Merge pull request #22193 from liewegas/wip-mon-ranks

mon: make rank ordering explicit (not tied to mon address sort order)

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

Merge pull request #22275 from tchaikov/wip-24321

qa: wait longer for osd to flush pg stats

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

mon/Monitor: use legacy ranks during bootstrap for initial seed monmap

We get various fabricated monmaps with epoch 0 to get things started.  Use
the legacy rank order so that we have a well-defined ordering and mon
ranks do not collide.

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

mon/MonMap: use legacy rank order for all bootstrap monmaps

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

Merge PR #22256 into master

* refs/pull/22256/head:
unittest_addrs: check parse() return value too
common/LogEntry: remove addrs from log entry operator<<
common/LogEntry: addr -> addrs
common/LogEntry: no need for LogEntryKey encode/decode
msg/msg_types: entity_addrvec_t ctors
msg/msg_types: add entity_addrvec_t::legacy_addr() helper
msg/msg_types: some comparators for entity_addrvec_t
msg/msg_types: entity_addrvec_t parse() and operator<<

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

Merge PR #22261 into master

* refs/pull/22261/head:
qa/suites/rados: move valgrind test to singleton-flat

Merge PR #22263 into master

* refs/pull/22263/head:
doc/releases: mimic 13.2.0

Reviewed-by: Abhishek Lekshmanan <abhishek.lekshmanan@gmail.com>

unittest_addrs: check parse() return value too

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

Merge pull request #22210 from rjfd/wip-dashboard-controller-refactor

mgr/dashboard: controller infrastructure refactor and new features

Reviewed-by: Ricardo Marques <rimarques@suse.com>
Reviewed-by: Sebastian Wagner <swagner@suse.com>
Reviewed-by: Volker Theile <vtheile@suse.com>

qa: wait longer for osd to flush pg stats

pg sends pg-stats to mgr every 5 seconds, so we cannot check for the
number of pgs right after creating the pool, at that moment, the number
of pgs could be 0, that's why manger.wait_for_clean() returns right
away, and leaves us with 0 pgs: the pgs serving the pool are still being
created. that's why `manager.get_num_active_clean()` returns `0`
sometimes. so, we should force osd to flush their stats to mgr, and wait
until the pg stats converages.

Fixes: http://tracker.ceph.com/issues/24321
Signed-off-by: Kefu Chai <kchai@redhat.com>

Merge pull request #22159 from b-ranto/wip-dashboard-shutdown

mgr/dashboard: Fix dashboard shutdown/restart

Reviewed-by: John Spray <john.spray@redhat.com>
Reviewed-by: Tatjana Dehler <tdehler@suse.com>

mgr/dashboard: hacking.rst: minor improvements and cleanup

Signed-off-by: Ricardo Dias <rdias@suse.com>

mgr/dashboard: hacking.rst: controller creation instructions update

Signed-off-by: Ricardo Dias <rdias@suse.com>

mgr/dashboard: test: controllers unit test

Signed-off-by: Ricardo Dias <rdias@suse.com>

mgr/dashboard: controllers: @Endpoint annotation implementation

With these changes we now have a single implementation for both the
BaseController and RESTController classes, with the respective
overrides.

Signed-off-by: Ricardo Dias <rdias@suse.com>

Merge pull request #21960 from socketpair/output

osd: write "debug dump_missing" output to stdout

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

Merge pull request #22267 from ZVampirEM77/wip-telegraf-typo-fix

doc: fix a typo in doc/mgr/telegraf.rst

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

doc: fix a typo in doc/mgr/telegraf.rst

Signed-off-by: Enming Zhang <enming.zhang@umcloud.com>

common/LogEntry: remove addrs from log entry operator<<

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

common/LogEntry: addr -> addrs

We want to switch to an addrvec.  This requires multiple parts:

 - switch the Key type to use just the rank
 - separate entity_name_t rank
 - compat encoding
 - graylog field naming has changed (includes name)
 - syslog output formatting has changed (includes name)
 - LogEntry operator<< modified a bit

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

common/LogEntry: no need for LogEntryKey encode/decode

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

msg/msg_types: entity_addrvec_t ctors

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

msg/msg_types: add entity_addrvec_t::legacy_addr() helper

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

msg/msg_types: some comparators for entity_addrvec_t

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

msg/msg_types: entity_addrvec_t parse() and operator<<

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

Merge pull request #22223 from wido/telemetry-bool

mgr/telemetry: Check if boolean is False or not present

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

Merge pull request #22252 from liewegas/wip-24308

mon/MDSMonitor: do not send redundant MDS health messages to cluster log

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

mon/MonMap: use c++17 structured bindings

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

mon/MonmapMonitor: add 'mon set-rank <name> <rank>' command

Add a command to manipulate the mon ranks.  This lets you control which
mon is preferred as the primary, in case you care about such things.
Mostly it doesn't matter.  (Previously we preferred the mon with the
lowest-sorting IP address, which is pretty arbitrary.)

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

mon/MonMap: separate rank ordering from entity_addr_t

We currently define the mon rank ordering based on the sort order of the
mon addresses.  Change that so that the rank order is explicitly encoded
in the ranks field of the MonMap.

If we load a legacy MonMap, calculate the legacy ordering.

If the monmap does not require nautilus features yet, force the legacy
ordering.  Once all mons are >= nautilus, we can reorder ranks.  Note that
the daemons and clients (MonClients) may see a different rank ordering.
That should be okay.

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

doc/releases: mimic 13.2.0

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

qa/suites/rados: move valgrind test to singleton-flat

No distro facet (or anything else) since we require centos for this test.

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

Merge pull request #22019 from majianpeng/cmake-fio

cmake: update fio version from 3.5 to 540e235dcd276e63c57.

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

Merge PR #22257 into master

* refs/pull/22257/head:
mon/MonCommands: allow nautilus for require-osd-release

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

mon/MonMap: simplify encode() flow

We only need legacy_mon_addr for the !MONENC case.

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

Merge PR #22243 into master

* refs/pull/22243/head:
Revert "mon: no delay for single message MSG_ALIVE and MSG_PGTEMP"

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

Merge PR #22228 into master

* refs/pull/22228/head:
mon: add send_mon_message() helper
mon: change timecheck_* key to int
mon/MonClient: make connection check more robust

Reviewed-by: João Eduardo Luís <joao@suse.de>

mon/MonCommands: allow nautilus for require-osd-release

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

Merge pull request #22255 from neha-ojha/wip-double-reservation

PG: release reservations after backfill completes

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

PG: add custom_reaction Backfilled and release reservations after backfill

After backfill completes, we directly go to the Recovered state without
releasing reservations. The outstanding reservations cause double reservation
issues.

Creating a custom_reaction Backfilled, allows us to release reservations,
before transiting to the Recovered state.

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

mon: add send_mon_message() helper

Avoid a zillion calls to monmap->get_inst()

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

mon: change timecheck_* key to int

We don't need to use entity_inst_t here since we're only paying attention
to mons in the quorum.

This involves changing the message format.  To save some time/effort and
complexity, we only support time checks on all-nautilus.  If you are
mid-upgrade, the time sync checks are skipped.

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

mon/MonClient: make connection check more robust

Do no rely on the address comparison; just check the Connection*
itself.

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

Merge PR #21934 into master

* refs/pull/21934/head:
qa/standalone/mon/misc: fix features test
qa/workunits: osd require-osd-release mimic -> nautilus
test/cli/monmaptool/feature-set-unset: fix unknown ids
test/cli/monmaptool: fix test
osd/OSDMap: make NAUTILUS feature bit significant
mon: add initial NAUTILUS feature/compatset bits

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

Merge PR #22037 into master

* refs/pull/22037/head:
filestore: Raise the priority of two counters

Reviewed-by: John Spray <john.spray@redhat.com>
Reviewed-by: Sage Weil <sage@redhat.com>

Merge PR #22133 into master

* refs/pull/22133/head:
osd/ReplicatedBackend: fix use-after-free on InProgressOp

Reviewed-by: David Zafman <dzafman@redhat.com>

Merge PR #22147 into master

* refs/pull/22147/head:
test: silence unused variable warning

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

Merge pull request #22187 from jdurgin/wip-pg-hard-limit

osd: increase default hard pg limit

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

qa/standalone/mon/misc: fix features test

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

qa/workunits: osd require-osd-release mimic -> nautilus

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

Merge pull request #22233 from neha-ojha/wip-fix-2580436

qa/workunits/suites/pjd.sh: use correct dir name

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

Merge pull request #21381 from liewegas/wip-seastore

doc/dev/seastore.rst: initial draft notes

Merge PR #22230 into master

* refs/pull/22230/head:
test/cli-integration/rbd: fixed unit formatting
qa/workunits/rbd: fix unit formatting tests
librbd: clones should not inherit implicit feature bits from parent

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

Merge pull request #22250 from falcon78921/wip-doc-falcon78921-cloudstack

doc: updated reference link for creating new disk offerings in cloudstack

mon/MDSMonitor: do not send redundant MDS health messages to cluster log

Fixes: http://tracker.ceph.com/issues/24308
Signed-off-by: Sage Weil <sage@redhat.com>

Merge pull request #22178 from aclamk/wip-bug-24225-fix

common/crc/aarch64: Added cpu feature pmull and make aarch64 specific…

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

Merge pull request #22236 from tchaikov/wip-rocksdb

rocksdb: sync with upstream

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

Merge pull request #22185 from ceph/wip-yuri-powercycle

qa/tests:  added supported distro for powercycle suite

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

doc: updated reference link for creating new disk offerings in cloudstack

Signed-off-by: James McClune <jmcclune@mcclunetechnologies.net>

Revert "mon: no delay for single message MSG_ALIVE and MSG_PGTEMP"

This change doesn't looks right and causing twice as much proposal as we targeted to (limited by paxos_propose_interval).

Imaging we have a sequence of pg_temp/up_thru during a large recovery.

now =T
The 1st up_thru/pg_temp will go through fast path and trigger propose at T + paxos_min_wait, last_attempted_minwait_time = T.

now = T+ paxos_min_wait
The [2, K] up_thru will failed by (now - last_attempted_minwait_time > g_conf->paxos_propose_interval)
and go through PaxosService::should_propose, which will schedule the propose at) T+paxos_propose_interval

now= T+ paxos_propose_interval + paxos_min_wait
The K+1 up_thru/pg_temp comes, both (now - last_attempted_minwait_time > g_conf->paxos_propose_interval
and now - paxos->get_last_commit_time() > g_conf->paxos_min_wait satisfied, so we trigger another propose
in now+ paxos_min_wait = T+ paxos_propose_interval +paxos_min_wait.

clearly we made TWO proposal in each paxos_propose_interval.

This reverts commit ca652104fe91ac41c7c1788a2907178c36fbe6ef.

Signed-off-by: Xiaoxi CHEN <xiaoxchen@ebay.com>

Merge pull request #22204 from tchaikov/wip-cmake-silence-bjam

cmake: always turn off bjam debugging output

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

Merge pull request #22126 from cbodley/wip-qa-rgw-no-ec-cache

qa/rgw: disable testing on ec-cache pools

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

Merge pull request #22124 from cbodley/wip-rgw-test-zonegroup-checkpoint

test/rgw: fix for bucket checkpoints

Reviewed-by: Yehuda Sadeh <yehuda@redhat.com>

Merge pull request #22246 from tchaikov/wip-mkfs-EINVAL

qa: use $TESTDIR for testing mkfs

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

qa: use $TESTDIR for testing mkfs

as tmpfs does not support O_DIRECT, but bluefs use this flag for
accessing the underlying file.

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

test/cli/monmaptool/feature-set-unset: fix unknown ids

Move 16 -> 64 so it is still unknown.

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

Merge PR #22163 into wip-sage-testing-20180525.131829

* refs/pull/22163/head:
CodingStyle: update my_type_t sample
CodingStyle: allow #pragma once
CodingStyle: a few updates

Reviewed-by: Kefu Chai <kchai@redhat.com>
Reviewed-by: Josh Durgin <jdurgin@redhat.com>
Reviewed-by: Jesse Williamson <jwilliamson@suse.de>

Merge pull request #22237 from a2batic/link

mgr/dashboard: Fixes documentation link- to open in new tab

Reviewed-by: Ricardo Marques <rimarques@suse.com>

Merge pull request #22182 from tspmelo/wip-unit-test-pipe

mgr/dashboard: Add unit tests for all frontend pipes

Reviewed-by: Ricardo Marques <rimarques@suse.com>
Reviewed-by: Volker Theile <vtheile@suse.com>

Merge pull request #22206 from theanalyst/wip-doc-tag-fix

doc: rgw: fix tagging support status

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

Merge pull request #22207 from tchaikov/wip-dpp-constexpr

dout: declare dpp using `decltype(auto)` instead of `auto`

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

mgr/dashboard: Fixes documentation link- to open in new tab
Adds 'target' attribute to open link in new tab.
Fixes : https://tracker.ceph.com/issues/24288

Signed-off-by: a2batic <murarkakanika@gmail.com>

rocksdb: sync with upstream

since
https://github.com/facebook/rocksdb/pull/3862 has been merged by
upstream.

Fixes: http://tracker.ceph.com/issues/23653
Signed-off-by: Kefu Chai <kchai@redhat.com>

Merge PR #22026 into master

* refs/pull/22026/head:
mds: send cap export message when exporting non-auth caps to auth mds

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

Merge pull request #22232 from batrick/cephfs-release-notes-mimic

doc/releases: update CephFS mimic notes

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

qa/workunits/suites/pjd.sh: use correct dir name

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

Merge pull request #22195 from neha-ojha/wip-minor-fixes

qa/suites/rados: minor fixes

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

doc/releases: update CephFS mimic notes

Fixes: http://tracker.ceph.com/issues/23775
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>

Merge pull request #22034 from mogeb/iostat-doc

doc: add documentation for iostat

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

test/cli-integration/rbd: fixed unit formatting

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

qa/workunits/rbd: fix unit formatting tests

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

librbd: clones should not inherit implicit feature bits from parent

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

CodingStyle: update my_type_t sample

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

osd/ReplicatedBackend: fix use-after-free on InProgressOp

- op in flight to disk...
- on_change() clears the InProgressOp
- C_OSD_OnOpCommit calls op_commit() w/ bare pointer
- crash!

Fix by refcounting InProgressOp and clearing on_commit when it is
canceled.

Fixes: http://tracker.ceph.com/issues/24219
Signed-off-by: Sage Weil <sage@redhat.com>

Merge pull request #22112 from szeestraten/fix_24199

common: fix typo in rados bench write JSON output

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

Merge pull request #22093 from liewegas/wip-24162

msg/async/Event: clear time_events on shutdown

Reviewed-by: Hoamai Wang <haomai@xsky.com>

Merge pull request #21491 from majianpeng/kerneldevice

os/bluestore/KernelDevice: misc cleanup

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

mgr/telemetry: Check if boolean is False or not present

Now we disable the module if the boolean is True

Signed-off-by: Wido den Hollander <wido@42on.com>

Merge pull request #22216 from votdev/fix_table_pagination_style

mgr/dashboard: Fix some datatable CSS issues

Approved-by: Ricardo Marques <rimarques@suse.com>

common/arch/arm: Improve CPU feature check.

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

test/cli/monmaptool: fix test

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

osd/OSDMap: make NAUTILUS feature bit significant

It's not actually significant to the encoding yet, but it will be, so put
this in place now.

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

mon: add initial NAUTILUS feature/compatset bits

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

mgr/dashboard: Fix some datatable CSS issues

Signed-off-by: Volker Theile <vtheile@suse.com>

Merge pull request #22214 from joscollin/wip-ceph-fuse-typo

doc: Fix typo in ceph-fuse(8)

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

doc: Fix typo in ceph-fuse

Fixes: https://github.com/ceph/ceph/pull/21616#pullrequestreview-122923127
Signed-off-by: Jos Collin <jcollin@redhat.com>

mgr/dashboard: Add unit tests for all frontend pipes

Signed-off-by: Tiago Melo <tmelo@suse.com>

Merge pull request #22205 from tchaikov/wip-librgw_admin_user

deb,rpm: package librgw_admin_user.{h,so.*}

mgr/dashboard: tox: ignore .vscode directory from linting

Signed-off-by: Ricardo Dias <rdias@suse.com>

mgr/dashboard: controller infrastructure refactor

Signed-off-by: Ricardo Dias <rdias@suse.com>