Merge pull request #42064 from batrick/i51231

pacific: pybind/mgr/snap_schedule: Invalid command: Unexpected argument 'fs=cephfs'

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

Merge pull request #42063 from batrick/i51230

pacific: qa: fs:bugs does not specify distro

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

Merge pull request #42045 from tchaikov/pacific-51354

pacific: osd/osd_type: use f->dump_unsigned() when appropriate

Reviewed-by: Neha Ojha <nojha@redhat.com>
Reviewed-by: Kamoltat (Junior) Sirivadhna <ksirivad@redhat.com>

Merge pull request #42019 from neha-ojha/wip-50791-pacific

pacific: osd: clear data digest when write_trunc

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

Merge pull request #41944 from ronen-fr/wip-ronenf-pac-50346

pacific: osd/scrub: replace a ceph_assert() with a test

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

Merge pull request #42062 from batrick/i51203

pacific: mds: CephFS kclient gets stuck when getattr() on a certain file

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

Merge pull request #42060 from batrick/i50875

pacific: mds: MDSLog::journaler pointer maybe crash with use-after-free

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

Merge pull request #42059 from batrick/i50848

pacific: mds: "cluster [ERR]   Error recovering journal 0x203: (2) No such file or directory" in cluster log"

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

Merge pull request #42058 from batrick/i50846

pacific: mds: journal recovery thread is possibly asserting with mds_lock not locked

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

Merge pull request #42057 from batrick/i50636

pacific: session dump includes completed_requests twice, once as an integer and once as a list

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

Merge pull request #42056 from batrick/i50630

pacific: mds: Error ENOSYS: mds.a started profiler

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

Merge pull request #41460 from idryomov/wip-qa-focal-pacific

pacific: qa/tests: replaced ubuntu_latest.yaml with ubuntu 20.04

Reviewed-by: Jeff Layton <jlayton@redhat.com>
Reviewed-by: Neha Ojha <nojha@redhat.com>
Reviewed-by: Casey Bodley <cbodley@redhat.com>
Reviewed-by: Patrick Donnelly <pdonnell@redhat.com>
Reviewed-by: Sage Weil <sage@redhat.com>
Reviewed-by: Kefu Chai <kchai@redhat.com>

Merge pull request #41052 from lxbsz/wip-50445-pacific

pacific: client: make Inode to inherit from RefCountedObject

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

Merge pull request #40856 from batrick/i50282

pacific: MDS slow request lookupino #0x100 on rank 1 block forever on dispatched

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

Merge pull request #42027 from rhcs-dashboard/wip-51148-pacific

pacific: mgr/dashboard: Include Network address and labels on Host Creation form

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

mgr/dashboard: Include Network address and labels on Host Creation form

The ability to create host by specifying network address and also create
labels.

https://tracker.ceph.com/issues/50318
Signed-off-by: Nizamudeen A <nia@redhat.com>
(cherry picked from commit 7c1df692f270ad0e7c80559e5b190c98eea44981)

Conflicts:
qa/tasks/mgr/dashboard/test_host.py
   -  Keep the {'status': ''} parameter on the request body of _post
src/pybind/mgr/dashboard/controllers/host.py
   -  Added the addr and labels as parameter in the create() method
      as well as the add_host() method.
src/pybind/mgr/dashboard/services/orchestrator.py
   -  Added  the addr and labels as parameter in the add() method and
      removed the status parameter
src/pybind/mgr/dashboard/openapi.yaml
   -  Regenerated the openapi.yaml spec

qa: update cli syntax to conventional

This was using an obscure syntax that worked at one time and wasn't
documented (AFAIK).

Fixes: https://tracker.ceph.com/issues/51182
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit a402b23c84aa399a989e1a6d9aa97046481c32c5)

qa: specify distro for fs:bugs

Fixes: https://tracker.ceph.com/issues/51184
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit 8eeb1455ee48c059a44e18b3fbae0c6a54d5d506)

mds: progress the recover queue immediately after the inode is enqueued

Fixes: https://tracker.ceph.com/issues/50840
Signed-off-by: "Yan, Zheng" <ukernel@gmail.com>
(cherry picked from commit 616ebc3f04620c22439a62f3af35845fa98e508f)

mds: switch to mdr->get_client()

The get_client() will check the mdr's type, if not a client request
it will return -1.

Signed-off-by: Xiubo Li <xiubli@redhat.com>
(cherry picked from commit 4357fa8028f320e973644b7b74aedab567afde2e)

mds: place the journaler pointer under the mds_lock

When the _recovery_thread is trying to reformat the journal, it
will delete the old journal pointer and assign with a new one,
during this the mds_lock is unlocked. That means in other threads,
such as when 'flush journal', who are using the MDSLog::journaler
pointer will potentially hit use-after-free bug.

Fixes: https://tracker.ceph.com/issues/50807
Signed-off-by: Xiubo Li <xiubli@redhat.com>
(cherry picked from commit 0c143922f8a7bad0f19a0a2c97121a470b8fcba5)

mds: save the metadata pool id MDSRank class's private member

There is one rare case that when mds daemon received a new mdsmap
and during decoding it, the metadata_pool will be reset to -1, if
other threads try to get it from old mdsmap it will get a invalid
pool id.

This can also help get rid of the mds_lock else where.

Fixes: https://tracker.ceph.com/issues/50389
Signed-off-by: Xiubo Li <xiubli@redhat.com>
Signed-off-by: Xiubo Li <xiubli@redhat.com>
(cherry picked from commit 5cc44b63f433c29c6fb933ae3072ac6270e7d333)

mds: fix possible mds_lock not locked assert

Fixes: https://tracker.ceph.com/issues/50744
Signed-off-by: Xiubo Li <xiubli@redhat.com>
(cherry picked from commit 59ab95c1a86823c372e62a147ecc6e9ed434edd2)

mds: session dumps num_completed_flushes

This can be useful when debugging clients failing to advance
oldest client/flush tid.

Signed-off-by: Dan van der Ster <daniel.vanderster@cern.ch>
(cherry picked from commit b59c7676ffb880cadd1906554c3445d77afa4389)

mds: completed_requests -> num_completed_requests

Rename this in the session dump so we don't collide with the
completed_requests dump in mdstypes session_info_t::dump.

Signed-off-by: Dan van der Ster <daniel.vanderster@cern.ch>
Fixes: https://tracker.ceph.com/issues/50559
(cherry picked from commit 52098f197e74490d5f2329674dfa93ab5a191063)

mds: reset the return value for heap command

Fixes: https://tracker.ceph.com/issues/50433
Signed-off-by: Xiubo Li <xiubli@redhat.com>
(cherry picked from commit e694f0a92367b7672fcbeab9c104fd055765a36c)

test: add test case for MDS privated inos accessing

Fixes: https://tracker.ceph.com/issues/50216
Signed-off-by: Xiubo Li <xiubli@redhat.com>
(cherry picked from commit 0ec4d781787df777bbfdeb742b394c73a0d76e08)

Conflicts:
src/test/libcephfs/test.cc: trivial conflict of mixed tests

mds: make the lost+found dir accessible from clients

Inode number 0x4 is reserved for the lost+found dir, and the apps
or recovery apps need to access it.

At the same time the 0x3 is reserved for the global snaprealm, which
will also be lookup by the clients when looking up the snaprealm inodes.

Here will make all the inode less than 100 accessible by clients.

Fixes: https://tracker.ceph.com/issues/50216
Signed-off-by: Xiubo Li <xiubli@redhat.com>
(cherry picked from commit 841071b3602aa2b8bcca57b95cdf671d48036a34)

mds: move the inos 1,2 and 3 macros to ceph_fs.h

For inos 1 and 2 there has two places are defining them, to make
sure we won't miss any one of them, let's define them in only one
header file.

Fixes: https://tracker.ceph.com/issues/50216
Signed-off-by: Xiubo Li <xiubli@redhat.com>
(cherry picked from commit 8615807237458e8cebf9497a7431097cdc56c784)

client: don't allow access to MDS-private inodes

Fixes: https://tracker.ceph.com/issues/50112
Signed-off-by: Xiubo Li <xiubli@redhat.com>
(cherry picked from commit 89c511356125f892477ef42bd14c0b447ff06106)

mds: reject lookup ino requests for mds dirs

Fixes: https://tracker.ceph.com/issues/49922
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit 682a55e8a22a137f80855f2cdd98f6f230d67fce)

test: add test for invalid lookup of mdsdir

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit 4778308eb9149594cb31519a5ff77413d6ec1fab)

Merge pull request #41753 from ifed01/wip-ifed-fix-alloc-init-add-free-0-len-pac

pacific: os/bluestore: tolerate zero length for allocators' init_[add/rm]_free()

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

Merge pull request #41752 from ifed01/wip-ifed-fix-repair-multithreading-pac

pacific: os/bluestore: introduce multithreading sync for bluestore's repairer

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

Merge pull request #42007 from rhcs-dashboard/wip-51343-pacific

pacific: mgr/dashboard: crushmap tree doesn't display crush type other than root

Reviewed-by: Alfonso Martínez <almartin@redhat.com>
Reviewed-by: Avan Thakkar <athakkar@redhat.com>
Reviewed-by: Ernesto Puerta <epuertat@redhat.com>
Reviewed-by: Kefu Chai <kchai@redhat.com>

Merge pull request #41450 from idryomov/wip-monc-rotating-timeout-pacific

pacific: mon/MonClient: tolerate a rotating key that is slightly out of date

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

Merge pull request #41249 from idryomov/wip-posix-memalign-fix-pacific

pacific: common/buffer: adjust align before calling posix_memalign()

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

Merge pull request #41995 from batrick/i51335

pacific: mds: avoid journaling overhead for setxattr("ceph.dir.subvolume") for no-op case

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

Merge pull request #41971 from trociny/wip-51313-pacific

pacific: osd: fix scrub reschedule bug

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

Merge pull request #41947 from vshankar/wip-cephfs-mirror-pacfiic-backports-21062021

pacific: cephfs-mirror: backports

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

Merge pull request #41019 from idryomov/wip-reset-authenticate-err-pacific

pacific: mon/MonClient: reset authenticate_err in _reopen_session()

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

mgr: expose CRUSHMap.find_roots()

so mgr module could use it to enumerate all nodes without parents

See-also: https://tracker.ceph.com/issues/50971
Signed-off-by: Kefu Chai <kchai@redhat.com>

mgr/dashboard: crushmap tree doesn't display crush type other than root

Fixes: https://tracker.ceph.com/issues/50971
Signed-off-by: Avan Thakkar <athakkar@redhat.com>
(cherry picked from commit de1e0ccd7a6c065ab2df7d3a108f4132f3942c41)

osd/osd_type: use f->dump_unsigned() when appropriate

it is more explicit than `dump_stream()`, as we can not tell if dumped
variable is an integer or not by reading the code. it helps us to figure
out the scheme of the dumped object.

Signed-off-by: Kefu Chai <kchai@redhat.com>
(cherry picked from commit 22128e3de697f3fdf66faf3fe3b701a3a599968f)

osd/scrub: replace a ceph_assert() with a test

We are using two distinct conditions to decide whether a candidate PG is already being scrubbed. The OSD checks pgs_scrub_active(), while the PG asserts on the value of PG_STATE_FLAG.
There is a time window when PG_STATE_FLAG is set but is_scrub_active() wasn't yet set. is_reserving() covers most of that period, but the ceph_assert is just before the is_reserving check.

fixes: https://tracker.ceph.com/issues/50346
Signed-off-by: Ronen Friedman <rfriedma@redhat.com>
(cherry picked from commit 50fbfeffbe76ad683df2dd97e4f65ca6d980e91d)

Merge pull request #41983 from neha-ojha/wip-50986-pacific

pacific: crush/crush: ensure alignof(crush_work_bucket) is 1

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

Merge pull request #41982 from neha-ojha/wip-50989-pacific

pacific: mon/OSDMonitor: drop stale failure_info even if can_mark_down()

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

Merge pull request #42016 from rhcs-dashboard/wip-50773-pacific

pacific: mgr/dashboard: fix typo: Filesystems to File Systems

Reviewed-by: Waad Alkhoury <walkhour@redhat.com>
Reviewed-by: Avan Thakkar <athakkar@redhat.com>
Reviewed-by: navinbarnwal <NOT@FOUND>
Reviewed-by: Pere Diaz Bou <pdiazbou@redhat.com>

Merge pull request #42008 from rhcs-dashboard/wip-51180-pacific

pacific: mgr/dashboard: fix for right sidebar nav icon not clickable

Reviewed-by: Waad Alkhoury <walkhour@redhat.com>
Reviewed-by: aaryanporwal <NOT@FOUND>
Reviewed-by: Avan Thakkar <athakkar@redhat.com>
Reviewed-by: Pere Diaz Bou <pdiazbou@redhat.com>

Merge pull request #41948 from rhcs-dashboard/wip-51273-pacific

pacific: mgr/dashboard: bucket details: show lock retention period only in days

Reviewed-by: Waad Alkhoury <walkhour@redhat.com>
Reviewed-by: Alfonso Martínez <almartin@redhat.com>
Reviewed-by: Avan Thakkar <athakkar@redhat.com>
Reviewed-by: Pere Diaz Bou <pdiazbou@redhat.com>

Merge pull request #41913 from sebastian-philipp/pacific-backport-41265-41746-41556-41632-41805-41799-41846-41820-41813-41346-

pacific: cephadm: june batch 3

Reviewed-by: Adam King <adking@redhat.com>
Reviewed-by: Michael Fritch <mfritch@suse.com>

osd: clear data digest when write_trunc

like CEPH_OSD_OP_TRUNCATE, the truncate within CEPH_OSD_OP_WRITE should
clear oi.data_digest too.

consider such circumstance:
[write 0~10]
[write 5~10 [1@5]]

the first write will generate data_digest, if the second write don't
clear digest, it will continue calc digest and make crc mismatch.

Fixes: https://tracker.ceph.com/issues/50763
Signed-off-by: Zengran Zhang <zhangzengran@sangfor.com.cn>
(cherry picked from commit 0e4ffe5e51291d1776aacfc0d88aa3b5f4c3ecca)

mgr/dashboard: fix typo: Filesystems to File Systems

Fixes: https://tracker.ceph.com/issues/50341
Signed-off-by: Navin Barnwal <knbarnwal@gmail.com>
(cherry picked from commit 9879b442acb122b8a5842a3ea9d979272e738572)

mgr/dashboard: fix typo: Filesystems to File Systems

Fixes: https://tracker.ceph.com/issues/50341
Signed-off-by: Navin Barnwal <knbarnwal@gmail.com>
(cherry picked from commit 537b5e70d97b8097b14d160355c6e4e23a18732e)

mgr/dashboard: fix typo: Filesystems to File Systems

Fixes: https://tracker.ceph.com/issues/50341
Signed-off-by: Navin Barnwal <knbarnwal@gmail.com>
(cherry picked from commit e55ff44dfbbfb4687a496e284ca42848d6b17761)

mgr/dashboard: fix typo: Filesystems to File Systems
Added breadcrumbs and edited Filesystem -> File Systems in dashboard
Navigation.

Fixes: https://tracker.ceph.com/issues/50341
Signed-off-by: Navin Barnwal <knbarnwal@gmail.com>
(cherry picked from commit 432e3279366e61e85f337c81609e4ec9a5592765)

Merge pull request #41503 from idryomov/wip-50931-pacific

pacific: rbd-mirror: fix segfault in snapshot replayer shutdown

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

Merge pull request #41304 from MrFreezeex/wip-50713-pacific

pacific: librbd/mirror/snapshot: avoid UnlinkPeerRequest with a unlinked peer

Reviewed-by: Ilya Dryomov <idryomov@redhat.com>
Reviewed-by: Mykola Golub <mgolub@mirantis.com>

Merge pull request #42010 from trociny/wip-51345-pacific

pacific: test/librbd: use really invalid domain

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

test/librbd: use really invalid domain

in TestMockMigrationHttpClient.OpenResolveFail

Fixes: https://tracker.ceph.com/issues/51342
Signed-off-by: Mykola Golub <mgolub@suse.com>
(cherry picked from commit 7a2a3fed4cd419ac14ab573e75c4d5b7cbc867aa)

Merge pull request #41987 from rhcs-dashboard/wip-51326-pacific

pacific: mgr/dashboard: Alertmanager fails to POST alerts

Reviewed-by: Aashish Sharma <aasharma@redhat.com>
Reviewed-by: Alfonso Martínez <almartin@redhat.com>
Reviewed-by: Avan Thakkar <athakkar@redhat.com>
Reviewed-by: Ernesto Puerta <epuertat@redhat.com>
Reviewed-by: Pere Diaz Bou <pdiazbou@redhat.com>

mgr/dashboard: fix for right sidebar nav icon not clickable

fixed the responsive sidebar not opening on click event, and close sidebar on clicking tasks and notification list item because it'll be over shadowed by the sidebar
Signed-off-by: Aaryan Porwal <aaryanporwal2233@gmail.com>
(cherry picked from commit 4e53a139d96215477d00eb709c1662d8277cba1d)

mds: avoid journaling overhead for ceph.dir.subvolume for no-op case

In preparation for acquiring the xlock on the directory inode, the MDS
must journal a few events before continuing on with the setvxattr. This
can cause significant delays in the volumes ceph-mgr module which needs
to regularly enable this vxattr from multiple code paths. We could cache
in that module whether the vxattr is set but it's also pretty easy to
adjust the MDS to acquire a rdlock on the directory to check if the
subvolume flag is already set. That is much lighter weight and the lock
is generally readily available.

Fixes: https://tracker.ceph.com/issues/51276
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit b5f736eee408c220ffdfb67b10667a7b553dac25)

Merge pull request #41768 from neha-ojha/wip-50797-pacific

pacific: mon: load stashed map before mkfs monmap

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

Merge pull request #41872 from kamoltat/wip-ksirivad-backport-pacific-fix-49988

pacific: pybind/mgr/progress: Disregard unreported pgs

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

Merge pull request #41893 from cfsnyder/wip-51001-pacific

pacific: rgw: crash on multipart upload to bucket with policy

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

Merge pull request #41877 from ideepika/wip-51241-pacific

pacific: debian/control: ceph-mgr-modules-core does not Recommend ceph-mgr-rook

Reviewed-by: Deepika Upadhyay <dupadhya@redhat.com>

Merge pull request #41863 from cfsnyder/wip-50710-pacific

pacific: rgw: radosgw_admin remove bucket not purging past 1,000 objects

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

Merge pull request #41862 from cfsnyder/wip-51094-pacific

pacific: mgr/DaemonServer.cc: prevent mgr crashes caused by integer underflow that is triggered by large increases to pg_num/pgp_num

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

Merge pull request #41755 from ifed01/wip-ifed-tchaikov-wip-rocksdb-pac

pacific: bluestore: use string_view and strip trailing slash for dir listing

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

Merge pull request #40819 from ivancich/wip-50301-pacific

pacific: rgw: radoslist incomplete multipart parts marker

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

Merge pull request #40804 from k0ste/wip-50294-pacific

pacific: mgr/prometheus:Improve the pool metadata

Reviewed-by: Paul Cuzner <pcuzner@redhat.com>
Reviewed-by: Laura Paduano <lpaduano@suse.com>

mgr/dashboard: Alertmanager fails to POST alerts

Fixes: https://tracker.ceph.com/issues/51312
Signed-off-by: Avan Thakkar <athakkar@redhat.com>
(cherry picked from commit bbff025b00a62d9292c939ecaf98d9d7464ffecf)

crush/crush: ensure alignof(crush_work_bucket) is 1

in do_rule(), we allocate the space for crush_work_bucket using

char work[crush_work_size(crush, maxout)];

where crush_work_size() calculate the size like:

map->working_size + result_max * 3 * sizeof(__u32);

so work is allocated on stack, but the alignment of the
crush_work_bucket struct is not taken into consideration, so in
crush_init_workspace(), point could point to an address which is not
aligned to 8 bytes, which is the alignment of crush_work_bucket by
default. so is its member variables, all of them are uint32_t, and hence
are also 8-bytes aligned.

to ensure the compiler generate the correct assembly for accessing
the member variables without assuming that the struct is 8-byte
aligned, we should specify the alignment explicitly.

in this change, `__attribute__ ((packed))` is specified for
crush_work_bucket, so that its alignment is 1.

this issue is spotted by ASan, it complains like:

../src/crush/mapper.c:881:22: runtime error: member access within misaligned address 0x7ffe051f90dc for type 'struct crush_work_bucket', which requires 8 byte alignment
0x7ffe051f90dc: note: pointer points here
  1d e5 77 3d 68 55 00 00  00 00 00 00 00 00 00 00  20 93 1f 05 fe 7f 00 00  10 91 1f 05 fe 7f 00 00
              ^
../src/crush/mapper.c:882:22: runtime error: member access within misaligned address 0x7ffe051f90dc for type 'struct crush_work_bucket', which requires 8 byte alignment
0x7ffe051f90dc: note: pointer points here
  1d e5 77 3d 00 00 00 00  00 00 00 00 00 00 00 00  20 93 1f 05 fe 7f 00 00  10 91 1f 05 fe 7f 00 00
              ^
../src/crush/mapper.c:883:20: runtime error: member access within misaligned address 0x7ffe051f90dc for type 'struct crush_work_bucket', which requires 8 byte alignment
0x7ffe051f90dc: note: pointer points here
  1d e5 77 3d 00 00 00 00  00 00 00 00 00 00 00 00  20 93 1f 05 fe 7f 00 00  10 91 1f 05 fe 7f 00 00
              ^

Fixes: https://tracker.ceph.com/issues/50978
Signed-off-by: Kefu Chai <kchai@redhat.com>
(cherry picked from commit 1fa8c925baeea549a824e463c89060dd60434f3f)

mon/OSDMonitor: drop stale failure_info even if can_mark_down()

in a124ee85b03e15f4ea371358008ecac65f9f4e50, we add a check to drop
stale failure_info reports. but if osdmap does not prohibit us from
marking the osd in question down, the branch checking the stale info
is not executed. in general, it is allowed to mark an osd down, so
the fix of a124ee85b03e15f4ea371358008ecac65f9f4e50 just fails to
work.

in this change, we check for stale failure report of osd in question
as long as the osd is not marked down in the same function. this should
address the slow ops of failure report issue.

Fixes: https://tracker.ceph.com/issues/50964
Signed-off-by: Kefu Chai <kchai@redhat.com>
(cherry picked from commit df6916a56841f89d66fd211729a0a7adc13042cf)

osd: fix scrub reschedule bug

not all element can be visited during reschedule traverse

Fixes: https://tracker.ceph.com/issues/49487
Signed-off-by: wencong wan <wanwc@chinatelecom.cn>
(cherry picked from commit d7561a6e58fc8043b77648a2cdd5d12bb637f92b)

Merge pull request #41714 from k0ste/wip-51105-pacific

pacific: ceph-volume: fix batch report and respect ceph.conf config values

mgr/dashboard: bucket details: show lock retention period only in days

Fixes: https://tracker.ceph.com/issues/51164
Signed-off-by: Alfonso Martínez <almartin@redhat.com>
(cherry picked from commit b927dc5bff56be742dc4f8608994f243ab18d2d2)

cephfs-mirror: silence warnings when connecting via mon host

When connecting to remote cluster via mon address, the cluster
confuration file is not required. However, the mirror daemon
fills up the ceph context with the cluster name resulting in
the following warning in (systemctl) logs:

    auth: unable to find a keyring on ...

These warning messages need to be silence as they can mislead
the user.

Fixes: http://tracker.ceph.com/issues/51204
Signed-off-by: Venky Shankar <vshankar@redhat.com>
(cherry picked from commit 98a8980e863814fd1d303e062e863b54ca3184f9)

test: add test to verify adding an active peer back to source

Fixes: http://tracker.ceph.com/issues/50447
Signed-off-by: Venky Shankar <vshankar@redhat.com>
(cherry picked from commit 074f55006edb605b35e3c9c20cdc70c4fa8d96e9)

pybind/mirroring: disallow adding a active peer back to source

Signed-off-by: Venky Shankar <vshankar@redhat.com>
(cherry picked from commit c8a5342e2b342dbfdad3ffe4f36b49d9ed192c47)

pybind/cephfs: interface to fetch file system id

Signed-off-by: Venky Shankar <vshankar@redhat.com>
(cherry picked from commit 2e9b840198041adee50b948ade89b9e7c837b6ce)

test: update log-ignorelist for fs:mirror test

Also, have a fs:mirror specific whitelist override since
the current one is a symlink to a common YAML probably used
by other tests.

Fixes: http://tracker.ceph.com/issues/50867
Signed-off-by: Venky Shankar <vshankar@redhat.com>
(cherry picked from commit d01b8c8566b21874ec5d3db700ffef45c6232330)

Merge pull request #41891 from s0nea/wip-51053-pacific

pacific: mgr/dashboard: show partially deleted RBDs

Reviewed-by: Avan Thakkar <athakkar@redhat.com>
Reviewed-by: Nizamudeen A <nia@redhat.com>
Reviewed-by: Tatjana Dehler <tdehler@suse.com>

Merge pull request #41818 from rhcs-dashboard/wip-api_label-pacific

pacific: .github/labeler: add api-change label

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

Merge pull request #41823 from batrick/i51186

pacific: qa: add async dirops testing

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

Merge pull request #41822 from batrick/i51185

pacific: qa: update RHEL to 8.4

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

Merge pull request #41811 from batrick/i51170

pacific: qa: AttributeError: 'RemoteProcess' object has no attribute 'split'

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

Merge pull request #41691 from kotreshhr/wip-51084-pacific

pacific: mds: Add full caps to avoid osd full check

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

Merge pull request #41294 from kotreshhr/wip-50627-pacific

pacific: client: Fix executeable access check for the root user

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

Merge pull request #41915 from rhcs-dashboard/wip-51259-pacific

pacific: mgr/dashboard: Fix 500 error while exiting out of maintenance

Reviewed-by: Alfonso Martínez <almartin@redhat.com>
Reviewed-by: Ernesto Puerta <epuertat@redhat.com>
Reviewed-by: Nizamudeen A <nia@redhat.com>
Reviewed-by: Pere Diaz Bou <pdiazbou@redhat.com>

mgr/dashboard: Fix 500 error while exiting out of maintenance

When you add a host in maintenance mode and then exit the maintenance
mode, a 500 server error will popup which will interrupt the whole
exit maintenance process and leave the host in an unknown/offline state.
It happened when I was setting the status of the host through the
HostSpec(). With this change, I am using the enter_maintenance api of
the orch to enable the maintenance.

Fixes: https://tracker.ceph.com/issues/51218
Signed-off-by: Nizamudeen A <nia@redhat.com>
(cherry picked from commit ea89b60e3f5e73a653eebf5cdcc3227acb7b2e62)

Conflicts:
     src/pybind/mgr/dashboard/controllers/host.py
 - Removed the addr and label variables since its not backported yet to
   pacific
 - Removed Versioning since for pacific this endpoint is not yet
   versioned to 0.1
src/pybind/mgr/dashboard/services/orchestrator.py
 - Removed the addr and label variables since its not backported yet to
   pacific
src/pybind/mgr/dashboard/frontend/src/app/shared/api/host.service.ts
 - Didn't included this file because the changes in there are irrelevant
   in pacific
src/pybind/mgr/dashboard/openapi.yaml
 - Regenerated the openapi file

doc/cephadm: removing "Octopus" from procedure

This PR removes "Octopus" from the curl-based installation
procedure.

After we moved on to Pacific, referring to Octopus looks wrong.
It looks wrong because it now is wrong.

Signed-off-by: Zac Dover <zac.dover@gmail.com>
(cherry picked from commit 1036edd69ab91144ab5e1185f11eae18fdd00f0b)

mgr/cephadm: skip ok-to-stop for mds in upgrade if not enough mds daemons

Fixes: https://tracker.ceph.com/issues/50817
Signed-off-by: Adam King <adking@redhat.com>
(cherry picked from commit 2bfc4bdccfc25bc336dc5d6a472b6fd94337beef)

mgr/cephadm: unit test for _enough_mds_for_ok_to_stop

Signed-off-by: Adam King <adking@redhat.com>
(cherry picked from commit dde64400ed6ab428c7dc60a0d92d7321d272f208)

cephadm: throw error when trying to add some unsupported repos

we son't have any repos available for fedora os always error there.

unlike octopus, el7 yum/dnf distros are not supported in pacific
so error there as well.

Signed-off-by: Adam King <adking@redhat.com>
(cherry picked from commit 32f4d8e0cfb3b013a8bfa5d51fb8da3fe9aff3a5)

cephadm: use gpg key for add-repo on ubuntu/debian

We were using the ascii version of the gpg key which
was marked as an unsupported filetype by apt-get which
caused apt-get to not make use of the repo source we
were adding.

Additionally, added aomething to make sure we update the
package list after adding the source and key

Fixes: https://tracker.ceph.com/issues/44972
Fixes: https://tracker.ceph.com/issues/45009
Signed-off-by: Adam King <adking@redhat.com>
(cherry picked from commit 54055381fd61d290bf3b44ec01e7d4f3a0eb5f1c)

cephadm: add log warn when unable to infer mon ip addr

fixup for `--mon-ip` bootstrap test (introduced b3e2c43):
  ValueError: 'eth0' does not appear to be an IPv4 or IPv6 address

Signed-off-by: Michael Fritch <mfritch@suse.com>
(cherry picked from commit ae7271d47da01dfeb90ed643da2cba91fc8fa2ce)

cephadm: raise `Error()` when unable to bind to an ip

Signed-off-by: Michael Fritch <mfritch@suse.com>
(cherry picked from commit db8d2a25c30dd25c468de598678dd15c81f3d19e)

cephadm: remove unit test inheritance

avoid running the `TestCephAdm` tests twice

Signed-off-by: Michael Fritch <mfritch@suse.com>
(cherry picked from commit 9939573adb60aa999ea7e6233d71ed119c058d7a)