cephadm: Make ceph-iscsi api user and password mandatory

The api user and password is required in order to use the API so let's
make these mandatory. The `ceph orch daemon add iscsi` now has them
mandatory:

  ceph orch daemon add iscsi <pool> <api_user> <api_password>

If your using apply with a yaml file, the validate_add in the spec now
checks for these too.

Signed-off-by: Matthew Oliver <moliver@suse.com>
(cherry picked from commit a36165b8dacaae3ab673a3e79f831a0de659341b)

qa/tasks/mgr: Don't disable cephadm module

Fixes: https://tracker.ceph.com/issues/45596
Signed-off-by: Varsha Rao <varao@redhat.com>
(cherry picked from commit 261c769333582247a149ade406ccffe2dce3be23)

cephadm: Detect stale and then recreate connections

Currently we make and cache connections to nodes during a check_host.
If a cached connection is disconnect from the other end the remoto
connection object doesn't track this, so further checks to the host
fail.

I have pushed up a PR[0] to remoto to add a `has_connection` method to
their `BaseConnection` class, which we now use in this patch to check to
see if the connection is stale. If it is it is then recreated.

There is some monkey patching happening so we can add the required
`has_connection` to remoto in this patch which we can remove as soon as
the other PR have landed and a new version of remoto is released.

[0] https://github.com/alfredodeza/remoto/pull/56
Fixes: https://tracker.ceph.com/issues/45627
Fixes: https://tracker.ceph.com/issues/45032
Signed-off-by: Matthew Oliver <moliver@suse.com>
(cherry picked from commit afabf82395361f5d2209a980578b4b6715d80f3d)

doc: New section in device management

A new section to describe the command used to blink lights in physical devices

Signed-off-by: Juan Miguel Olmo Martínez <jolmomar@redhat.com>
(cherry picked from commit e0257273dd4c83049ab9cedf703f165b63b52d01)

mgr/dashboard: document Prometheus' security model

Fixes: https://tracker.ceph.com/issues/45377
Signed-off-by: Patrick Seidensal <pseidensal@suse.com>
(cherry picked from commit 34765e9215f8a0f1476b09993a0703c38a691e4b)

mgr/orch: add explicit naming tests (MDS / RGW)

Signed-off-by: Michael Fritch <mfritch@suse.com>
(cherry picked from commit 959ee9bb68eb35daa60c8d065f1951abf0cf658d)

mgr/orch: fixup chars before hostname without '.' test

Signed-off-by: Michael Fritch <mfritch@suse.com>
(cherry picked from commit 41c0ccd638dff4f46a5b315a299b4725dbea9c5d)

mgr/orch: allow MDS with explicit naming

Fixes: https://tracker.ceph.com/issues/45617
Signed-off-by: Michael Fritch <mfritch@suse.com>
(cherry picked from commit ac61f51f6a80f1abdde3bb3b548724e43e809039)

mgr/orch: allow RGWs with explicit naming

Signed-off-by: Sebastian Wagner <sebastian.wagner@suse.com>
(cherry picked from commit 3c3a97eaaa9d3939209abe492d1bc7c5ff0f26a8)

qa/cephadm: Add roleless mon names to ctx.cluster.remotes role dict

Fixes: https://tracker.ceph.com/issues/45584
Signed-off-by: Varsha Rao <varao@redhat.com>
(cherry picked from commit 58eaeb53e4c7f0b5ed1131778304f72264f20e46)

Merge pull request #35343 from sebastian-philipp/octopus-qa-py2-filenotfound

octopus: qa/cephadm: FileNotFoundError doesn't exist in py2

Reviewed-by: Georgios Kyratsas <gkyratsas@suse.com>

Merge pull request #35188 from sebastian-philipp/octopus-backport-34745-34658-34814-34778-34703-34773-34856-34898-34887-34820-34891-3490

octopus: Batch backport May (2)

Reviewed-by: Tim Serong <tserong@suse.com>

qa/cephadm: FileNotFoundError doesn't exist in py2

Signed-off-by: Sebastian Wagner <sebastian.wagner@suse.com>
(cherry picked from commit eddb6354e157a653962e68d100c6881b5a0fdc24)

Merge pull request #35246 from yuriw/wip-yuriw-octopus-p2p-octopus

qa/tests: added v15.2.3 to parallel tests

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

Merge pull request #35255 from smithfarm/wip-45578-octopus

octopus: librbd: copy API should not inherit v1 image format by default

Reviewed-by: Mykola Golub <mgolub@mirantis.com>
Reviewed-by: Jason Dillaman <dillaman@redhat.com>

Merge pull request #35254 from smithfarm/wip-45643-octopus

octopus: rgw/notifications: version id was not sent in versioned buckets

Reviewed-by: Yuval Lifshitz <yuvalif@yahoo.com>

Merge pull request #34991 from smithfarm/wip-45489-octopus

octopus: rgw: add "rgw-orphan-list" tool and "radosgw-admin bucket radoslist ..."

Reviewed-by: J. Eric Ivancich <ivancich@redhat.com>

Merge pull request #34538 from smithfarm/wip-45074-octopus

octopus: build/ops: address SElinux denials observed in rgw/multisite test run

Reviewed-by: Kefu Chai <kchai@redhat.com>

Merge pull request #35253 from smithfarm/wip-45708-octopus

octopus: mds: preserve ESlaveUpdate logevent until receiving OP_FINISH

Reviewed-by: Patrick Donnelly <pdonnell@redhat.com>
Reviewed-by: Yan, Zheng <zyan@redhat.com>

Merge pull request #35252 from smithfarm/wip-45685-octopus

octopus: mds: take xlock in the order requests start locking

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

Merge pull request #35251 from smithfarm/wip-45678-octopus

octopus: cephfs: allow pool names with hyphen and period

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

Merge pull request #35250 from smithfarm/wip-45674-octopus

octopus: qa: use py3 compat list from range

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

Merge branch 'octopus' of https://github.com/ceph/ceph into wip-yuriw-octopus-p2p-octopus
Changed to v15.2.3

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

Merge branch 'octopus-saved' into octopus

15.2.3

bluestore, common/options.cc: disable bluefs_preextend_wal_files

This is a temporary fix for https://tracker.ceph.com/issues/45613
to avoid WAL corruption because of enabling bluefs_preextend_wal_files
and bluefs_buffered_io together.

Fixes: https://tracker.ceph.com/issues/45613
Signed-off-by: Neha Ojha <nojha@redhat.com>

Merge pull request #35260 from sebastian-philipp/octopus-backport-35235

octopus: qa/cephadm: Add local dockerhub mirror

Reviewed-by: Kefu Chai <kchai@redhat.com>
Reviewed-by: Kyr Shatskyy <kyrylo.shatskyy@suse.com>

qa/cephadm: sudo_write_file doesn't support unicode

Signed-off-by: Sebastian Wagner <sebastian.wagner@suse.com>

qa/cephadm: Docker doesn't ship a registries.conf

I guess not using the mirror for docker based tests is ok for now.

For adding docker support, we need to:

1. change the docker config
2. restart the docker daemon

Note that Docker's config only supports to mirror the docker.io registry.

Signed-off-by: Sebastian Wagner <sebastian.wagner@suse.com>
(cherry picked from commit 9310b6e278c78bdf3d4a6f46e3b17f6837f828ef)

qa/cephadm: Add local registry mirror

Signed-off-by: Sebastian Wagner <sebastian.wagner@suse.com>
(cherry picked from commit d6edd8199f0ad3e34260e3d7eeb05e64ce726ca5)

Conflicts:
qa/tasks/cephadm.py
qa/tox.ini

Changes:
  qa/tox.ini converted to py2

Merge pull request #35150 from Vicente-Cheng/wip-45599-octopus

octopus: client: expose Client::ll_register_callback via libcephfs

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

Merge pull request #35148 from Vicente-Cheng/wip-45603-octopus

octopus: mds: Handle blacklisted error in purge queue

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

Merge pull request #35147 from Vicente-Cheng/wip-45601-octopus

octopus: mds: don't shallow copy when decoding xattr map

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

Merge pull request #34999 from smithfarm/wip-45495-octopus

octopus: cephfs: Client: fix Finisher assert failure

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

Merge pull request #34997 from smithfarm/wip-45477-octopus

octopus: client: only set MClientCaps::FLAG_SYNC when flushing dirty auth caps

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

Merge pull request #34996 from smithfarm/wip-45473-octopus

octopus: bash_completion: Do not auto complete obsolete and hidden cmds

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

Merge pull request #34727 from kotreshhr/octopus-ceph-fs-status-output

octopus: mgr/status: Fix "ceph fs status" json format writing to stderr

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

Merge pull request #34509 from smithfarm/wip-45028-octopus

octopus: mds: don't assert empty io context list when shutting down

Reviewed-by: Yan, Zheng <zyan@redhat.com>
Reviewed-by: Patrick Donnelly <pdonnell@redhat.com>

librbd: copy API should not inherit v1 image format by default

When copying from a v1 image, by default the new destination image
would be created using the v1 format. Since the creation of v1 images
is disallowed, this has been updated to default to using the v2
image format.

Fixes: https://tracker.ceph.com/issues/45518
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit 69b6d5997e8c6a11739e4d5a32564e61eb2f470f)

rgw/notifications: version id was not sent in versioned buckets

Signed-off-by: Yuval Lifshitz <ylifshit@redhat.com>
Fixes: https://tracker.ceph.com/issues/45440
(cherry picked from commit 973b0b1ed8767767f0c4446d97248c1b8002f1d3)

mds: preserve ESlaveUpdate::OP_PREPARE logevent before doing commit

Fixes: https://tracker.ceph.com/issues/45024
Signed-off-by: songxinying <songxinying@sensetime.com>
(cherry picked from commit 4940ab62e0d19ce36e53bcc67b2a2161c47f6c6d)

Conflicts:
src/mds/Mutation.h
- where master has "ceph::buffer::list", octopus has "bufferlist"

mds: take xlock in the order requests start locking

this avoid assertion in MutaionImpl::finish_locking()

Fix: https://tracker.ceph.com/issues/45261
Signed-off-by: "Yan, Zheng" <zyan@redhat.com>
(cherry picked from commit 2e11a35d5b06312e0b2d0aecd83e8eb882ddf719)

cephfs: allow pool names with '-' and '.'

... to be set as pool layouts for files and directories.

Fixes: https://tracker.ceph.com/issues/45521
Signed-off-by: Ramana Raja <rraja@redhat.com>
(cherry picked from commit d9f8f8e3d073de8e1cfd5075159420ca8dfdcf2d)

qa: use py3 compat list from range

Fixes: https://tracker.ceph.com/issues/45590
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit 8d51b33e5d70feec3b65ac3eb819a3515befcdac)

rgw: remove unused function obj_force_ns

Removing the function eliminates a compiler warning.

Signed-off-by: J. Eric Ivancich <ivancich@redhat.com>
(cherry picked from commit 9a1e89173d398bf6e39a431d4b4ef9ab1f20e2bb)

qa/tests: added v15.2.2 to parallel tests

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

cephadm: fixup command_deploy log line

fixup log line to reflect the deploy action
(e.g. deploy, reconfig, redeploy)

Signed-off-by: Michael Fritch <mfritch@suse.com>
(cherry picked from commit 2571d1bd6da97ff852d22d18f6bffb42b2180d23)

cephadm: skip `port_in_use` during redeploy

skip the `port_in_use` check during reconfig/redeploy
of the monitoring components and nfs-ganesha

Signed-off-by: Michael Fritch <mfritch@suse.com>
(cherry picked from commit c85a34aaa70aca55a79e11396b57b31ded14d0b3)

cephadm: remove unused `reconfig` param from create_daemon_dirs

Signed-off-by: Michael Fritch <mfritch@suse.com>
(cherry picked from commit 7a2e4f4a8b2a39877f983eb18dcfbe37fb8dc5f8)

cephadm: don't check port in use during nfs reconfig

Signed-off-by: Michael Fritch <mfritch@suse.com>
(cherry picked from commit 6f4b5ac9dd4839741e1ceb8d48bf33bd6a5d7793)

mgr/cephadm: Fix OSDs not being created

Fixes: https://tracker.ceph.com/issues/45560
This was caused by completions no longer getting
executed synchonously.

Also

* We don't want to have a busy loop when applying OSDs
* Added call to `_apply_all_services` to OSD test

This is a follow-up on f7bf4d0c597080f45bb8466bd6bab3b9cecb03c4

Signed-off-by: Sebastian Wagner <sebastian.wagner@suse.com>
(cherry picked from commit 91c1c6181671719e958ab787f8f3c96850167d82)

doc/cephadm: accessing the admin socket

Signed-off-by: Sebastian Wagner <sebastian.wagner@suse.com>
(cherry picked from commit f0f86ce8bad6a9d6b892cedf9de4a9f36cb717cd)

qa/tasks/cephadm: ignore tarfile.ReadError

there is chance that /var/log/ceph (and other directories) is not
created when cephadm fails, but we always try to collect /var/log/ceph
by tar the directory on the test node and transfer it back for archiving
it. when we fail to do so, an exception is thrown, and this exception is
in turn logged as the cause of the failure instead of the one which
fails the test.

in this change, `tarfile.TarError`s is caught and ignored if it is
thrown by `misc.pull_directory()`, as normally, we don't consider it a fatal
failure when `misc.pull_directory()` fails because of `tar`. the
failures for the test are.

Signed-off-by: Kefu Chai <kchai@redhat.com>
(cherry picked from commit d67b0f1e3ac12d2b38ab6455b5311bcb051acb05)

qa/cephadm: rm ubuntu_latest for now

Fixes: https://tracker.ceph.com/issues/45462
Signed-off-by: Sebastian Wagner <sebastian.wagner@suse.com>
(cherry picked from commit 3742a28415836234a047d4c24c7c2fe910dd3fad)

doc/cephadm: Synopsis: add `cephadm unit`

Signed-off-by: Sebastian Wagner <sebastian.wagner@suse.com>
(cherry picked from commit de278519040a7ff6a407daddfa74678959b7388a)

cephadm: `ls`: also return systemd_unit

Let users directly know the systemd unit name.

Enable things like

```
cephadm ls | jq -r '.[].systemd_unit' | xargs -n 1 systemctl status | cat
```

Signed-off-by: Sebastian Wagner <sebastian.wagner@suse.com>
(cherry picked from commit d8f81540600ee1687030259ed535a390a5b65f09)

qa/tasks/cephadm: Add support for raw disks

Currently tests on raw disks are failing due to stripping '/dev/' from
device name. (ie https://tracker.ceph.com/issues/45383#change-165721)

Signed-off-by: Georgios Kyratsas <gkyratsas@suse.com>
(cherry picked from commit 35e158605ef557671e8feec691fd245b17ea6189)

qa/tasks/cephadm.py: Support bootstrapped clusters

Add the possibility of being able to run cephadm.py on clusters that
were already bootstrapped by some other task/way so you can deploy the
rest of the daemons/services.

Signed-off-by: Georgios Kyratsas <gkyratsas@suse.com>
(cherry picked from commit a002b28bc1aa75e087d850759138763a9a2a3916)

mgr/cephadm: upgrade: use check_mon_command

Signed-off-by: Sebastian Wagner <sebastian.wagner@suse.com>
(cherry picked from commit ec5b8de70a246276d892ab8d1fad7ee1bde1ca2a)

mgr/dashboard: support setting iSCSI gateways from orchestrator

- Dashboard doesn't actively query iSCSI gateways from orchestrator anymore.
  Instead, Orchestrator backend invokes iscsi-gateway-add CLI after any
  iSCSI daemons are deployed.
- Add an optional argument `name` to `iscsi-gateway-add` command. If
  `name` is provided, Dashboard won't issue API request to iSCSI gateway
  to get the hostname. This is required because when orchestrator sets
  the gateway via CLI, the gateway is just spawned and might be not
  ready yet.

Fixes: https://tracker.ceph.com/issues/45163
Signed-off-by: Kiefer Chang <kiefer.chang@suse.com>
(cherry picked from commit 28483b3be1f04b71464cce8f98a526c8670f630d)

mgr/cephadm: config iSCSI gateways in Dashboard

- Add `daemon_check_post` fuction to allow services implement actions
  needed to be done after daemons are checked.
- After iSCSI daemons are spawned, call dashboard CLI to
  set the API URL of the daemons.

Fixes: https://tracker.ceph.com/issues/45163
Signed-off-by: Kiefer Chang <kiefer.chang@suse.com>
(cherry picked from commit 5674f46f0cec041d5a502182066a097b214dd928)

mgr/cephadm: move iSCSI service class to its own file

Signed-off-by: Kiefer Chang <kiefer.chang@suse.com>
(cherry picked from commit 1aea3a0850ea7fcc12c15b39b4bd9039dab2c594)

mgr/orchestrator: adapting orch interface to preview changes

Signed-off-by: Joshua Schmid <jschmid@suse.de>
(cherry picked from commit 15076923684f6933dfb9c20671ecf3234b68b5fc)

mgr/cephadm: adapt cephadm to osdspec preview caching

Signed-off-by: Joshua Schmid <jschmid@suse.de>
(cherry picked from commit 2d4b8c0b791ee74ae98f114018149bd5f74864be)

mgr/cephadm: add preview caching to inventory.py

Signed-off-by: Joshua Schmid <jschmid@suse.de>
(cherry picked from commit 9498b5f77b7c7c32bf3bc0f08617d39b047c4132)

mgr/cephadm: increase global scrape interval

Signed-off-by: Patrick Seidensal <pseidensal@suse.com>
(cherry picked from commit b4a2afa8d2437e7c19cf33863e8db65047e02634)

 mgr/cephadm/nfs: Instead of CACHEINODE use MDCACHE

As CACHEINODE will be deprecated soon.

Signed-off-by: Varsha Rao <varao@redhat.com>
(cherry picked from commit f5e4631dc32bf1b3baa764991450ef6e9a459d14)

cephadm: enable logging to stderr in ceph-iscsi

All the other ceph servies when deployed are deployed with:

  --default-log-to-stderr=true --default-log-stderr-prefix=debug
  --default-log-to-file=false

Ceph-iscsi doesn't have these as commandline params but once
https://github.com/ceph/ceph-iscsi/pull/186 lands it'll have something
similar as config options. This patch sets these options

Signed-off-by: Matthew Oliver <moliver@suse.com>
(cherry picked from commit f65584637f8ee44e79e0976dcecc086d2ba7fe45)

doc/dev/cephadm: Note regarding network calls form CLI handlers

Signed-off-by: Sebastian Wagner <sebastian.wagner@suse.com>
(cherry picked from commit 4985a9ddfca1510fc612a6d115dc00e8bdba229d)

Incorporating sebastian-philipp's suggestions.

Signed-off-by: Zac Dover <zac.dover@gmail.com>
(cherry picked from commit 845b05419ae870f87c66dcef4858d68e2ec8dd5c)

Add ceph orch zap command

Signed-off-by: Zac Dover <zac.dover@gmail.com>
(cherry picked from commit 0455c0e608a02266e8f98fef1643f03d9293ad0a)

mgr/cephadm: disable env_var passing until pr#34835 is merged

Signed-off-by: Joshua Schmid <jschmid@suse.de>
(cherry picked from commit 8f275834ff4a5adcda165ea39e263840e5dd139a)

mgr/cephadm: generate and pass env vars to cephadm

Signed-off-by: Joshua Schmid <jschmid@suse.de>
(cherry picked from commit ca9aafbccf70af466b841c7ae322410e8f3c0caf)

python-common: do not pass envs after entrypoint

Signed-off-by: Joshua Schmid <jschmid@suse.de>
(cherry picked from commit 726cabc75b30e4163f486ac014d193fa76ebf958)

bin/cephadm: move --env-var handling to global argparser

Signed-off-by: Joshua Schmid <jschmid@suse.de>
(cherry picked from commit 84fdccdbeec72f326270b11be0034dff42a44471)

mgr/cephadm: adapt env_var handling

Signed-off-by: Joshua Schmid <jschmid@suse.de>
(cherry picked from commit 6d7a65f2936c6e92330aeb466b28638c33a80e13)

cephadm: return absolute path from pathify

normalizes the user defined paths a bit better to avoid inconsistent
results with os.path.basename(), os.path.join(), etc.

For example:
$ cephadm shell --mount ~/

Signed-off-by: Michael Fritch <mfritch@suse.com>
(cherry picked from commit 669840abc037fc8c49255a68bebae9a8279eac86)

qa/workunits/cephadm/test_cephadm.sh: add shell mount test

Signed-off-by: Michael Fritch <mfritch@suse.com>
(cherry picked from commit 629a9eeeb54bd0bbbdc4e29dc4473ae08244c20a)

cephadm: fix shell mount for special files

block/char devs, named pipes, sockets etc

Signed-off-by: Michael Fritch <mfritch@suse.com>
(cherry picked from commit 72ae06f117d50009baf1dbddacd0805d419f76bc)

mgr/cephadm: Remove AsyncCompletion

Simplyfy things a lot by not using multiprocessing.pool.

Signed-off-by: Sebastian Wagner <sebastian.wagner@suse.com>
(cherry picked from commit f7bf4d0c597080f45bb8466bd6bab3b9cecb03c4)

mgr/cephadm: Proper error for OSErrors

Previsously, an OSError resulted in

    OSError: cannot send (already closed?)

which is not helpful.

Fixes: https://tracker.ceph.com/issues/45032
Signed-off-by: Sebastian Wagner <sebastian.wagner@suse.com>
(cherry picked from commit 7540ab7ba0921fd48071c31fad8f8d8e4a887be7)

cephadm: Fix error parsing <command> arguments

Optional arguments in the <command> parameter used in:
 <cephadm shell>
 <cephadm enter>
 <cephadm ceph-volume>
are not parsed properly. This fixes this problem.

Signed-off-by: Juan Miguel Olmo Martínez <jolmomar@redhat.com>
(cherry picked from commit ba07fa257cbdcb06037ccbab25f853cc3dfbb1e2)

mgr/cephadm: add some more type checks

Signed-off-by: Sebastian Wagner <sebastian.wagner@suse.com>
(cherry picked from commit 85c812c4ed6a7eec25d0d7749dae376bea003fa8)

mgr/cephadm: cleanup async_map_completion etc.

This breaks the progress integration, but we haven't
used that anyway till now.

Signed-off-by: Sebastian Wagner <sebastian.wagner@suse.com>
(cherry picked from commit 24de3e21b911ca7dd3f2390104b10535e2eb2a6c)

mgr/cephadm: replace async_completion with trivial_completion

Signed-off-by: Sebastian Wagner <sebastian.wagner@suse.com>
(cherry picked from commit 397d062759eeabad36a01ccee8bf0db3dfa1415b)

mgr/cephadm: replace async_map_completion with a simple wrapper

There is no need to wrap everything into completions.

Signed-off-by: Sebastian Wagner <sebastian.wagner@suse.com>
(cherry picked from commit a78f42e225bbc041324dc3e565deb207feddad26)
(cherry picked from commit 4bdbbeb6fd7db670fce4f39e6afac15bfa42ad54)

mgr/cephadm: ceph_default_alerts.yml can be utf-8 encoded

non-ascii chars introduced by 653c3f66823

Fixes: https://tracker.ceph.com/issues/45458
Signed-off-by: Michael Fritch <mfritch@suse.com>
(cherry picked from commit bc26b1ab8cc71e36f6d3e3aa041a7b339d41a1f1)

cephadm: extend systemd stop timeout

daemon stop + unit.poststop might take longer to complete

Fixes: https://tracker.ceph.com/issues/45417
Signed-off-by: Michael Fritch <mfritch@suse.com>
(cherry picked from commit 3af9f0d9b99c8e9f9b9ec0b15f1932e15abf7f79)

pybind/mgr: properly mock the mon store for pytest

Signed-off-by: Sebastian Wagner <sebastian.wagner@suse.com>
(cherry picked from commit fbc089755c3ae22854b284c1927a40204d0133fc)

mgr/cephadm: add extra function for the auth entity

Signed-off-by: Sebastian Wagner <sebastian.wagner@suse.com>
(cherry picked from commit f18523fa219ef6e39902625ed2ae0da47695625a)

mgr/cephadm: node-exporter is not a ceph service

Fixes: https://tracker.ceph.com/issues/45427
Signed-off-by: Sebastian Wagner <sebastian.wagner@suse.com>
(cherry picked from commit f9321ebbb42319aa47a0bac2d29790fa262ca4e9)

mgr/cephadm: move HostAssignment to new module

(speeds up my editor)

Signed-off-by: Sebastian Wagner <sebastian.wagner@suse.com>
(cherry picked from commit 524efe899426b73dc8c665b226a89dfd8c7ae05b)

mgr/cephadm: move upgrade code to upgrade.py

Make it easier for everyone to read the upgrade algorithm.

Signed-off-by: Sebastian Wagner <sebastian.wagner@suse.com>
(cherry picked from commit cedf9dbb2b29a787c92e38f2214cc9881a442d61)

mgr/cephadm: refresh osd config when mons change

OSDs don't have a service spec, so the previous implementation
was not reconfiguring them when the list of mons changed.

Fixes: https://tracker.ceph.com/issues/45393
Signed-off-by: Tim Serong <tserong@suse.com>
(cherry picked from commit e4aa556fd0e2859b3992274b64d40e0f0561c76f)

qa/workunits/cephadm: test grafana adoption

now includes https://github.com/ceph/cephadm-adoption-corpus/commit/af85d49307e1652cd0007bf3074030bf0a63c9b7

Fixes: https://github.com/ceph/cephadm-adoption-corpus/pull/3
Signed-off-by: Sebastian Wagner <sebastian.wagner@suse.com>
(cherry picked from commit b62dd11e3d92ec0770389aadd16b4fcb39832480)

mgr/orch: fix mypy errors

orchestrator/module.py: note: In member "_daemon_add_misc" of class "OrchestratorCli":
orchestrator/module.py:655: error: Incompatible types in assignment (expression has type "PlacementSpec", variable has type "Optional[str]")
orchestrator/module.py:656: error: Argument 1 to "ServiceSpec" has incompatible type "Optional[str]"; expected "str"
orchestrator/module.py:656: error: Argument "placement" to "ServiceSpec" has incompatible type "Optional[str]"; expected "Optional[PlacementSpec]"

orchestrator/module.py: note: In member "_apply_misc" of class "OrchestratorCli":
orchestrator/module.py:881: error: Incompatible types in assignment (expression has type "PlacementSpec", variable has type "Optional[str]")
orchestrator/module.py:882: error: Argument 1 to "ServiceSpec" has incompatible type "Optional[str]"; expected "str"
orchestrator/module.py:882: error: Argument "placement" to "ServiceSpec" has incompatible type "Optional[str]"; expected "Optional[PlacementSpec

Signed-off-by: Michael Fritch <mfritch@suse.com>
(cherry picked from commit e6d7fb67282dfd8d1f614980821f9cb7ff4fdf51)

mgr/orch: add missing `subcluster`, `port`, and `ssl`

subcluster, port, and ssl arguments are missing for `orch apply rgw`

Signed-off-by: Michael Fritch <mfritch@suse.com>
(cherry picked from commit d2066b6ecddc9a4d4dadb68cc2dcfadde38aa94a)

mgr/orch: add type checking for add/apply commands

Signed-off-by: Michael Fritch <mfritch@suse.com>
(cherry picked from commit 02a3762e0e8b0db6727ff5533749f34c5c17763b)

mgr/orch: raise validation error for unexpected inbuf

ceph orch apply mds x -i nfs.yaml
Error ENOENT: unrecognized command -i; -h or --help for usage

Signed-off-by: Michael Fritch <mfritch@suse.com>
(cherry picked from commit 393b1a422020094a661b6935c570fb40aa316fb5)

mgr/orch: remove inbuf logic during daemon add (rgw/iscsi)

Simplifies logic for handling the ServiceSpec

Allows for a ServiceSpec:
$ ceph orch daemon add -i iscsi.yaml

But not when combined with a service_type:
$ ceph orch daemon add iscsi <poolname> -i iscsi.yaml

Signed-off-by: Michael Fritch <mfritch@suse.com>
(cherry picked from commit 601ca2710e871012dbbe3eba6328c9c71e62be3b)