16.2.0

Merge pull request #40452 from smithfarm/wip-50029-pacific

pacific: rpm: drop extraneous explicit sqlite-libs runtime dependency

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

Merge pull request #40455 from liewegas/pacific-final-status

pacific: ceph_release: rc -> stable

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

Merge pull request #40217 from liewegas/cephadm-default-image-pacific

pacific: cephadm: update default image (DNM until right before final release)

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

Merge PR #40473 into pacific

* refs/pull/40473/head:
mgr/cephadm/upgrade: ignore deployed_by until mgr is upgraded

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

mgr/cephadm/upgrade: ignore deployed_by until mgr is upgraded

Until we upgrade the mgr itself, we will never be able to make our
deployed daemons have a deployed_by == target_digests.  Ignore those
daemons until the mgr is the right version.

Signed-off-by: Sage Weil <sage@newdream.net>
(cherry picked from commit bb00de5c40bf70e2b9cd14c3890d0be05d7d984b)

ceph_release: rc -> stable

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

Merge PR #40437 into pacific

* refs/pull/40437/head:
mgr/cephadm: make upgrade progress bar mention target version, not repo digest
doc/cephadm: fix rgw realm and zone flags
mgr/volumes: do not overwrite existant mds specs
mgr/cephadm: no-overwite flag for apply command
mgr/orchestrator: remove image name field from 'orch ps' and 'orch ls'
cephadm: fix parsing of keepalived version (drop leading 'v')
cephadm: keepalived needs --cap-add=NET_RAW
cephadm: fix --cap-add=NET_ADMIN
cephadm: fix quoting for keepalived env var
mgr/cephadm: ha-rgw: use correct port
cephadm: validate fsid during cephadm shell command

Reviewed-by: Juan Miguel Olmo <jolmomar@redhat.com>

Merge PR #40436 into pacific

* refs/pull/40436/head:
auth: require CEPHX_V2 by default

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

Merge PR #40425 into pacific

* refs/pull/40425/head:
test_ipaddr: check that we correctly skip loopback
common/ipaddr: also skip just `lo`

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

mgr/cephadm: make upgrade progress bar mention target version, not repo digest

The repo digest is super long and meaningless for a human user.  Instead,
use the target version (as soon as we know what it is--until then, use
the target image name).

Signed-off-by: Sage Weil <sage@newdream.net>
(cherry picked from commit 521cb8ab225561e0d29e215ffc63f293555ca00d)

doc/cephadm: fix rgw realm and zone flags

Signed-off-by: Daniel Pivonka <dpivonka@redhat.com>
(cherry picked from commit 41a1df1b96903bf283cf4ffc40e3d18eb54c447e)

mgr/volumes: do not overwrite existant mds specs

Signed-off-by: Adam King <adking@redhat.com>
(cherry picked from commit f6f7ae4c4c88324969f554d99b42b27d50209cfb)

mgr/cephadm: no-overwite flag for apply command

no-overwrite flag makes it so cephadm will only apply the spec
if the spec does not already exist.

Signed-off-by: Adam King <adking@redhat.com>
(cherry picked from commit 248173efcdebf67d4dae3bf8b4d26122d6eca911)

rpm: drop extraneous explicit sqlite-libs runtime dependency

Commit 75980798f19b8c11efd75ba4aae3e491d4c99f98 introduced a new package,
libcephsqlite, with a hard RPM dependency on a package "sqlite-libs" which
does not exist in openSUSE.

Since the runtime library dependencies of libcephsqlite are handled by RPM
transparently, this line is not needed.

Fixes: https://tracker.ceph.com/issues/50007
Signed-off-by: Nathan Cutler <ncutler@suse.com>
(cherry picked from commit 4025858570f270f8997478f95ffeec4eda8beafc)

Merge pull request #40432 from smithfarm/wip-opensuse-usrmerged-pacific

pacific: ceph.spec: prepare openSUSE usrmerge (boo#1029961)

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

Merge pull request #40372 from batrick/49935

pacific: libcephfs: test termination "what(): Too many open files"

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

Merge PR #40422 into pacific

* refs/pull/40422/head:
src/tools/cephfs-top: fix flake8 line too long error
cephfs-top: include additional metrics reported by `fs perf stats`.

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

Merge pull request #40417 from racpatel/wip-cepfstop-fixinterval

pacific: cephfs-top: allow configurable stats refresh interval

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

Merge pull request #40402 from joscollin/wip-49973-pacific

pacific: cephfs-top: be resilient to missing client metadata keys

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

Merge pull request #40358 from tchaikov/pacific-pr-39937

pacific: mgr: add mon metada using type of "mon"

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

Merge pull request #40371 from batrick/i49932

pacific: MDS should return -ENODATA when asked to remove xattr that doesn't exist

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

Merge pull request #40308 from varshar16/wip-49685-pacific

pacific: tools/cephfs-shell: fix listing of symbolic links

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

Merge pull request #40307 from varshar16/wip-49713-pacific

pacific: mgr/volumes/nfs: Add command to update cephfs exports

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

Merge pull request #40252 from singuliere/wip-49852-pacific

pacific: mds: fix race of fetching large dirfrag

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

Merge pull request #40251 from singuliere/wip-49854-pacific

pacific: client: avoid cct being released while instances are still using it

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

Merge pull request #40109 from lxbsz/wip-49520-pacific

pacific: client: wake up the front pos waiter

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

Merge pull request #40108 from lxbsz/wip-49609-pacific

pacific: qa: remove passed_validation check for test_damage

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

Merge pull request #40419 from badone/wip-pacific-revert-pr-39983

Revert "mgr/dashboard:test prometheus rules through promtool"

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

Merge pull request #40416 from neha-ojha/wip-onode-resiliant-to-split-cache-pacific

pacific: os/bluestore: Make Onode::put/get resiliant to split_cache

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

Merge pull request #40391 from neha-ojha/wip-49965-pacific

pacific: common/options: bluefs_buffered_io=true by default

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

mgr/orchestrator: remove image name field from 'orch ps' and 'orch ls'

Now that we're typically using the image digests the name isn't as helpful. We also
end up in scenarios where some images use tags for their name and others use the
digest so the image name comes out as "mix" in orch ls despite it being the same image.

Fixes: https://tracker.ceph.com/issues/47333
Signed-off-by: Adam King <adking@redhat.com>
(cherry picked from commit 30c9210d9763ff01f4780194342fcfad6c6b6344)

cephadm: fix parsing of keepalived version (drop leading 'v')

Signed-off-by: Sage Weil <sage@newdream.net>
(cherry picked from commit f33bf0647bab441993777827903bf046c1c3d80a)

cephadm: keepalived needs --cap-add=NET_RAW

This makes

Mar 24 12:00:32 dael conmon[3969650]: Wed Mar 24 16:00:32 2021: cant open raw socket. errno=1

go away and allows it to enter the MASTER state.

Signed-off-by: Sage Weil <sage@newdream.net>
(cherry picked from commit 3368844d02f52dddc25db982e586d2c2b303abf4)

cephadm: fix --cap-add=NET_ADMIN

Podman wants the = sign.  This aligns us with the other --cap-add user
(SYS_PTRACE), which uses =.

Signed-off-by: Sage Weil <sage@newdream.net>
(cherry picked from commit 6a176b02b13c7551705ecffcff4285d52b58e526)

cephadm: fix quoting for keepalived env var

This was broken by 3ea514c5525cd21722ef3fe9b90363c21e483596

Signed-off-by: Sage Weil <sage@newdream.net>
(cherry picked from commit 40e29b97863950a474adc20fccadf145e75dd8f6)

mgr/cephadm: ha-rgw: use correct port

The DaemonDescription includes the port that RGW is bound to; use that
in the haproxy configuration.

Signed-off-by: Sage Weil <sage@newdream.net>
(cherry picked from commit 401e7255062ef45e2c15d7f30dcddc3822cb67ed)

cephadm: validate fsid during cephadm shell command

Signed-off-by: Daniel Pivonka <dpivonka@redhat.com>
(cherry picked from commit 9118b08365b24c3fe26f3dcdc4bf88d8ccbcbce0)

auth: require CEPHX_V2 by default

It's been almost three years and support is present in all relevant
clients.

From the security perspective, roughly the same could be achieved
with "ceph osd set-require-min-compat-client nautilus", but this is
more user friendly as the client gets ENOTSUP instead of spinning on
"feature set mismatch" faults.

Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
(cherry picked from commit e5744672dbde2a897f5f4959339472b7b10c5688)

Merge PR #40355 into pacific

* refs/pull/40355/head:
mgr/cephadm: Fix dashboard gateway configuration when using IPV6
qa/workunits/cephadm/test_cephadm: specify image separately
mgr/cephadm: retry after JSONDecodeError in wait_for_mgr_restart()
cephadm: prevent podman from breaking socket.getfqdn()
qa/tasks/cephadm: use 'orch apply mon' to deploy mons
qa/suites/rados/cephadm/upgrade: add centos upgrade on latest octopus
mgr/cephadm/upgrade: do not crash if error races with user cancellation
doc/cephfs/nfs: Add note about cephadm NFS-Ganesha daemon port
cephadm: only bootstrap using image that matches cephadm version
mgr/cephadm: redeploy daemons deployed using old image during upgrade
mgr/cephadm: add container digests of mgr that deployed daemon to unit.meta

Reviewed-by: Juan Miguel Olmo <jolmomar@redhat.com>

Merge PR #40410 into pacific

* refs/pull/40410/head:
os/bluestore: separate omap per-pool vs per-pg alerts

Reviewed-by: Igor Fedotov <ifedotov@suse.com>

src/tools/cephfs-top: fix flake8 line too long error

Fixes: https://tracker.ceph.com/issues/50005
Signed-off-by: Varsha Rao <varao@redhat.com>
(cherry picked from commit 51cf6c3d14883797faa2c2ea4a94f65f26bdf2ef)

ceph.spec: prepare openSUSE usrmerge (boo#1029961)

The compat symlink in /sbin is no longer required and actually in
the way in the usrmerge case.

Signed-off-by: Ludwig Nussel <ludwig.nussel@suse.de>
(cherry picked from commit e4c4a4ce97fff8a5b4efa747d9cffeabcceedd25)

test_ipaddr: check that we correctly skip loopback

We should skip devices named 'lo' or of the form 'lo:0' regardless
of their IP address.

Signed-off-by: Dan van der Ster <daniel.vanderster@cern.ch>
Related-to: https://tracker.ceph.com/issues/49938
(cherry picked from commit 780125d1ed93cd7b17172752b3e76186a524103b)

common/ipaddr: also skip just `lo`

Skip iface's with name like 'lo' or of the form 'lo:0', 'lo:1'. This
brings back the original behavior from b6d0fc9e0e515e50894c08217d688a8c94db7570

Signed-off-by: Dan van der Ster <daniel.vanderster@cern.ch>
Fixes: https://tracker.ceph.com/issues/49938
(cherry picked from commit 6147c0917157efd2d35610e759685656a4989abb)

cephfs-top: include additional metrics reported by `fs perf stats`.

Without this, `cephfs-top` hits an exception since the additional
metrics keys were not configured.

Also, include a validation suring selftest that checks if `ceph
fs perf stats` metrics match what cephfs-top is configured to
report.

Fixes: http://tracker.ceph.com/issues/49974
Signed-off-by: Venky Shankar <vshankar@redhat.com>
(cherry picked from commit 5a119dcc13354549071dddae96e9fb6f7df747e5)

Merge pull request #40405 from tchaikov/pacific-pr-40400

pacific: run-make-check.sh: let ctest generate XML output

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

Revert "mgr/dashboard:test prometheus rules through promtool"

Reverts: https://github.com/ceph/ceph/pull/39983

This is currently blocking testing on ubuntu on the eve of a pacific
release. The problems associated with this PR have been resolved
upstream but have not been backported yet and are non-trivial.

This reverts commit be7f9e704c8d9ab70713a78c9a83481b5e26ee79.

Signed-off-by: Brad Hubbard <bhubbard@redhat.com>

Merge pull request #40361 from sseshasa/wip-pacific-release-note

pacific: PendingReleaseNotes: Document mclock scheduler refinements and profiles

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

cephfs-top: signal main thread to exit on SIGINT/SIGTERM

Fixes: http://tracker.ceph.com/issues/49953
Signed-off-by: Rachana Patel <racpatel@redhat.com>
(cherry picked from commit 6ea5cb5ff3380ccf6c2e0ca3f838f803173d0582)

cephfs-top: allow configurable stats refresh interval

Signed-off-by: Rachana Patel <racpatel@redhat.com>
(cherry picked from commit abd4ae9f9b1bdf1f4d7ee7b10baa9c8ec03303fc)

os/bluestore: acquire proper lock in split_cache()

Fixes: https://tracker.ceph.com/issues/49900
Signed-off-by: Igor Fedotov <ifedotov@suse.com>
(cherry picked from commit 35a3f7be8f2f204ad3b5e720d0534ca3e2a8587c)

os/bluestore: Make Onode::put/get resiliant to split_cache

In
  OnodeCacheShard* ocs = c->get_onode_cache();
  std::lock_guard l(ocs->lock);
while waiting for lock, split_cache might have changed OnodeCacheShard.
This will result in adding Onode to improper OnodeCacheShard.
Such action is obviously bad, as we will operate in future (at least once) on
different OnodeCacheShard then we got lock for. Particulary sensitive to this
are _trim and split_cache functions, as they iterate over elements.

Signed-off-by: Adam Kupczyk <akupczyk@redhat.com>
(cherry picked from commit 343b049a1328d39a69a8c4c9e9cb93ac6ac77280)

Merge pull request #40325 from batrick/49930

pacific: mon/MDSMonitor: standby-replay daemons should be removed when the flag is turned off

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

Merge pull request #40267 from kotreshhr/wip-49905-pacific

pacific: mgr/volumes: Retain suid guid bits in clone

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

Merge pull request #40091 from batrick/i49610

pacific: qa: mds removed because trimming for too long with valgrind

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

Merge pull request #40069 from petrutlucian94/wip-49634-pacific

pacific: cephfs: Add ceph-dokan, providing Windows support

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

Merge pull request #39974 from singuliere/wip-49346-pacific

pacific: vstart: fix ganesha cluster id and rados url errors

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

Merge pull request #39972 from singuliere/wip-49687-pacific

pacific: mds: add opened files/inodes and pinned i_caps metric payload support

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

Merge pull request #40206 from vshankar/wip-cephfs-mirror-pacific-backport-2

pacific: cephfs-mirror: peer bootstrap

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

Merge pull request #40349 from neha-ojha/wip-39729-pacific

pacific: os/bluestore: Add use_direct_io function

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

os/bluestore: separate omap per-pool vs per-pg alerts

Currently the health alert raised does not match the docs, and the docs
do not describe what the health alert indicates.

Octopus added per-pool omap storage.  This improves space accounting
and reporting.

Pacific added per-pg omap storage (object hash in key).  This speeds up
PG removal.

Separate everthing out into two distinct alerts raised from bluestore
and surfaced as health alerts, with corresponding config options to
disable, and update the docs accordingly.

Also update the fsck options for warn vs error, and raise separate
errors for the per-pg and per-pool cases.

Signed-off-by: Sage Weil <sage@newdream.net>
(cherry picked from commit f755e353e88b65fb923289464935ec635daf13b2)

mgr/cephadm: Fix dashboard gateway configuration when using IPV6

Fixes: https://tracker.ceph.com/issues/49957
Signed-off-by: Juan Miguel Olmo Martínez <jolmomar@redhat.com>
(cherry picked from commit 1b18f4f9cb28708b544c62b3d07f9e1b4c701e41)

qa/workunits/cephadm/test_cephadm: specify image separately

On master, this will be IMAGE_MASTER.  On pacific branch, we'll make it
IMAGE_PACIFIC so that we don't hit the release mismatch error.

Signed-off-by: Sage Weil <sage@newdream.net>
(cherry picked from commit 4d5c6044b062d7e61b1bb0c5c1efb705878517bf)

 - adjusted IMAGE_DFEAULT to be pacific

run-make-check.sh: let ctest generate XML output

to enable XUnit plugin of jenkins to consume the ctest output and
publish it in the dashboard, we need to

* let ctest generate XML output instead of plain text output
* do not fail the test if any test case fails. this allows the publisher
  to do its job by checking the XML output.
* prevent ctest from compressing the output. see
  https://issues.jenkins.io/browse/JENKINS-21737

Signed-off-by: Kefu Chai <kchai@redhat.com>
(cherry picked from commit 48ba39987d3958531589d7969750ea749e6a6d30)

tools/cephfs: fix missing keys in cephfs-top client_metadata

Fixes: https://tracker.ceph.com/issues/49736
Signed-off-by: Jos Collin <jcollin@redhat.com>
(cherry picked from commit 3fd63e97b41b9f648390e796d60451ddee2143db)

common/options: bluefs_buffered_io=true by default

Enable bluefs_buffered_io again because it makes a huge user-visible
improvement in metadata intensive scenarios, such as but not limited to
PG deletion.

In our environment, deleting PGs from 4 hybrid OSDs (sharing one SATA SSD block.db) saturates
the block.db at 350MB/s reads and causes slow reqs and flapping on the OSDs.
Those OSDs have 3GB osd_target_memory.
Enabling bluefs_buffered_io drops the SSD IO down to <1MBps and the OSDs
are performant again. (The underlying PG deletion inefficiency is being
solved separately, but the page cache is so much more effective than
the bluestore cache in this scenario).

Lastly, remove the comment about swap. We should separately advise
operators to disable swap on OSD machines, as it is much better in
our experience to OOM and restart than to chug along swapping.

Signed-off-by: Dan van der Ster <daniel.vanderster@cern.ch>
Related-to: https://tracker.ceph.com/issues/45765
Related-to: https://tracker.ceph.com/issues/47044
(cherry picked from commit 5ec8e8e63d409860c35e24a192090ac2b70af8f6)

Merge PR #40317 into pacific

* refs/pull/40317/head:
cephsqlite: add julian day offset in milliseconds
doc: add libcephsqlite
ceph.spec,debian: package libcephsqlite
test/libcephsqlite,qa: add tests for libcephsqlite
libcephsqlite: rework architecture and backend
SimpleRADOSStriper: wait for finished aios after write
SimpleRADOSStriper: add new minimal async striper
mon: define simple-rados-client-with-blocklist profile
librados: define must renew lock flag
common: add timeval conversion for durations
Revert "libradosstriper: add function to read into char*"
test_libcephsqlite: test random inserts
cephsqlite: fix compiler errors
cmake: improve build inst for cephsqlite
libcephsqlite: sqlite interface to RADOS
libradosstriper: add function to read into char*

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

mgr/cephadm: retry after JSONDecodeError in wait_for_mgr_restart()

'ceph mgr dump' does not always return valid JSON so cephadm
will throw an exception sometimes when applying a spec as per
the issue this PR closes. Add a try/except to catch a possible
JSONDecodeError and retry after sleeping.

Fixes: https://tracker.ceph.com/issues/49870
Signed-off-by: John Fulton <fulton@redhat.com>
(cherry picked from commit 0aba5704d9eb1a2df6dd437785fc1f8c558c0990)

cephadm: prevent podman from breaking socket.getfqdn()

socket.getfqdn() will return the reverse lookup for 127.0.1.1, which is
the last item listed for that IP in /etc/hosts.  Podman, by default, will
append the container name (ceph-$fsid-$name) to that line, which is not
a valid hostname, and not what we want the dashbaord to use for the URI
it advertises in the service map.

Pass --no-hosts to podman to disable this.

Docker does not appear to modify /etc/hosts by default--or, more
importantly, does not add the container name there.

Explicitly instruct podman (and docker) to add a

Fixes: https://tracker.ceph.com/issues/49890
Signed-off-by: Sage Weil <sage@newdream.net>
(cherry picked from commit cfc1f914ce74f1fd1f45e2efd3ba2ddcb2da129a)

PendingReleaseNotes: Document mclock scheduler refinements and profiles

Signed-off-by: Sridhar Seshasayee <sseshasa@redhat.com>

Merge pull request #39999 from rhcs-dashboard/wip-49708-pacific

pacific: mgr/dashboard: cluster > hosts: replace inventory->devices table title

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

test: reduce number of threads to 32 in LibCephFS.ShutdownRace

We're still occasionally hitting file descriptor limits when running
this test. Reduce the thread count to 32 for now, since it was possible
to reproduce the original problem with 10 or so threads.

Fixes: https://tracker.ceph.com/issues/49559
Signed-off-by: Jeff Layton <jlayton@redhat.com>
(cherry picked from commit 5aec283a1c33b6c21f877a27f57a1bc03b4894a0)

test: add test for removing non-existent xattr

We should get back -ENODATA.

Signed-off-by: Jeff Layton <jlayton@redhat.com>
(cherry picked from commit b94b668b7e340fa660462b75c97ab77dda9d5273)

mds: fix removexattr logic when there aren't any

The MDS currently returns success on a removexattr if the xattr map is
completely empty. Fix the subtle logic bug and have it return -ENODATA
in that case.

Fixes: https://tracker.ceph.com/issues/49833
Signed-off-by: Jeff Layton <jlayton@redhat.com>
(cherry picked from commit 85e73c7c7509cefbc50902436aca07a9a333eb23)

Merge PR #39988 into pacific

* refs/pull/39988/head:
doc: fix broken links multimds and kcephfs

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

Merge PR #40250 into pacific

* refs/pull/40250/head:
doc: mds cap acquisition readdir throttle documentation

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

Merge PR #40306 into pacific

* refs/pull/40306/head:
doc/cephfs/nfs: Add rook pod restart note, export and log block example

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

Merge pull request #40255 from singuliere/wip-49743-pacific

pacific: rgw: limit rgw_gc_max_objs to RGW_SHARDS_PRIME_1

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

Merge pull request #40261 from singuliere/wip-49310-pacific

WIP: pacific: rgw: add support for SSL encrypted AMQP connections

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

Merge pull request #40338 from cbodley/wip-rgw-lc-lock-pacific

pacific: rgw/lc: Fix use-after-free in RGWLC::process

Reviewed-by: Daniel Gryniewicz <dang@redhat.com>

Merge PR #40362 into pacific

* refs/pull/40362/head:
doc/cephfs/nfs: Add note about cephadm NFS-Ganesha daemon port

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

Merge PR #40094 into pacific

* refs/pull/40094/head:
rgw/kms/vault - PendingReleaseNotes pointer
rgw/kms/vault - s3tests for both old and new test logic.
rgw/kms/vault - rework unit test logic for new transit logic.
rgw/kms/vault - 0 terminate before rapidjson
rgw/kms/vault - document configuration for new transit logic
rgw/kms/vault - new transit logic - fix compat logic
rgw/kms/vault - define attribute for new transit logic
rgw/kms/vault - "compat" option
rgw/kms/vault - encryption context - first part
rgw/kms/vault - define attribute to store encryption context
rgw/kms/vault - share get/set attr between rgw_crypt.cc and rgw_kms.cc
rgw/kms/vault - relax configuration parsing for rgw_crypt_vault_secret_engine
rgw/kms/vault - need libicu to make canonical json for encryption contexts.
rgw/kms/kmip - document configuration for a new feature: kmip kms
rgw/kms/kmip - rgw / kmip test integration.
rgw/kms/kmip - rgw / kmip test integration.
rgw/kms/kmip - rgw / kmip test integration.
rgw/kms/kmip - correct documentation.
rgw/kms/kmip - pykmip.py needs to make keys too.
rgw/kms/kmip - pykmip.py should actually run pykmip.
rgw/kms/kmip - python3 changes for testing.
rgw/kms/kmip - string handling cleanup.
teuthology/rgw: pykmip task
kmip: first pass at implementation logic.
kmip: configuration options.
Including cmake build logic inside of libkmip.
cmake glue to build libkmip.
Added libkmip as a submodule.

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

cephsqlite: add julian day offset in milliseconds

This magic number was copied from another VFS but was not adjusted for
the xCurrentTimeInt64 interface.

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit 9107e1a34743167d8f0cf969a440f5aaed2848af)

Merge pull request #40262 from singuliere/wip-49176-pacific

pacific: rgw: Fix spurious error on empty datalog shard

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

Merge pull request #40256 from singuliere/wip-49722-pacific

pacific: qa/rgw: add librgw_file unit tests to rgw/verify task

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

Merge pull request #40257 from singuliere/wip-49601-pacific

pacific: rgw/sts: fix for encoding/decoding user namespace

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

Merge pull request #40258 from singuliere/wip-49587-pacific

pacific: rgw: added missing documentation on "pubsub" in rgw_enable_apis

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

Merge pull request #40259 from singuliere/wip-49381-pacific

pacific: rgw: multisite: fix single-part-MPU object etag misidentify problem

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

Merge pull request #40260 from singuliere/wip-49362-pacific

pacific: rgw: fix sts get_session_token duration check failed

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

Merge pull request #40339 from dillaman/wip-librbd-backports-pacific-10

pacific: librbd: miscellaneous backports

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

Merge pull request #40340 from rhcs-dashboard/wip-49941-pacific

pacific: mgr/dashboard: fix error shown when no rgw daemons are running.

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

Merge pull request #40360 from rhcs-dashboard/wip-49944-pacific

pacific: mgr/dashboard: Display users current quota usage

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

doc/cephfs/nfs: Add note about cephadm NFS-Ganesha daemon port

Fixes: https://tracker.ceph.com/issues/49921
Signed-off-by: Varsha Rao <varao@redhat.com>
(cherry picked from commit fec8bbf5cd3e3dd72d05c4eb23209c84f056edda)

mgr/dashboard: Display users current quota usage

Fixes: https://tracker.ceph.com/issues/45965
Signed-off-by: Avan Thakkar <athakkar@redhat.com>
(cherry picked from commit 9456884c2996d31ce9c39e39011d295c5e62bcf7)

mgr: add mon metada using type of "mon"

this change addresses a regression introduced by
c037f4cb5d7436879d58c34748ef516b5269781f

also remove the "P" before the json command.

see also: https://tracker.ceph.com/issues/48905

Fixes: https://tracker.ceph.com/issues/49661
Signed-off-by: Kefu Chai <kchai@redhat.com>
(cherry picked from commit 8fc290bfba4d71a60d30c2374ce4bcba37e649de)

qa/tasks/cephadm: use 'orch apply mon' to deploy mons

The 'orch daemon add ...' command is not idempotent and can cause
duplicate (and failing) attempts to add the same mon.

Maintain the ability to add mons the old way for the benefit of testing
upgrades from early octopus versions of cephadm.

Signed-off-by: Sage Weil <sage@newdream.net>
(cherry picked from commit a17593a177798e5b5371c224ede40dc411626d76)

qa/suites/rados/cephadm/upgrade: add centos upgrade on latest octopus

centos currently has the semi-broken podman 2.2.1 (no json function
for inspect), but the latest octopus (15.2.10) works around it.

Signed-off-by: Sage Weil <sage@newdream.net>
(cherry picked from commit c6a75d05565f327e7b7a3971e89f5a3e730299a6)

mgr/cephadm/upgrade: do not crash if error races with user cancellation

If the user cancels the upgrade just before the upgrade thread runs into
a problem (and these things may be correlated!), ignore the failure
instead of crashing the module.

Signed-off-by: Sage Weil <sage@newdream.net>
(cherry picked from commit 2580de1a5db476b01093a269056d2f5bd0cc3d75)

doc/cephfs/nfs: Add note about cephadm NFS-Ganesha daemon port

Fixes: https://tracker.ceph.com/issues/49921
Signed-off-by: Varsha Rao <varao@redhat.com>
(cherry picked from commit fec8bbf5cd3e3dd72d05c4eb23209c84f056edda)