cls/rgw: when object is versioned and lc transition it, the object is becoming non-current

when object is versioned and lc transition it, the object is becoming non-current

Signed-off-by: Or Friedmann <ofriedma@redhat.com>
Fixes: https://tracker.ceph.com/issues/43444
(cherry picked from commit 5f37ca283141df76a6f698ae4d7a23772b8b9e1b)

rgw: apply_olh_log ignores RGW_ATTR_OLH_VER decode error

this resolves test failures that return 400 Bad Request after:

  apply_olh_log failed to decode olh ver ''

because olh_init_modification_impl() is writing an empty attribute:

  bufferlist verbl;
  op.setxattr(RGW_ATTR_OLH_VER, verbl);

on such decode errors, link_epoch will be set to 0, and will correctly
be overwritten by real epochs from the olh log

Fixes: https://tracker.ceph.com/issues/39142
Signed-off-by: Casey Bodley <cbodley@redhat.com>
(cherry picked from commit 36c54a601c62894d0dccec49322cf7ea6bbb7d02)

rgw: allow apply_olh_log to rewrite the same olh version

if two zones write an object version with the same olh epoch,
apply_olh_log needs to be able to write the more-current instance at the
same epoch

Fixes: https://tracker.ceph.com/issues/39142
Signed-off-by: Casey Bodley <cbodley@redhat.com>
(cherry picked from commit d78aa3f4d4e283726d9c3b35a477b238df6b5af1)

rgw: apply_olh_log filters out older instances of same epoch

when apply_olh_log() detects multiple instances with the same epoch, we
only promote the one that the bucket index lists as more current

Fixes: https://tracker.ceph.com/issues/39142
Signed-off-by: Casey Bodley <cbodley@redhat.com>
(cherry picked from commit 15aa3ad341ab0df128464d22cd97bd33f185e41a)

rgw: factor out decode_olh_info()

Signed-off-by: Casey Bodley <cbodley@redhat.com>
(cherry picked from commit ef039b27df93bf89c95724d13ba82324813c3aae)

cls/rgw: only promote instance entries if they sort after current instance

to support bidirectional replication, where each zone could create its
own instances with the same olh epoch, the sort order of those instances
should be used to determine whether an entry should be promoted over the
current instance

Fixes: https://tracker.ceph.com/issues/39142
Signed-off-by: Casey Bodley <cbodley@redhat.com>
(cherry picked from commit 6f52900c817c536431f76de8f72ea9c94d5f2a99)

test/rgw: add test_concurrent_versioned_object_incremental_sync

Signed-off-by: Casey Bodley <cbodley@redhat.com>
(cherry picked from commit a078120512260c6ccf74acefe6f9f50e78803996)

Merge pull request #32743 from leseb/fast-shutdown-bkp

nautilus: osd: add osd_fast_shutdown option (default true)

Merge pull request #31554 from jan--f/wip-42703-nautilus

nautilus: ceph-volume: reject disks smaller then 5GB in inventory

Merge pull request #31553 from jan--f/wip-42741-nautilus

nautilus: ceph-volume: use fsync for dd command

qa/standalone/ceph-helpers.sh: remove osd down check

A kill doesn't induce a mark-down of the OSD with osd_fast_shutdown=true.

Signed-off-by: Sage Weil <sage@redhat.com>
(cherry picked from commit 3a62d166a7af5a62d5210417a5c10f101ce75229)

qa/standalone/ceph-helpers.sh: destroy_osd: mark osd down

Stopping the OSD doesn't guarantee that it will be marked down.

Signed-off-by: Sage Weil <sage@redhat.com>
(cherry picked from commit ede1d3677310bd152b3493e8a84da1e648fd3ea9)

osd: add osd_fast_shutdown option (default true)

If we get a SIGINT or SIGTERM or are deleted from the OSDMap, do a fast
shutdown by exiting immediately.  This has a few important benefits:

 - We immediately stop responding (binding) to any sockets, which means
   other OSDs will immediately decide we are down (and dead!).  This
   minimizes IO interruption.
 - We avoid the complex "clean" shutdown process, which is historically a
   source of bugs.

In reality, the only purpose of the "clean" shutdown is to try to tear down
everything in memory so we can do memory leak checking with valgrind.  Set
this option to false for valgrind QA runs so we can still do that.

Not that with the new read leases in octopus, we rely on the default
behavior that a ECONNREFUSED is taken to mean that the OSD is fully dead,
so that we don't have to wait for any leases to time out.  This works in
sane environments with normal IP networks, but that behavior could
conceivably be a bad idea if there are some weird network shenanigans
going on.  If osd_fast_fail_on_connection_refused were disabled, then this
fast shutdown procedure might be *worse* than the clean shutdown because
we would have to wait for the heartbeat timeout.

Signed-off-by: Sage Weil <sage@redhat.com>
(cherry picked from commit cf352c3ac0bd87d8b7e0c52ac724f94576ae5aa7)

Merge pull request #32078 from smithfarm/wip-43141-nautilus

nautilus: tools/cephfs: make 'cephfs-data-scan scan_links' fix dentry's first

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

Merge pull request #32077 from smithfarm/wip-43138-nautilus

nautilus: mds: remove unnecessary debug warning

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

Merge pull request #32075 from smithfarm/wip-43001-nautilus

nautilus: tests/cephfs: Ignore "ceph.dir.pin: No such attribute errors" in getfattr tests for old kernel client

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

Merge pull request #32073 from smithfarm/wip-42949-nautilus

nautilus: mds: properly evaluate unstable locks when evicting client

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

Merge pull request #32072 from smithfarm/wip-43170-nautilus

nautilus: tests/multimds: ignore RECENT_CRASH for multimds snapshot testing

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

Merge pull request #32233 from s0nea/wip-43159-nautilus

nautilus: mgr/dashboard: KeyError on dashboard reload

Reviewed-by: Lenz Grimmer <lgrimmer@suse.com>
Reviewed-by: Yuri Weinstein <yweinste@redhat.com>
Reviewed-by: Patrick Seidensal <pseidensal@suse.com>
Reviewed-by: Volker Theile <vtheile@suse.com>
Reviewed-by: Laura Paduano <lpaduano@suse.com>

Merge pull request #31789 from rhcs-dashboard/wip-42808-nautilus

nautilus:  mgr/dashboard: key-value-table doesn't render booleans

Reviewed-by: Ernesto Puertat <epuertat@redhat.com>
Reviewed-by: Patrick Seidensal <pseidensal@suse.com>

Merge pull request #31741 from joscollin/wip-42886-nautilus

nautilus: mgr/volumes: allow setting uid, gid of subvolume and subvolume group during creation

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

Merge pull request #32071 from smithfarm/wip-42650-nautilus

nautilus: mds: no assert on frozen dir when scrub path

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

Merge pull request #31802 from batrick/i42943

nautilus: mds: free heap memory may grow too large for some workloads

Reviewed-by: Ramana Raja <rraja@redhat.com>
Reviewed-by: Nathan Cutler <ncutler@suse.com>
Reviewed-by: Mark Nelson <mnelson@redhat.com>

Merge pull request #32065 from smithfarm/wip-42631-nautilus

nautilus: cephfs: client: add warning when cap != in->auth_cap.

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

Merge pull request #32012 from rhcs-dashboard/wip-43123-nautilus

nautilus: mgr/dashboard: Cross sign button not working for some modals

Reviewed-by: Ernesto Puerta <epuertat@redhat.com>
Reviewed-by: Lenz Grimmer <lgrimmer@suse.com>
Reviewed-by: Volker Theile <vtheile@suse.com>
Reviewed-by: Tatjana Dehler <tdehler@suse.com>

Merge pull request #32128 from votdev/issue_43079

nautilus: mgr/dashboard: Hardening accessing the metadata

Reviewed-by: Stephan Müller <smueller@suse.com>
Reviewed-by: Ernesto Puerta <epuertat@redhat.com>

Merge pull request #31784 from rhcs-dashboard/wip-42773-nautilus

nautilus: mgr/dashboard: Remove compression mode unset in pool from

Reviewed-by: Stephan Müller <smueller@suse.com>
Reviewed-by: Ernesto Puerta <epuertat@redhat.com>
Reviewed-by: Laura Paduano <lpaduano@suse.com>

Merge pull request #31999 from rhcs-dashboard/wip-43115-nautilus

nautilus: mgr/dashboard: grafana charts match time picker selection.

Reviewed-by: Lenz Grimmer <lgrimmer@suse.com>
Reviewed-by: Laura Paduano <lpaduano@suse.com>
Reviewed-by: Ernesto Puerta <epuertat@redhat.com>

Merge pull request #32658 from tchaikov/nautilus-remove-seastar-test

nautilus: cmake: remove seastar tests from "make check"

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

Merge pull request #32063 from smithfarm/wip-42878-nautilus

nautilus: common/admin_socket: Increase socket timeouts

Reviewed-by: Brad Hubbard <bhubbard@redhat.com>
Reviewed-by: Kefu Chai <kchai@redhat.com>
Reviewed-by: Sage Weil <sage@redhat.com>

Merge pull request #31850 from smithfarm/wip-43007-nautilus

nautilus: common/options: remove unused ms_msgr2_{sign,encrypt}

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

Merge pull request #32059 from smithfarm/wip-41954-nautilus

nautilus: tests/krbd: avoid hexdump skip and length options

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

Merge pull request #31848 from smithfarm/wip-42989-nautilus

nautilus: tests: kernel.sh: update for read-only changes

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

Merge pull request #31782 from rhcs-dashboard/wip-42809-nautilus

nautilus: mgr/dashboard: Handle always-on Ceph Manager modules correctly

Reviewed-by: Volker Theile <vtheile@suse.com>
Reviewed-by: Laura Paduano <lpaduano@suse.com>
Reviewed-by: Ernesto Puerta <epuertat@redhat.com>

Merge pull request #31697 from k0ste/wip-42850-nautilus

nautilus: pybind/mgr: Cancel output color control

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

cmake: remove seastar tests from "make check"

because crimson is released yet, it does not make sense to run seaetar
based tests in LTS branches.

Signed-off-by: Kefu Chai <kchai@redhat.com>
Conflicts: this change is not cherry-picked from master. as explained
  in the commit message above. we need to use these seastar based
  tests to verify that the crimson facilities are not broken in master.
  but it does not help to run them in nautilus.

mgr/dashboard: KeyError on dashboard reload

When reloading the dashboard using `ceph mgr module disable/enable
dashbaord`, the error is thrown shortly after the dashboard module has
been reactivated, because `dashboard` is not yet available in
`services`. A second attempt works.

Fixes: https://tracker.ceph.com/issues/42684
Signed-off-by: Patrick Seidensal <pseidensal@suse.com>
(cherry picked from commit 808f0bf5e0adf4abda192758edbccdee3d4113d2)

Merge pull request #32651 from tchaikov/nautilus-dts-8-on-arm64

nautilus: build/ops: install-deps,rpm: enable devtoolset-8 on aarch64 also

Reviewed-by: David Galloway <dgallowa@redhat.com>
Reviewed-by: Nathan Cutler <ncutler@suse.com>

Merge pull request #31791 from rhcs-dashboard/wip-42935-nautilus

nautilus: mgr/dashboard: sort monitors by open sessions correctly.

Reviewed-by: Ernesto Puertat <epuertat@redhat.com>
Reviewed-by: Laura Paduano <lpaduano@suse.com>

install-deps,rpm: enable devtoolset-8 on aarch64 also

in 5ae3b06e, we left aarch64 with devtoolset-7, but aarch64 builders are
also suffering from http://tracker.ceph.com/issues/38892. so i installed
devtoolset-8 on all aarch64 builders manually from the rpm packages
downloaded from koji builder:

- https://cbs.centos.org/koji/buildinfo?buildID=24923
- https://cbs.centos.org/koji/buildinfo?buildID=24931

libasan5, libubsan1 are not installed.

Signed-off-by: Kefu Chai <kchai@redhat.com>
(cherry picked from commit 00f19923b4225c25480ddb3c09a898a2444fbeb9)

Merge pull request #31646 from batrick/i42818

nautilus: global: disable THP for Ceph daemons

Merge pull request #31082 from smithfarm/wip-42279-nautilus

nautilus: tests/ceph.py: tolerate ECONNRESET errcode during logrotate

Merge pull request #30820 from liewegas/wip-alerts-nautilus

nautilus: mgr/alerts: simple health alerts

Merge pull request #30765 from smithfarm/wip-42129-nautilus

nautilus: doc/ceph-fuse: mention -k option in ceph-fuse man page

Reviewed-by: Jan Fajerski <jfajerski@suse.com>

Merge pull request #30739 from smithfarm/wip-41849-nautilus

nautilus: tools/setup-virtualenv.sh: do not default to python2.7

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

Merge pull request #30009 from smithfarm/wip-40126-nautilus

nautilus: mgr/zabbix: Adds possibility to send data to multiple zabbix servers.

Reviewed-by: Wido den Hollander <wido@widodh.nl>

make-dist: drop Python 2/3 autoselect

virtualenv will prefer Python 3 over Python 2 if both are available.

Signed-off-by: Nathan Cutler <ncutler@suse.com>
(cherry picked from commit 249bf81a32ff6d7046e525b947718d49b2314c13)

tools/setup-virtualenv.sh: do not default to python2.7

This commit changes the script's behavior when run without a --python option.
Before, we were running virtualenv with "--python python2.7" in that case.

That is not a reasonable default in September 2019, given that Python 2 EOL is
coming on January 1, 2020. Also, it breaks setup-virtualenv.sh in py3-only
environments.

With this patch, the script will not pass any "--python" option to virtualenv
(i.e., it will let virtualenv do whatever it thinks is best) in this case.

Fixes: https://tracker.ceph.com/issues/41821
Signed-off-by: Nathan Cutler <ncutler@suse.com>
(cherry picked from commit 78468e49ab2110d69bc34333c0ce90acb832de1a)

Merge pull request #32593 from trociny/wip-43500-nautilus

nautilus: rbd-mirror: make logrotate work

Reviewed-by: Jason Dillaman <dillaman@redhat.com>

Merge pull request #32447 from trociny/wip-43428-nautilus

nautilus: rbd-mirror: fix 'rbd mirror status' asok command output

Reviewed-by: Jason Dillaman <dillaman@redhat.com>

Merge pull request #32086 from dillaman/wip-42630-nautilus

nautilus: mgr: add 'rbd' profiles to support 'rbd_support' module commands

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

Merge pull request #32167 from rzarzynski/wip-fips-zeroize-memset_bzero_nautilus

nautilus: common: FIPS: audit and switch some memset & bzero users

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

Merge pull request #31332 from joscollin/wip-41182-nautilus

nautilus: mgr/volumes: fs subvolume resize command

Merge pull request #31302 from smithfarm/wip-42142-nautilus

nautilus: mds: split the dir if the op makes it oversized, because some ops maybe in flight

Merge pull request #31084 from smithfarm/wip-42424-nautilus

nautilus: tests: whitelist "Error recovering journal" for cephfs-data-scan

Merge pull request #31083 from smithfarm/wip-42422-nautilus

nautilus: tests/cephfs.py: allow client mount to reset fully

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

Merge pull request #31081 from smithfarm/wip-42158-nautilus

nautilus: cephfs: osdc/objecter: Fix last_sent in scientific format and add age to ops

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

Merge pull request #30769 from smithfarm/wip-41888-nautilus

nautilus: cephfs: "client: fix lazyio_synchronize() to update file size" and "libcephfs: Add Tests for LazyIO"

Merge pull request #30767 from smithfarm/wip-42147-nautilus

nautilus: cephfs: mds returns -5(EIO) error when the deleted file does not exist

Merge pull request #30766 from smithfarm/wip-42145-nautilus

nautilus: cephfs: client: remove Inode.dir_contacts field and handle bad whence value to llseek gracefully

Merge pull request #30764 from smithfarm/wip-42121-nautilus

nautilus: cephfs: client: add procession of SEEK_HOLE and SEEK_DATA in lseek

Merge pull request #30763 from smithfarm/wip-42040-nautilus

nautilus: cephfs: client: _readdir_cache_cb() may use the readdir_cache already clear

Merge pull request #30762 from smithfarm/wip-42035-nautilus

nautilus: cephfs: client: EINVAL may be returned when offset is 0

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

Merge pull request #30761 from smithfarm/wip-41899-nautilus

nautilus: mds: "apply configuration changes through MDSRank" and "recall caps from quiescent sessions" and "drive cap recall while dropping cache"

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

Merge pull request #31518 from trociny/wip-42726-nautilus

nautilus: rbd-mirror: clone v2 mirroring improvements

Reviewed-by: Jason Dillaman <dillaman@redhat.com>

Merge pull request #32133 from dillaman/wip-43212-nautilus

nautilus: mgr/dashboard: block mirroring page results in internal server error

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

logrotate: also sighup rbd-mirror

Signed-off-by: Mykola Golub <mgolub@suse.com>
(cherry picked from commit 86424fc3c895995d1d45f067c7852e6dce993027)

Conflicts:
src/cephadm/cephadm (does not exist)

rbd-mirror: reopen all contexts logs on SIGHUP

Signed-off-by: Mykola Golub <mgolub@suse.com>
(cherry picked from commit 9ddf111506611a596c713ffe861a41aeda05e7a5)

Conflicts:
src/tools/rbd_mirror/Mirror.cc (std::lock_guard vs Mutex::Locker)
src/tools/rbd_mirror/PoolReplayer.cc (std::lock_guard vs Mutex::Locker)

rbd-mirror: delay local/remote rados initialization until context created

We rely on that if a rados ref is initialized it contains the
valid context.

Signed-off-by: Mykola Golub <mgolub@suse.com>
(cherry picked from commit f3b49191771d2a3a20a7f55a14e0a7482ee96172)

PendingReleaseNotes: reference new MGR profile cap support

Signed-off-by: Jason Dillaman <dillaman@redhat.com>

pybind/mgr: use custom exception to handle authorization failures

Python2 does not include the PermissionError exception so use a
custom exception to support both versions.

Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit 7683f5f259bbeb6fcfc6a18c8e1a9551e068e847)

qa/workunits/rbd: add permission tests for mgr profile

Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit 7cccdef25e3c35a743414d5a7a6e0b10b8878b9b)

mgr: added placeholder 'osd' and 'mds' profiles

CephFS documents a previously unhandled profile named 'mds' that
needs to be accepted as valid (but currently provides no caps).
Similarly, the 'osd' cap is deployed in several scenarios as well.

Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit 2dc7a959509325f8da23d0baef2509174477398c)

doc/rbd: add new 'profile rbd' mgr caps to examples

Fixes: https://tracker.ceph.com/issues/42265
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit 5137cd80db7c9b699378ce22a91fbe25aba321f1)

Conflicts:
doc/rbd/rbd-kubernetes.rst: doc does not exist

pybind/mgr: test session authorization against specific pools/namespaces

Use the optional "pool" and "namespace" constraints for the RBD cap
profiles to ensure the user has access to touch the specified
pool/namespace combo.

Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit cd71413f57fa2285e9349a07449e3fcf345eeb48)

mgr: python modules can now perform authorization tests

In cases where the python service or individual python modules are
enabled via caps, the module might want to perform finer grained
tests to ensure specific commands are allowed. An example of this is
the 'rbd_support' module limiting access by pools and namespaces.

Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit 282c31c383856b45caadcefb876a71e39fe4b219)

Conflicts:
src/mgr/ActivePyModules.cc: tweaked due to changes introduced in 4878509652
src/mgr/DaemonServer.cc: tweaked due to changes introduced in 4878509652
src/mgr/MgrSession.h: trivial resolition

mgr: validate that profile caps are actually valid

Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit 9193d8762347af65bebdaea673175113f7ccd01a)

mgr: added 'profile rbd/rbd-read-only' cap

These profiles can be further restricted via 'pool' and
'namespace' argument optionals.

Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit b0d73aed19a9f3f242d5a13757016759835c164c)

Conflicts:
doc/rados/operations/user-management.rst: trivial resolution

mgr: support optional arguments for module and profile caps

This allows an optional, arbitrary key/value constraint clauses to
be appended to "profile XYZ" and "allow module XYZ" caps. A module
can then provide additional validatation against these meta-arguments.

Example:
profile rbd pool=rbd
allow module rbd_support with pool=rbd

Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit cb534e0049afc4f9259812a6624744bbdbaabafa)

mgr: add new 'allow module' cap to MgrCap

This allows specific python add-on modules to be whitelisted instead
of manually adding each command exported by the module.

  allow module {module-name} {access-spec}

Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit 3463613bd43d5fb7f8094c3b9b8a3163325ff196)

mon: dropped daemon type argument for MonCap

This was a placeholder for handling MGR caps within the MonCap
class. Now that the MGR has its own MgrCap class, this is no
longer required.

Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit dbb1b54d492134b677b4e239415c0579e34032b4)

Conflicts:
src/mon/MonCap.h: trivial resolution
src/mon/Monitor.cc: trivial resolution

mgr: stop re-using MonCap for handling MGR caps

Use the existing MonCap as the basis for a new custom MgrCap
processor.

Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit 6350bee5c873cb1cef0ca1345b89bff66817e0c1)

rbd-mirror: fix 'rbd mirror status' asok command output

This was broken by def50d04796, and implicitly fixed during
refactoring in the master (octopus) by adf1486e46c, hence it is a
direct commit to nautilus branch.

Fixes: https://tracker.ceph.com/issues/43429
Signed-off-by: Mykola Golub <mgolub@suse.com>

Merge pull request #32520 from sseshasa/wip-43495-nautilus

nautilus: mon/OSDMonitor: Don't update mon cache settings if rocksdb is not used

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

Merge pull request #32056 from smithfarm/wip-42735-nautilus

nautilus: rgw: allow reshard log entries for non-existent buckets to be cancelled

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

Merge pull request #32055 from smithfarm/wip-42733-nautilus

nautilus: rgw: auto-clean reshard queue entries for non-existent buckets

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

Merge pull request #31683 from smithfarm/wip-42840-nautilus

nautilus: rgw: when resharding store progress json

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

Merge pull request #32058 from smithfarm/wip-43161-nautilus

nautilus: rgw: url encode common prefixes for List Objects response

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

Merge remote-tracking branch 'origin/nautilus-saved' into nautilus

14.2.6

doc: update inf/infinite option for subvolume resize

Fixes: https://tracker.ceph.com/issues/42479
Signed-off-by: Jos Collin <jcollin@redhat.com>
(cherry picked from commit e5bad7d4a6fa598042e718924fadbddf9ac64e55)

doc: fs subvolume resize command

Fixes: https://tracker.ceph.com/issues/41182
Signed-off-by: Jos Collin <jcollin@redhat.com>
(cherry picked from commit 04b6b737153940c852c4b6771fb53a74a0c73bbe)

doc: uid, gid for subvolume create and subvolumegroup create commands

Fixes: https://tracker.ceph.com/issues/40959
Signed-off-by: Jos Collin <jcollin@redhat.com>
(cherry picked from commit 596c1fef6bacfc336ce0d11b7452e4bdab9f4b51)

qa/tasks/cbt: install python3 deps

Signed-off-by: Sage Weil <sage@redhat.com>
(cherry picked from commit 299375852a4b3210cf5b95d3559d69080f14339f)

Conflicts:
qa/tasks/cbt.py: install python36-* instead of python3-*
packages. as in EL7, python packages is still versioned using minor
version, and they don't "Provide" python3-*.

mon/OSDMonitor: Don't update mon cache settings if rocksdb is not used

Currently any update to mon cache config settings like 'mon_memory_target'
or 'rocksdb_cache_size' will cause a crash in case rocksdb and priority
cache manager is not being used.

If rocksdb (and therefore priority cache manager) is not being used, then
the update to the mon caches is ignored. Also, reduce the log levels of
some important log messages to help in debugging.

Fixes: https://tracker.ceph.com/issues/43454
Signed-off-by: Sridhar Seshasayee <sseshasa@redhat.com>
(cherry picked from commit b9e7a7af98ca284f4febc37fede58345ecb8a1d1)

Merge pull request #32436 from tchaikov/nautilus-cbt-py3

nautilus: qa/tasks/cbt: install python3 deps

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

qa/tasks/cbt: install python3 deps

Signed-off-by: Sage Weil <sage@redhat.com>
(cherry picked from commit 299375852a4b3210cf5b95d3559d69080f14339f)

Conflicts:
qa/tasks/cbt.py: install python36-* instead of python3-*
packages. as in EL7, python packages is still versioned using minor
version, and they don't "Provide" python3-*.

mon/PGMap.h: disable network stats in dump_osd_stats

Signed-off-by: Neha Ojha <nojha@redhat.com>
(cherry picked from commit ffb8a2a904426c8a3d6fc2fd70266418a98623c5)

 Conflicts:
src/mon/PGMap.h - trivial resolution

osd_stat_t::dump: Add option for ceph-mgr pythonn callers to skip ping network

Due to the overhead of this on large clusters, we removed the network
information was wasn't getting used anyway.

Fixes: https://tracker.ceph.com/issues/43364
Signed-off-by: David Zafman <dzafman@redhat.com>
(cherry picked from commit 1e2fe722c41d4cc34094afb157b3eb06b4a50972)

 Conflicts:
src/mon/PGMap.cc - trivial resolution
src/mon/PGMap.h - trivial resolution
src/osd/osd_types.h  - trivial resolution

Merge pull request #32466 from neha-ojha/wip-43364-2-nautilus

nautilus: mon/PGMap.h: disable network stats in dump_osd_stats

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