fix typos

Signed-off-by: wangyunqing <wangyunqing@inspur.com>
(cherry picked from commit 92eb799a952db4f2fe2290aef56d2f66b8f64802)

mgr/cephadm: check spec host when adding osd
Fixes: https://tracker.ceph.com/issues/47872
Signed-off-by: Redouane Kachach <rkachach@redhat.com>
(cherry picked from commit b87c966697d36ef51f1e62425d77200667e651ae)

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

mgr/cephadm: offline host watcher

To be able to detect if certain offline hosts go
offline quicker. Could be useful for the NFS
HA feature as this requires moving nfs daemons from
offline hosts within 90 seconds.

Signed-off-by: Adam King <adking@redhat.com>
(cherry picked from commit bd9eb596570cfcc7fea793c2b380bc66dd719439)

Conflicts:
src/pybind/mgr/cephadm/module.py
src/pybind/mgr/cephadm/ssh.py
src/pybind/mgr/cephadm/tests/fixtures.py
src/pybind/mgr/cephadm/utils.py

mgr/cephadm: Reschedule nfs daemons from offline hosts

In order to improve nfs availability, if there are other
hosts we can place an nfs daemon on or if there is a host
with a lower rank nfs daemon when a higher rank one is on
an offline host, we should reschedule the nfs daemons

Signed-off-by: Adam King <adking@redhat.com>
(cherry picked from commit 9febc21c14b7ad26e6d811444e7daf0b0a292afb)

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

mgr/cephadm: checking service name before removal
Fixes: https://tracker.ceph.com/issues/54503
Signed-off-by: Redouane Kachach <rkachach@redhat.com>
(cherry picked from commit b26c114c8456941d6cccf7d4355445f21cb373a7)

cephadm: verify config file exists when inferring it

Fixes: https://tracker.ceph.com/issues/54571
Signed-off-by: Adam King <adking@redhat.com>
(cherry picked from commit 1568875a281d56b413e75b244c9c75311cf353a0)

mgr/cephadm: adding HostSpec validation
Fixes: https://tracker.ceph.com/issues/54342
Signed-off-by: Redouane Kachach <rkachach@redhat.com>
(cherry picked from commit 15ba147a2a4cae8ca69437382136d328a1f416f2)

doc/cephadm/adoption.rst: fix typos

Signed-off-by: wangyunqing <wangyunqing@inspur.com>
(cherry picked from commit e4db28f6b294909e0f177e82dbda8cfcc8129846)

cephadm: still set container_image when --no-assimilate-config is provided

Fixes: https://tracker.ceph.com/issues/54141
Signed-off-by: Adam King <adking@redhat.com>
(cherry picked from commit 59d004cb901eb6d84fb6907cb88314fd31b87904)

qa/tasks/cephadm_cases: increase timeouts in test_cli.py

These seem to be failing sometimes but in my testing
sometimes these events are happening a few seconds after
we hit the timeout. Trying to see if this makes the tests
more consistent. No need to mark the test as failed
if we report something up in 34 seconds vs 25 especially
when cephadm works on a cyclic daemon refresh.

Signed-off-by: Adam King <adking@redhat.com>
(cherry picked from commit 46f939f057bd05a885eaf750663310375f9dd929)

Conflicts:
qa/tasks/cephadm_cases/test_cli.py

pacific: osd/scrub: restart snap trimming only after scrubbing is done

Snap trimming that was postponed as the target PG was scrubbing
must be restarted at scrub completion.
PR #38111 moved trimming restart to just before the scrub fully
terminated. The current PR fixes that.

Trimming is also restarted in those cases where scrub was
queued but aborted immediately.

Fixes: https://tracker.ceph.com/issues/52026
Signed-off-by: Ronen Friedman <rfriedma@redhat.com>
(cherry picked from commit 948d3266c67bf896d1c20472977b849178d233d3)

Conflicts:
src/osd/pg_scrubber.cc

Conflict resolved by removing a clear_queued_or_active() call that
was dragged in.

mgr/telemetry: fix waiting for mgr to warm up

1. The implementation of config_notify() in telemetry module sets the
flag for event, which is supposed to wake up the 'serve' thread whenever
a config option is changed. The problem is that we call config_notify()
at the beginning of serve(), before we enter its 'run' loop. This call
sets the event which cancels the 10 seconds wait for the mgr to warm up.
To fix this, we extract the logic of updating the config options to a
separate function (config_update_module_option()), and call it on
__init__, instead of calling config_notify() in serve().

2. We should always wait for the mgr to warm up here (10 seconds). In
case of a sporadic event (e.g. a config option change via CLI) the event
will be set, and wait will return immediately. We enforce this wait by
using time.sleep(10) instead of event.wait(10).

Fixes: https://tracker.ceph.com/issues/53204
Signed-off-by: Yaarit Hatuka <yaarit@redhat.com>
(cherry picked from commit fa5cc0ca081ca3cce552e0cb21a1e17273cf3482)

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

- Several options under __init__ had to be removed that were not present
  in Pacific
- No type checking in Pacific

Merge pull request #45654 from ljflores/wip-pacific-fast-shutdown-backports

Pacific fast shutdown backports

Reviewed-by: Neha Ojha <nojha@redhat.com>
Reviewed-by: Sridhar Seshasayee <sseshasa@redhat.com>
Reviewed-by: Nitzan Mordechai <nmordech@redhat.com>

Merge pull request #45586 from idryomov/wip-pool-reverse-lookup-osdmap-pacific

pacific: librados: check latest osdmap on ENOENT in pool_reverse_lookup()

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

Merge pull request #45753 from ceph/wip-pacific-debug

build: Add some debugging messages

Merge pull request #45638 from idryomov/wip-diff-iterate-striping-fix-pacific

pacific: librbd: make diff-iterate in fast-diff mode sort and merge reported extents

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

build: Add some debugging messages

Having a unique string like "CI_DEBUG" will help me know where we are in the build process in Jenkins logs.

Signed-off-by: David Galloway <dgallowa@redhat.com>
(cherry picked from commit 57edb76ea46893294a70aa080916bc723fb35f9e)

osd/scrub: destruct the scrubber shortly before the PG is destructed

By destructing the scrubber when the PG is still intact, we guarantee that
Scrubber's code can refer to the PG object - especially in dout()s.

Signed-off-by: Ronen Friedman <rfriedma@redhat.com>
(cherry picked from commit bcd13e134c1f335506e425800170d55cd8a2af1b)

osd/scrub: remove reliance of Scrubber objects' logging on the PG

Modify the Scrubber's sub-objects to use their own gen_prefix()
functions, instead of using PG::gen_prefix().

Fixes: https://tracker.ceph.com/issues/51843
Signed-off-by: Ronen Friedman <rfriedma@redhat.com>
(cherry picked from commit 2aeb9263d643b19d59219e8e187e1a0fa0292693)

Conflicts:
        src/osd/PG.h
        src/osd/PrimaryLogScrub.cc
        src/osd/pg_scrubber.cc
        src/osd/pg_scrubber.h

Conflict resolution:
- manually removing some scrub scheduling changes from
  PR #40984
- pg_scrubber.h: removing some irrelevant lines that were dragged
  in.
- PG.h: restoring lines removed by the merge.

mgr/dashboard: Remove padding in search highlighted text

Signed-off-by: Sarthak0702 <sarthak.0702@gmail.com>

Merge pull request #45620 from s0nea/wip-55036-pacific

pacific: mgr/cephadm: try to get FQDN for configuration files

Reviewed-by: Michael Fritch <mfritch@suse.com>
Reviewed-by: Adam King adking@redhat.com

mgr/dashboard: highlight the search text in cluster logs

Fixes: https://tracker.ceph.com/issues/54445
Signed-off-by: Sarthak0702 <sarthak.0702@gmail.com>
(cherry picked from commit a878c7442059d11ac14edd226d71abbabda9a3c4)

Merge pull request #45374 from ronen-fr/wip-rf-42684-pacific

pacific: osd/scrub: tag replica scrub messages to identify stale events

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

Merge pull request #45355 from mgfritch/backport-45347-pacific

pacific: cephadm: preserve `authorized_keys` file during upgrade

Reviewed-by: Adam King adking@redhat.com

Merge pull request #45591 from vumrao/wip-vumrao-55020

pacific: osd/PrimaryLogPG.cc: CEPH_OSD_OP_OMAPRMKEYRANGE should mark omap dirty

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

Merge pull request #45203 from rhcs-dashboard/wip-54113-pacific

pacific: mgr/dashboard: perform daemon actions

Reviewed-by: Pere Diaz Bou <pdiazbou@redhat.com>
Reviewed-by: Aashish Sharma <aasharma@redhat.com>

Merge pull request #45173 from kamoltat/wip-ksirivad-backport-pacific-44054

pacific: osd: add pg_num_max value & pg_num_max reordering

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

Merge pull request #45588 from ljflores/wip-pacific-perfcounter-fix

pacific: common: fix missing name in PriorityCache perf counters

Merge pull request #45561 from idryomov/wip-readv-writev-overflow-pacific

pacific: librbd: readv/writev fix iovecs length computation overflow

Reviewed-by: Christopher Hoffman <choffman@redhat.com>

Merge pull request #45474 from nmshelke/wip-54573-pacific

pacific: mgr/volumes: the 'mode' should honor idempotent subvolume creation

Reviewed-by: Venky Shankar vshankar@redhat.com

Merge pull request #45464 from cfsnyder/wip-53471-pacific

pacific: common: avoid pthread_mutex_unlock twice

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

Merge pull request #45436 from cfsnyder/wip-51783-pacific

pacific: qa/rgw: add failing tempest test to blocklist

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

Merge pull request #45163 from dvanders/wip-52635-pacific

pacific: mds: ensure that we send the btime in cap messages

Reviewed-by: Venky Shankar vshankar@redhat.com

Merge pull request #45565 from ajarr/wip-52875-pacific

pacific: qa: fsync dir for asynchronous creat on stray tests

Reviewed-by: Venky Shankar vshankar@redhat.com

Merge pull request #45564 from ajarr/wip-52427-pacific

pacific: qa: ignore expected metadata cluster log error

Reviewed-by: Venky Shankar vshankar@redhat.com

Merge pull request #44771 from pdvian/wip-53944-pacific

pacific: osd/OSD: Log aggregated slow ops detail to cluster logs

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

Merge pull request #44664 from gregsfortytwo/wip-53933-pacific

pacific: osd: PeeringState: fix selection order in calc_replicated_acting_stretch

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

Merge pull request #45491 from cfsnyder/wip-54077-pacific

pacific: rgw: bucket chown bad memory usage

Merge pull request #45557 from vshankar/wip-53911

pacific: Revert "mds: kill session when mds do ms_handle_remote_reset"

qa/standalone: Fix test_activate_osd() test in ceph-helpers.sh

Modify test_activate_osd() to get the type of scheduler in use and then
verify the value of osd_max_backfills. This is because mclock scheduler
overrides this option to 1000 upon OSD initialization.

The test earlier used to pass because the OSD daemon was killed but not
marked down and upon being brought up, the wait for OSD up check was
passing quickly. But the OSD still didn't have the latest config values.

But now upon killing the OSD, the osd_fast_shutdown sequence notifies the
mon (see PR: https://github.com/ceph/ceph/pull/44807) and is marked down
and dead. Upon bringing it up, the wait for OSD up check takes a longer
time and this is sufficient for the config values to be updated. This
results in the correct values being read from the config 'Values' map.

Signed-off-by: Sridhar Seshasayee <sseshasa@redhat.com>
(cherry picked from commit 3aa2df2e0f6f5bafadc96fd72935e5cf8b2fcf17)

osd/OSD: osd_fast_shutdown_notify_mon not quite right

When osd_fast_shutdown and osd_fast_shutdown_notify_mon set as true, OSD marked as Down
it should be marked as Dead,

Fixed: https://tracker.ceph.com/issues/53327

Signed-off-by: Nitzan Mordechai <nmordech@redhat.com>
nd

nd

(cherry picked from commit 07302d5e41c49c885c9398c1c478638023e3f264)

osd: make osd_fast_shutdown_notify_mon option true by default

osd_fast_shutdown_notify_mon option is false by default. So users suffer
from error log flood, slow ops, and the long I/O timeouts on voluntary OS
shutdown before they are aware of the existence of this option. Let's
make this option true by default.

Fixes: https://tracker.ceph.com/issues/53328
Signed-off-by: Satoru Takeuchi <satoru.takeuchi@gmail.com>
(cherry picked from commit 729a5b85a6586b47d16acbba2cf8e765e498cd65)

 Conflicts:
src/common/options/global.yaml.in
- global.yaml.in does not exist in Pacific; rather,
  these configs were handled in options.cc.

Merge pull request #45489 from cfsnyder/wip-54041-pacific

pacific: rgw: remove bucket API returns NoSuchKey than NoSuchBucket

Merge pull request #45543 from ajarr/rraja-pacific-backport-pr-44795

pacific: mgr/nfs: allow dynamic update of cephfs nfs export

Merge pull request #45502 from cfsnyder/wip-54148-pacific

pacific: rgw: RGWPostObj::execute() may lost data.

Merge pull request #45500 from cfsnyder/wip-54147-pacific

pacific: rgw/admin: fix radosgw-admin datalog list max-entries issue

Merge pull request #45497 from cfsnyder/wip-54092-pacific

pacific: rgwlc:  warn on missing RGW_ATTR_LC

Merge pull request #45495 from cfsnyder/wip-54083-pacific

pacific: librgw: make rgw file handle versioned

Merge pull request #45487 from mchangir/wip-54532-pacific

pacific: mds,client: add new getvxattr op

Merge pull request #45555 from idryomov/wip-rbd-du-validate-pacific

pacific: test/librbd: add test to verify diff_iterate size

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

Merge pull request #44551 from lxbsz/wip-53759

pacific: mds: reset heartbeat in each MDSContext complete()

Reviewed-by: Venky Shankar vshankar@redhat.com

librbd: make diff-iterate in fast-diff mode sort and merge reported extents

Various users, the most notable example being the QEMU driver, assume
that extents are reported in image offset order.

Fixes: https://tracker.ceph.com/issues/53885
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
(cherry picked from commit 85e7075d5f021bd2d11024e6646d74a8a9f96e15)

mgr/cephadm: add keep-alive requests to ssh connections

Fixes: https://tracker.ceph.com/issues/51733
Signed-off-by: Adam King <adking@redhat.com>

Merge pull request #45509 from cfsnyder/wip-54152-pacific

pacific: rgw: in bucket reshard list, clarify new num shards is tentative

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

Merge pull request #45508 from phlogistonjohn/jjm-backport-pacific-pr-44845-44870-45211

pacific: mgr/nfs: support managing exports without orchestration enabled

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

Merge pull request #45569 from mgfritch/backport-45420-pacific

pacific: cephadm: infer the default container image during pull

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

Merge pull request #45582 from idryomov/wip-rbd-nbd-cookie-pacific

pacific: backport nbd cookie support

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

Merge pull request #45580 from idryomov/wip-fix-pids-limit-pacific

pacific: cephadm: Remove containers pids-limit

Reviewed-by: Adam King adking@redhat.com

Merge pull request #45477 from cfsnyder/wip-53637-pacific

pacific: test/rgw: disable cls_rgw_gc test cases with defer_gc()

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

Merge pull request #45476 from cfsnyder/wip-53518-pacific

pacific: rgw: add object null point judging when listing pubsub  topics

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

Merge pull request #45484 from cfsnyder/wip-53835-pacific

pacific: rgw: document S3 bucket replication support

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

Merge pull request #45483 from cfsnyder/wip-53818-pacific

pacific: rgw: RadosBucket::get_bucket_info() updates RGWBucketEnt

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

Merge pull request #45480 from cfsnyder/wip-53654-pacific

pacific: rgw: init bucket index only if putting bucket instance info succeeds

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

Merge pull request #45478 from cfsnyder/wip-53639-pacific

pacific: rgw/amqp: add default case to silence compiler warning

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

Merge pull request #45463 from cfsnyder/wip-53254-pacific

pacific: rgw: fix rgw.none statistics

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

mgr/cephadm: try to get FQDN for inventory address

Fixes: https://tracker.ceph.com/issues/54502
Signed-off-by: Tatjana Dehler <tdehler@suse.com>
(cherry picked from commit 4f14993b1667fff309cd9cd6f9dad638a5a7e502)

Conflicts:
src/pybind/mgr/cephadm/services/monitoring.py
Fixed conflict because ceph#44751 has not been backported to
pacific (yet).

mgr/cephadm: unify way to get the host address

There are two different ways to get the host address. From the
inventory of the mgr object directly or via the `_inventory_get_addr`
method of `CephadmService`. Update the code in order to use the
`_inventory_get_addr` method only.

Signed-off-by: Tatjana Dehler <tdehler@suse.com>
(cherry picked from commit 30385068eed7a4f93179e8d2748dd6e01bba6ffd)

Merge pull request #45596 from neha-ojha/wip-45512-pacific

pacific: ceph/admin: s/master/main

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

Merge pull request #45532 from idryomov/wip-fix-schedule-status-pacific

pacific: mgr/rbd_support: cast pool_id from int to str when collecting LevelSpec

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

rgw: fix rgw.none statistics

Only unaccount for objects being removed if they exist.

Also, clean up some formatting.

Signed-off-by: J. Eric Ivancich <ivancich@redhat.com>
(cherry picked from commit 95b662fa176cd841e26b76341f91bf2e6c266003)

ceph/admin: s/master/main

This PR changes the name "master" to "main" so
that builds (and, I assume, a great many other
things) will not fail.

Signed-off-by: Zac Dover <zac.dover@gmail.com>
(cherry picked from commit 6a1dd3a8a2f3dc9fe8615d402c9041273516ff89)

osd/PrimaryLogPG.cc: CEPH_OSD_OP_OMAPRMKEYRANGE should mark omap dirty

We should mark_omap_dirty() for all omap write ops, just like we did
in cb927925af1f3df4b9c31df85cf31f982aae1988.

Currently, for CEPH_OSD_OP_OMAPRMKEYRANGE ops, clean_omap gets set to true,
which results in incomplete recovery of objects and results in
inconsistent PGs after a scrub.

Fixes: https://tracker.ceph.com/issues/54592
Signed-off-by: Neha Ojha <nojha@redhat.com>
(cherry picked from commit f7fd5895fd3d7d7c4691be91434868d90f7a4e0f)

common: fix missing name in PriorityCache perf counters

There was a problem with PriorityCache perf counters, where part of the name was missing (i.e. "mon.a.cache_bytes" instead of "mon.a.prioritycache.cache_bytes"). The problem was happening because a 'this' pointer was missing in the original implementation.

Signed-off-by: Laura Flores <lflores@redhat.com>
(cherry picked from commit 9e07175b3c9a5c69b8694bc05e0baa26a351582d)

librados: check latest osdmap on ENOENT in pool_reverse_lookup()

Avoid spurious ENOENT errors from rados_pool_reverse_lookup() and
Rados::pool_reverse_lookup().

This makes lookup by id consistent with lookup by name: the latter
has been checking latest osdmap since commit 7e5669b11b14 ("rados: we
need to get the latest osdmap when pool does not exists").

Fixes: https://tracker.ceph.com/issues/54593
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
(cherry picked from commit 1f837e233af32c8a66f88508cde534c361ecfcbc)

mon, pybind/mgr: Add additional debug level logs for pool options

We find that these logs helped with
the debugging process issues like:
https://tracker.ceph.com/issues/54263.

Added debug level logs to `do_set_pool()` in
src/mon/OSDMonitor.cc.

Added debug level logs to `_maybe_adjust()` in
src/pybind/mgr/pg_autoscaler/module.py.

Signed-off-by: Kamoltat <ksirivad@redhat.com>
(cherry picked from commit ee40c2d2431825f36a01108115b3913112e2ef54)

upgrade/pacific-x/parallel: Added mds.a and mds.b

Added mds daemons so that it can create
cephFS pools and set options using
`do_set_pool()` in FSCommand.cc. Such that
we can cover corner cases like that in

https://tracker.ceph.com/issues/54263

Signed-off-by: Kamoltat <ksirivad@redhat.com>
(cherry picked from commit 1bc51f057fa80b3e34d8bac06ea22ea168fb8cf8)

Conflicts:
qa/suites/upgrade/pacific-x/parallel/0-start.yaml
- File doesn't exist in pacific,
- pacific-p2p-parallel/point-to-point-upgrade.yaml
- is already running mds, so we can delete
- pacific-x/parallel/0-start.yaml.

osd/osd_types: reorder pg_num_max

moved `pg_num_max` to be at the end of the
list in src/osd/osd_types.cc and
src/osd/osd_types.h.

Added comments to `opt_mapping` and `pool_opts_t`
about the importance of the order of options
in the list and class.

Fixes: https://tracker.ceph.com/issues/54263
Signed-off-by: Kamoltat <ksirivad@redhat.com>
(cherry picked from commit e44c469f59eaef18ecd3c3b348981939043eae02)