src/test: Removing unrecognized option from bootstrap file

Removing '--distribute' option from a command in bootstrap file which seemed to be the reason of failure in:
http://qa-proxy.ceph.com/teuthology/dang-2021-04-19_12:28:43-rgw-wip-dang-zipper-sysobj-distro-basic-smithi/6059513/teuthology.log

Signed-off-by: Kalpesh Pandya <kapandya@redhat.com>

Merge pull request #40911 from tchaikov/wip-doc-confval

doc/rados/configuration/auth-config-ref: drop section of keyrings

Reviewed-by: Sage Weil <sage@redhat.com>

doc/man/8/ceph-authtool: fix the indent of option

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

Merge pull request #40899 from malcolmholmes/malcolmholmes/fix-json-dashboard-error

mgr/Dashboard: Remove erroneous elements in hosts-overview Grafana dashboard

Reviewed-by: Avan Thakkar <athakkar@redhat.com>
Reviewed-by: Ernesto Puerta <epuertat@redhat.com>
Reviewed-by: p-se <NOT@FOUND>

Merge pull request #40898 from rzarzynski/wip-crimson-only-mgrclt-noconn-report

crimson/mgr: don't report if there is no connection available.

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

doc/rados/configuration/mon-config-ref:  use confval directive

for defining options

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

doc/rados/configuration/auth-config-ref: drop section of keyrings

as per Sage, the caps are out of date (should be based on profiles) and
the keyring locations aren't useful information

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

Merge pull request #40897 from tchaikov/wip-doc-more-conf-vals

doc/rados/configuration: use confval directive to define options.

Reviewed-by: Reviewed-by: Sage Weil <sage@redhat.com>

Merge pull request #40905 from idryomov/wip-rbd-kms-config-map-doc

doc/rbd/rbd-kubernetes: update for recent ceph-csi versions

Reviewed-by: Madhu Rajanna <madhupr007@gmail.com>
Reviewed-by: Sage Weil <sage@redhat.com>

Merge pull request #40378 from zdover23/wip-doc-cephadm-osd-declarative-state-rewrite-2021-Mar-25

doc/cephadm: rewrite "delcarative state" in osd.rst

Reviewed-by: Sage Weil <sweil@redhat.com>

Merge PR #40817 into master

* refs/pull/40817/head:
mgr/MgrStandby: fix config observer
mgr/MgrStandby: respawn if mgr_standby_modules changes
qa/tasks/mgr/test_dashboard: skip test_standby if mgr_standby_modules=false
qa/suites/rados/cephadm/smoke-*: use cephadm.wait_for_service
qa/suites/rados/cephadm/smoke-singlehost: test --single-host-defaults
cephadm: add --single-host-defaults option to bootstrap
mgr/cephadm: allow mgr colo if mgr_standby_modules=false
mgr/MgrStandby: add mgr_standby_modules option

Reviewed-by: Sebastian Wagner <swagner@suse.com>

Merge PR #40893 into master

* refs/pull/40893/head:
mgr/cephadm: don't list non ceph daemons as needing upgrade in upgrade check

Reviewed-by: Sage Weil <sage@redhat.com>

Merge pull request #37730 from thotz/rgwkmsvaultsslsupport

rgw: extending existing ssl support for vault KMS

Merge pull request #40881 from tchaikov/wip-50393

qa/workunits/mon/test_mon_config_key: use subprocess.run() instead of…

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

doc/rbd/rbd-kubernetes: expose volume expansion functionality

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

doc/rbd/rbd-kubernetes: imageFeatures is required

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

doc/rbd/rbd-kubernetes: mention KMS config map

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

doc/rados/configuration/mon-config-ref:  use confval directive

for defining options

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

doc/rados/configuration/msgr2: use confval directive

for defining options

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

doc/rados/configuration/auth-config-ref: use "describe" to define default keyrings

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

doc/_ext: use iec_size() instead of filesizeformat()

to format size options in the same format supported by our C++
strict_iec_cast() parser. so they are more consistent from user's
perspective.

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

doc/rados/configuration: merge ms-ref into network-conf-ref

the latter provides more context, and "messenger" is how the networking
infra is implemented.

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

doc/rados/configuration/ceph-conf: use "describe" to define option types

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

doc/rados/configuration: fix formatting of code block

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

doc: use "confsec" role for configure section

and rename confval_section to confsec for shorter name

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

doc/rados/configuration/ceph-conf: use confval directive

for defining options

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

doc: use ":confval:" role to reference options defined using confval

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

doc/rados/configuration/osd-config-ref: use confval directive

for defining options

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

doc/_ext: render see_also field

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

doc/_ext: render more fields in an option

* print "enum_value" as a list
* print "note" and "warning"

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

doc/rados/configuration/mclock-config-ref: use confval directive

for defining options

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

doc/rados/configuration/pool-pg-config-ref: use confval directive

for defining options.

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

doc/rados/configuration/bluestore-config-ref: use confval directive

for defining options.

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

Merge pull request #40906 from tchaikov/rzarzynski/wip-crimson-internal_client_request-cleanup

crimson/osd: various cleanups

Reviewed-by: Radoslaw Zarzynski <rzarzyns@redhat.com>

Merge pull request #40874 from rhcs-dashboard/fix-cephfs-pool-size-stat

mgr/dashboard: filesystem pool size should use stored stat

Reviewed-by: Aashish Sharma <aasharma@redhat.com>

Merge pull request #40879 from myoungwon/wip-fix-50299

osd: fix reference leak when ManifestOp is not used

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

Merge pull request #40878 from tchaikov/wip-mop-cleanup

osd: remove C_SetManifestRefCountDone::mop

Reviewed-by: Myoungwon Oh <myoungwon.oh@samsung.com>
Reviewed-by: Sam Just <sjust@redhat.com>

Merge pull request #40896 from jdurgin/wip-yaml-options-fix

common/options: fix several out of date defaults and options added during yaml conversion

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

Merge pull request #40884 from majianpeng/osd-watch-remove-unused-code

osd/Watch.h: remove unused code.

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

qa/workunits/mon/test_mon_config_key: use subprocess.run() instead of proc.communicate()

the loop of proc.communicate() on python3.6, where we always are able to
get something out of stdout and/or stderr PIPEs. and the `stdout` and
`stderr` keep growing until out of memory. and teuthology considers
the command crashed after a while.

Fixes: https://tracker.ceph.com/issues/50393
Signed-off-by: Kefu Chai <kchai@redhat.com>

crimson/osd: fix formatting around ClientRequest::do_process().

Signed-off-by: Radoslaw Zarzynski <rzarzyns@redhat.com>

crimson/osd: fix header inclusion order in ops_executer.h.

Signed-off-by: Radoslaw Zarzynski <rzarzyns@redhat.com>

crimson/osd: PG::with_locked_obc() doesn't depend on Op anymore.

Signed-off-by: Radoslaw Zarzynski <rzarzyns@redhat.com>

crimson/osd: drop namespace for arg in PG::with_locked_obc().

It's unnecessary.

Signed-off-by: Radoslaw Zarzynski <rzarzyns@redhat.com>

crimson/osd: avoid unneeded obc catching in PG::do_osd_ops_execute().

Actually, `OpsExecuter` already holds `ObjectContextRef` and even
has a (private till now) getter for `hobject_t` extraction.

Signed-off-by: Radoslaw Zarzynski <rzarzyns@redhat.com>

Merge pull request #40872 from wjwithagen/wjw-fix-pidfile-order

core: fix compiler warning due to difference in order of struct members

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

Merge pull request #40749 from tchaikov/wip-python3-spec

ceph.spec: consolidate python3-{setuptools,Cython} builds deps

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

Merge pull request #40869 from tchaikov/wip-run-make-check.sh-py3

run-make-check.sh: do not specify -DWITH_PYTHON=3

Reviewed-by: Josh Durgin <jdurgin@redhat.com>
Reviewed-by: Neha Ojha <nojha@redhat.com>
Reviewed-by: Patrick Donnelly <pdonnell@redhat.com>

cmake: rearrange configure_file()

AddCephTest and googletest's CMake scripts also call
find_package(Python3...), but they do not specify the required minor
version of Python3. by default, find_package(Python3...) picks the highest
available python3. so, if we have multiple python3 versions installed in the
system, and the highest python3 version is not the one specified by the
-DWITH_PYTHON3=3.x.y in the cmake command line, we might end up using a
different python3 for the ceph CLI. and even worse, the required python3
package might not available for the picked python3 interpreter found by
googletest. as, in general, only a single python3 has the full access to
prepackaged python3-* shipped by a GNU/Linux distro.

in this change, the configure_file() calls are rearranged to the top of
src/CMakeLists.txt, so they have less chance to use the "polluted" cmake
variable for their subvars.

this change address the test failure where we have, for instance, python3.8
installed on RHEL8/CentOS8, where python3.6 is the python3 which has
the access to the python3-* packages.

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

run-make-check.sh: do not specify -DWITH_PYTHON=3

should leave it to do_cmake.sh to decide which python3 version to use,
there is case that we have multiple python3 installed, but only one of them
is fully supported by the distro, in the sense that python3-* packages
are packaged for that python3.

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

Merge pull request #40901 from tchaikov/wip-mgr-rook

cmake: let WITH_MGR_ROOK_CLIENT depend on WITH_MGR

Reviewed-by: Reviewed-by: Sage Weil <sage@redhat.com>
Reviewed-by: Willem Jan Withagen <wjw@digiware.nl>
Reviewed-by: Patrick Donnelly <pdonnell@redhat.com>

cmake: let WITH_MGR_ROOK_CLIENT depend on WITH_MGR

it does not depend on WITH_MGR_DASHBOARD_FRONTEND, which is disabled by
default and is used for enable/disable the inclusion of dashboard
support. while root client is used by orchestrator. so it should depend
on WITH_MGR not WITH_MGR_DASHBOARD_FRONTEND.

this change addresses the regression introduced by
1003f1ffeee0d016e353f2c8c4ba692b5fdc2805

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

monitoring/grafana: Remove erroneous elements in hosts-overview Grafana dashboard

The hosts-overview Grafana dashboard json file contains a repeated element, making
it invalid JSON. Some JSON parsers handle this. However, this prevents Jsonnet
from parsing the dashboard, which prevents the deployment of this dashboard via
Jsonnet.

Fixes: https://tracker.ceph.com/issues/50410
Signed-off-by: Malcolm Holmes <mdh@odoko.co.uk>

crimson/mgr: don't report if there is no connection available.

During a teuthology run [1] following crash happended:

```
rzarzynski@teuthology:/home/teuthworker/archive/rzarzynski-2021-04-08_10:14:11-rados-master-distro-basic-smithi/6028696$ less remote/smithi052/log/ceph-osd.3.log.gz
...
DEBUG 2021-04-08 10:32:58,548 [shard 0] ms - [osd.3(client) v2:172.21.15.52:6813/30889@62168 >> mon.0 v2:172.21.15.52:3300/0] <== #3 === mgrmap(e 4) v1 (1796)
INFO  2021-04-08 10:32:58,549 [shard 0] ms - [osd.3(client) v2:172.21.15.52:6813/30889@62056 >> mgr.4100 v2:172.21.15.52:6800/30259] closing: reset no, replace no
DEBUG 2021-04-08 10:32:58,549 [shard 0] ms - [osd.3(client) v2:172.21.15.52:6813/30889@62056 >> mgr.4100 v2:172.21.15.52:6800/30259] TRIGGER CLOSING, was READY
INFO  2021-04-08 10:32:58,549 [shard 0] ms - [osd.3(client) v2:172.21.15.52:6813/30889@62056 >> mgr.4100 v2:172.21.15.52:6800/30259] execute_ready(): protocol aborted at CLOSING -- std::system_error (error crimson::net:4, read eof)
DEBUG 2021-04-08 10:32:58,549 [shard 0] ms - [osd.3(client) v2:172.21.15.52:6813/30889@62056 >> mgr.4100 v2:172.21.15.52:6800/30259] closed!
Segmentation fault on shard 0.
Backtrace:
  0x000000000151765c
  0x00000000014d9600
  0x00000000014d9902
  0x00000000014d9972
  /lib64/libpthread.so.0+0x0000000000012b1f
  0x0000000000e59cba
  0x00000000014dc8a6
  0x00000000014cdd1c
  0x0000000001503053
  0x000000000149fab7
  0x00000000006e0ef5
  /lib64/libc.so.6+0x00000000000237b2
  0x000000000072a23d
daemon-helper: command crashed with signal 11
```
[1]: http://pulpito.front.sepia.ceph.com/rzarzynski-2021-04-08_10:14:11-rados-master-distro-basic-smithi/6028696/

GDB testifies the `conn` during the execution of `ceph::mgr:report()` was null:

```
(gdb) frame 7
154 in /usr/src/debug/ceph-17.0.0-2935.g4153f8c2.el8.x86_64/src/crimson/mgr/client.cc
(gdb) print conn
$1 = {_b = 0x0, _p = 0x0}
```

Taken altogether with the `mgr.4100 v2:172.21.15.52:6800/30259] closed!`
debug this suggests that a call to `report()` occurred (likely from the
timer) but we were in the middle of the unatomic reconnect sequence:

```cpp
seastar::future<> Client::reconnect()
{
  if (conn) {
    conn->mark_down();
    conn = {};
  }
  // ...
  return seastar::sleep(a_while).then([this] {
    // ...
    conn = msgr.connect(peer, CEPH_ENTITY_TYPE_MGR);
  });
}
```

This commit alters the `mgr::report()` to skip reporting is the `conn`
is unavailable.

Signed-off-by: Radoslaw Zarzynski <rzarzyns@redhat.com>

Merge PR #40892 into master

* refs/pull/40892/head:
qa/crontab: reduce pacific priority

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

Merge pull request #40890 from tchaikov/wip-yamlize-options-doc

common/options,doc: extract formatted desc into .yaml.in

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

common/options/global.yaml.in: remove unused mon options

These were removed in a5f9ca1571cb2b9965d57c9102239aac28a5cc2a

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

Merge PR #40888 into master

* refs/pull/40888/head:
qa/tasks/cephadm: ignore --keep-logs failure
qa/tasks/cephadm: use yaml.dump_all()
qa/suites/rados/cephadm/smoke-*: use cephadm.wait_for_service
qa/tasks/cephadm: tear down clsuter before gathering logs
qa/suites/rados/cephadm/smoke-roleless: test rgw-ingress
mgr/cephadm: remove virtual_ip check during scheduling
mgr/orchestrator: orch ls: leave off virtual_ip prefixlen
qa/tasks/cephadm: add wait_for_service
qa/tasks/cephadm: allow skip_monitor_stack=true
qa/tasks/cephadm: do subst_vip for cephadm.shell and .apply
qa/tasks/vip: add vip task to allocate virtual IPs
qa/suites/rados/cephadm/smoke-roleless: add rgw-ingress test case
qa/tasks/cephadm: shell: take 'all-roles' or 'all-hosts'
qa/tasks/cephadm: let cephadm.shell take string or list

Reviewed-by: Sebastian Wagner <swagner@suse.com>

common/options/y2c.py: add newline after set_enum

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

common/options/immutable-object-cache.in.yaml: fix watermark after yaml conversion

This was set higher by 2093c882298e0dc363fc56727b8e1c4a14d0c2f2

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

common/options/global.yaml.in: fix bluestore cache ratios from yaml conversion

These were updated two years ago in 303d42533d3ffe6d4cf29bce3dded8f1e7ceb9a8

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

Merge pull request #40732 from neha-ojha/wip-50217

common/options/global.yaml.in: increase default value of bluestore_cache_trim_max_skip_pinned

Reviewed-by: Josh Durgin <jdurgin@redhat.com>
Reviewed-by: Mark Nelson <mnelson@redhat.com>
Reviewed-by: Adam Kupczyk <akupczyk@redhat.com>
Reviewed-by: Igor Fedotov <ifedotov@suse

mgr/MgrStandby: fix config observer

Otherwise handle_conf_change() is never called.

Signed-off-by: Sage Weil <sage@newdream.net>

mgr/MgrStandby: respawn if mgr_standby_modules changes

If we have standby modules and this goes false, or do not have modules
and this goes true, then respawn.

Signed-off-by: Sage Weil <sage@newdream.net>

qa/tasks/mgr/test_dashboard: skip test_standby if mgr_standby_modules=false

Signed-off-by: Sage Weil <sage@newdream.net>

qa/suites/rados/cephadm/smoke-*: use cephadm.wait_for_service

Signed-off-by: Sage Weil <sage@newdream.net>

qa/suites/rados/cephadm/smoke-singlehost: test --single-host-defaults

Signed-off-by: Sage Weil <sage@newdream.net>

cephadm: add --single-host-defaults option to bootstrap

[global] osd_crush_choose_leaf_type = 0
[global] osd_pool_default_size = 2
[mgr] mgr_standby_modules = false

Signed-off-by: Sage Weil <sage@newdream.net>

mgr/cephadm: allow mgr colo if mgr_standby_modules=false

If the standby mgr daemons' modules aren't listening on any ports, then we
can schedule multiple on the same host.

Note that this may make 'orch ps' output misleading, as ports will be
reported for each mgr instance, but only one of them will actually be
listening at any one time (if they are behaving, at least!).  Treat a
mgr port check error as non-fatal.

Signed-off-by: Sage Weil <sage@newdream.net>

mgr/MgrStandby: add mgr_standby_modules option

Add config option to control whether the standby modules are started.
Default to true (no change in behavior), but if set to false the standby
mgr modules don't do the redirect business.

Signed-off-by: Sage Weil <sage@newdream.net>

common/options/global.yaml.in: remove obsolete bluefs options

These were removed in d17cd6604b4031ca997deddc5440248aff451269 and
1f04b4589d4e96012313fe8a9069b958297b42f1

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

qa/tasks/cephadm: ignore --keep-logs failure

Signed-off-by: Sage Weil <sage@newdream.net>

common/options/global.yaml.in: fix some defaults that got changed by the yaml conversion

Presumably git didn't complain during a rebase because the whole file
was removed.

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

common/options/global.yaml.in: remove duplicate line

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

doc/radosgw/vault: add documentation for ssl authentication

Now RGW server can use custom ssl certs to authenticate with vault
server, adding steps for the same

Signed-off-by: Jiffin Tony Thottan <jthottan@redhat.com>

rgw: add support for client cert and key for vault

Authenticate vault with help of user supplied client cert and keys.

Signed-off-by: Jiffin Tony Thottan <jthottan@redhat.com>

common/options/global.yaml.in: increase default value of bluestore_cache_trim_max_skip_pinned

This option controls the rate of trimming of onodes and the earlier default of
64 has been seen to be too low for large clusters, leading to buildup of
onodes resulting in memory growth.

Increase the default value to 1000, since there are no known downsides to it.

Fixes: https://tracker.ceph.com/issues/50217
Signed-off-by: Neha Ojha <nojha@redhat.com>

rgw: update debug log for curl error

The req_data->error_buf stores error returned by the curl command, add
that information to existing debug log.

Signed-off-by: Jiffin Tony Thottan <jthottan@redhat.com>

rgw: add support to consume user given ca cert for vault

Currently RGW can authenticate with vault via SSL using system certs.
With this patch user can provide custom ca cert and location of the file
can be specified in ceph.conf like this :
rgw_crypt_require_ssl = <file path>

Fixes: https://tracker.ceph.com/issues/47776
Signed-off-by: Jiffin Tony Thottan <jthottan@redhat.com>

rgw: add seperate option for verify ssl for vault KMS engine

Signed-off-by: Jiffin Tony Thottan <jthottan@redhat.com>

Merge pull request #40189 from dang/wip-dang-zipper-rename

RGW Zipper - The Great Rename

Merge pull request #40891 from idryomov/wip-y2c-reject-duplicate-keys

common: make y2c.py choke on duplicate keys

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

mgr/cephadm: don't list non ceph daemons as needing upgrade in upgrade check

Fixes: https://tracker.ceph.com/issues/50401
Signed-off-by: Adam King <adking@redhat.com>

qa/crontab: reduce pacific priority

Now that it is released, we can reduce its priority to be in line with
other nightlies (but still higher, for now).

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

doc/rados/configuration/mon-osd-interaction: use confval directive to define options

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

common/options,doc: extract formatted desc into .yaml.in

* add a field named "fmt_desc", which is the description formatted using
  reStructuredText. it is preserved as it is if it's different from the
  desc or long_desc of an option. we can consolidate it with long_desc
  in future, and use pretty printer which has minimal support for
  reStructuredText for printing the formatted descriptions for a better
  user experience of command line. but at this moment, fmt_desc has
  only one consumer: the "ceph_confval" sphinx extension which extracts
  and translate the options yaml file to reStructuredText, which is in
  turn rendered by sphinx.
* remove unused options from the doc
  - journal_queue_max_ops
  - journal_queue_max_bytes

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

Merge pull request #40748 from tchaikov/wip-selftest-ann

pybind/mgr/selftest: add selftest to mypy and cleanups

Reviewed-by: Sebastian Wagner <swagner@suse.com>

Merge PR #40712 into master

* refs/pull/40712/head:
doc/cephadm: wrong command for single daemon events

Reviewed-by: Sebastian Wagner <swagner@suse.com>
Reviewed-by: Juan Miguel Olmo <jolmomar@redhat.com>

qa/tasks/cephadm: use yaml.dump_all()

Signed-off-by: Sage Weil <sage@newdream.net>

common: make y2c.py choke on duplicate keys

Commit 5505fc0051a3 ("common: generate legacy_config_opts.h from
.yaml.in files") inadvertently reverted a change of a default value by
adding a second "default" key with the old value.  This was corrected
in commit 75e07f8638ef ("common/options/global: correct default of
auth_mon_ticket_ttl"), but highlights that mis-merging a yaml file
is rather easy.

To prevent this happening again, fail the build if duplicate keys
exist in any of src/common/options/*.yaml.in files.

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

qa/suites/rados/cephadm/smoke-*: use cephadm.wait_for_service

Signed-off-by: Sage Weil <sage@newdream.net>

qa/tasks/cephadm: tear down clsuter before gathering logs

We dont' always stop all services, because teuthology doesn't know about
things it didn't start.  Use rm-cluster to tear things down, but do not
remove the logs themselves.  After we get logs, we'll clean up completely.

Signed-off-by: Sage Weil <sage@newdream.net>

qa/suites/rados/cephadm/smoke-roleless: test rgw-ingress

Test this properly by downing each rgw and haproxy in turn and ensuring
that things remain up.

Signed-off-by: Sage Weil <sage@newdream.net>

mgr/cephadm: remove virtual_ip check during scheduling

In 2f33c6ebbc8e2a6c3844a6921c857fb0796a1552 we made the keepalived task
set the necessary sysctls  to add a virtual_ip, so we don't need this
check anymore.

Signed-off-by: Sage Weil <sage@newdream.net>

mgr/orchestrator: orch ls: leave off virtual_ip prefixlen

Signed-off-by: Sage Weil <sage@newdream.net>

qa/tasks/cephadm: add wait_for_service

Signed-off-by: Sage Weil <sage@newdream.net>

qa/tasks/cephadm: allow skip_monitor_stack=true

(Useful for roleless when we want to go faster)

Signed-off-by: Sage Weil <sage@newdream.net>

qa/tasks/cephadm: do subst_vip for cephadm.shell and .apply

Signed-off-by: Sage Weil <sage@newdream.net>

qa/tasks/vip: add vip task to allocate virtual IPs

Signed-off-by: Sage Weil <sage@newdream.net>

Merge PR #40411 into master

* refs/pull/40411/head:
doc: add note about removal of the `cephfs` nfs cluster type
mgr/volumes/nfs: drop `type` param during cluster create

Reviewed-by: Sage Weil <sage@redhat.com>
Reviewed-by: Varsha Rao <varao@redhat.com>