mds/quiesce-db: always clear the db if a membership is lost

Fixes: https://tracker.ceph.com/issues/64912
Signed-off-by: Leonid Usov <leonid.usov@ibm.com>
(cherry picked from commit e1c8e08961e8199f1da74dc9d1eb2a940e2908d5)

Merge pull request #56601 from zdover23/wip-doc-2024-04-01-backport-56600-to-squid

squid: doc/dev: update leads list

doc/dev: update leads list

Remove Joao Luis from the list of component leads.

Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit 158b0403bd66cdf6c4b456146705f42d69f56730)

Merge pull request #56593 from zdover23/wip-doc-2024-03-31-backport-56592-to-squid

squid: doc/rados/operations: Improve crush_location docs

doc/rados/operations: Improve crush_location docs

* Fix incorrect syntax
* Use underscores for config options, like other ceph docs did
* Fix incorrect statement that crush_location_hook adds fiels; it replaces
* Explain `root=default host=HOSTNAME` is not set if `crush_location` is given
* Remove duplication across sections
* Point out that `root=default` is important

Signed-off-by: Niklas Hambüchen <mail@nh2.me>
(cherry picked from commit d91e75e1e92b73bf967740575b3c27a86ef844d5)

Merge pull request #56588 from zdover23/wip-doc-2024-03-30-backport-56581-to-squid

squid: doc/dev: update leads list

doc/dev: update leads list

Update the list of component leads in
doc/dev/developer_guide/essentials.rst.

Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit 57e9427eac08b4057999f5afc44cf0fc77a3068f)

Merge pull request #56547 from rhcs-dashboard/wip-65193-squid

squid: mgr/dashboard: fix rgw display name validation

Reviewed-by: Avan Thakkar <athakkar@redhat.com>
Reviewed-by: Ankush Behl <cloudbehl@gmail.com>

Merge pull request #56556 from afreen23/wip-65204-squid

squid: mgr/dashboard: make ceph logo redirect to dashboard

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

Merge pull request #56559 from afreen23/wip-64829-squid

squid: mgr/dashboard: Locking improvements in bucket create form

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

Merge pull request #56522 from zdover23/wip-doc-2024-03-27-backport-56173-to-squid

squid: doc/dev: osd_internals/snaps.rst: add clone_overlap doc

Reviewed-by: Anthony D'Atri <anthony.datri@gmail.com>

Merge pull request #56505 from rhcs-dashboard/wip-65169-squid

squid: mgr/dashboard: bump follow-redirects from 1.15.3 to 1.15.6 in /src/pybind/mgr/dashboard/frontend

Reviewed-by: afreen23 <NOT@FOUND>

Merge pull request #56545 from rhcs-dashboard/wip-65191-squid

squid: mgr/dashboard: s/active_mds/active_nfs in fs attach form

Reviewed-by: Ankush Behl <cloudbehl@gmail.com>

Merge pull request #56549 from rhcs-dashboard/wip-65195-squid

squid: mgr/dashboard: fix clone unique validator for name validation

Reviewed-by: Ankush Behl <cloudbehl@gmail.com>

mgr/dashboard: Locking improvements in bucket create form

Fixes https://tracker.ceph.com/issues/64658
- Addition of help texts
- Addition of info/warnings related to modes and versioning
- change of Locking section layout
- renaming locking to 'Object Locking'
- changes default retention period to 10
- edit bucket only shows lock when its enabled

Signed-off-by: Afreen <afreen23.git@gmail.com>
(cherry picked from commit 014d4468fce5aadfbc63bd9651c0c66db0bf7f49)

Conflicts:
src/pybind/mgr/dashboard/frontend/src/app/shared/components/components.module.ts

mgr/dashboard: make ceph logo redirect to dashboard

Fixes https://tracker.ceph.com/issues/64734

Clicking on ceph logo in navigation takes back to overview dashboard.

Signed-off-by: Afreen <afreen23.git@gmail.com>
(cherry picked from commit 44e7bb336c26da77f49fb027578509975b5b2633)

mgr/dashboard: fix clone unique validator for name validation

Fixes: https://tracker.ceph.com/issues/65145
Signed-off-by: Nizamudeen A <nia@redhat.com>
(cherry picked from commit 8070dfd80be9de15d2234fafa78a25569fde30a7)

mgr/dashboard: fix rgw display name validation

the dot(.) character is allowed in full name

Fixes: https://tracker.ceph.com/issues/65146
Signed-off-by: Nizamudeen A <nia@redhat.com>
(cherry picked from commit 8118a32ec578455b6c2cb4000f2bbee234ae9117)

mgr/dashboard: s/active_mds/active_nfs in fs attach form

Fixes: https://tracker.ceph.com/issues/65173
Signed-off-by: Nizamudeen A <nia@redhat.com>
(cherry picked from commit 5f97e4eaed3ff25cc01fdaed39f73542e8196677)

Merge pull request #56502 from rhcs-dashboard/wip-65166-squid

squid: mgr/dashboard: fix cephfs name validation

Reviewed-by: afreen23 <NOT@FOUND>

Merge pull request #56500 from Svelar/wip-65127-squid

squid: crimson: add crimson-osd rpm and deb runtime dependencies for protobuf inherited from seastar

Merge pull request #56314 from adk3798/wip-64621-squid

squid: cephadm: Add nvmeof to autotuner calculation

Reviewed-by: John Mulligan <jmulligan@redhat.com>

Merge pull request #56324 from adk3798/wip-64993-squid

squid: mgr/cephadm: refresh public_network for config checks before checking

Reviewed-by: John Mulligan <jmulligan@redhat.com>

Merge pull request #56321 from adk3798/wip-64687-squid

squid: mgr/cephadm: fix placement with label and host pattern

Reviewed-by: John Mulligan <jmulligan@redhat.com>

Merge pull request #56319 from adk3798/wip-64633-squid

squid: cephadm: rm podman-auth.json if removing last cluster

Reviewed-by: John Mulligan <jmulligan@redhat.com>

Merge pull request #56317 from adk3798/wip-64628-squid

squid: mgr/cephadm: catch CancelledError in asyncio timeout handler

Reviewed-by: John Mulligan <jmulligan@redhat.com>

Merge pull request #56315 from adk3798/wip-64623-squid

squid: mgr/cephadm is not defining haproxy tcp healthchecks for Ganesha

Reviewed-by: John Mulligan <jmulligan@redhat.com>

Merge pull request #56313 from adk3798/squid-cephadm-sudo-cmd-auditing

squid: cephadm: sudo cmd auditing

Reviewed-by: John Mulligan <jmulligan@redhat.com>

Merge pull request #56309 from adk3798/squid-bootstrap-no-container-engine

squid: cephadm: bootstrap fails if no container engine installed

Reviewed-by: John Mulligan <jmulligan@redhat.com>

Merge pull request #56305 from adk3798/squid-enable-black-cephadm

squid: cephadm: enforce black formatting for cephadmlib

Reviewed-by: John Mulligan <jmulligan@redhat.com>

Merge pull request #56303 from adk3798/squid-ingress-haproxy-bind-ipv4v6

squid: mgr/cephadm: make HAProxy to listen on IPv4 and IPv6 by default

Reviewed-by: John Mulligan <jmulligan@redhat.com>

Merge pull request #56289 from adk3798/squid-cephadm-pull-usage

squid: cephadm: improve cephadm pull usage message

Reviewed-by: John Mulligan <jmulligan@redhat.com>

Merge pull request #56288 from adk3798/squid-grafana-mount-dashboards

squid: mgr/cephadm: mount grafana dashboards from the path

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

Merge pull request #56285 from adk3798/squid-haproxy-healthcheck-interval

squid: cephadm: adjust the ingress ha proxy health check interval

Reviewed-by: John Mulligan <jmulligan@redhat.com>

Merge pull request #56038 from asm0deuz/wip-64697-squid

squid: mgr/cephadm: Allow idmap overrides in nfs-ganesha configuration

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

Merge pull request #55839 from adk3798/squid-cephadm-device-class

squid: cephadm: remove restriction for crush device classes

Reviewed-by: John Mulligan <jmulligan@redhat.com>

Merge pull request #55821 from adk3798/squid-no-agent-workunits

squid: qa/cephadm: don't test certain workunits with agent

Reviewed-by: John Mulligan <jmulligan@redhat.com>

Merge pull request #55818 from adk3798/squid-nvmeof-log-file-mount

squid: cephadm: Add mount for nvmeof log location

Reviewed-by: John Mulligan <jmulligan@redhat.com>

Merge pull request #55814 from adk3798/squid-debian/postint

squid: debian/*.postinst: add adduser as a dependency and specify --home when adduser

Reviewed-by: Kefu Chai <tchaikov@gmail.com>

doc/dev: osd_internals/snaps.rst: add clone_overlap doc

snapdir concept was removed in: https://github.com/ceph/ceph/pull/17579

Co-authored-by: Zac Dover <zac.dover@proton.me>
Signed-off-by: Matan Breizman <mbreizma@redhat.com>
(cherry picked from commit 6a7f2b0fda74aedb213067abe5c0a64b68fbbdc7)

Merge pull request #56494 from zdover23/wip-doc-2024-03-27-backport-56450-to-squid

squid: docs/rados: remove incorrect ceph command

Reviewed-by: Anthony D'Atri <anthony.datri@gmail.com>

Merge pull request #56442 from rhcs-dashboard/wip-65102-squid

squid: mgr/dashboard: rm warning/error threshold for cpu usage

Reviewed-by: afreen23 <NOT@FOUND>

mgr/dashboard: bump follow-redirects

Bumps [follow-redirects](https://github.com/follow-redirects/follow-redirects) from 1.15.3 to 1.15.6.
- [Release notes](https://github.com/follow-redirects/follow-redirects/releases)
- [Commits](https://github.com/follow-redirects/follow-redirects/compare/v1.15.3...v1.15.6)

---
updated-dependencies:
- dependency-name: follow-redirects
  dependency-type: indirect
...

Signed-off-by: dependabot[bot] <support@github.com>
(cherry picked from commit b0275929c07ebcfba3e7877020f182c6d8b74435)

Merge pull request #56464 from aaSharma14/wip-64608-squid

squid: mgr/dashboard: replace piechart plugin charts with native pie chart panel

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

mgr/dashboard: fix cephfs name validation

allow volume name to start with dot (.)

Fixes: https://tracker.ceph.com/issues/65143
Signed-off-by: Nizamudeen A <nia@redhat.com>
(cherry picked from commit 242318452c8111717e49b896c7f9cf0ec5f5b262)

debian/control: add runtime protobuf dependency for crimson-osd inherited from seastar

Signed-off-by: Samuel Just <sjust@redhat.com>
(cherry picked from commit ab4c5daead7f26d41028625453d50bb58d3b02be)

ceph.spec.in: add runtime protobuf dependency for crimson-osd inherited from seastar

Signed-off-by: Samuel Just <sjust@redhat.com>
(cherry picked from commit 57602aad351b21dde06db8632a9c8fdc2ab541d6)

docs/rados: remove incorrect ceph command

The removed line was (incorrectly changed) output of the previous command.

Signed-off-by: Taha Jahangir <mtjahangir@gmail.com>
(cherry picked from commit 3cd39e3582ed830f86c55d2203ece72fa0f9426b)

Merge pull request #56226 from cbodley/wip-64949-squid

squid: rgw/multisite: add x-rgw-replicated-at header to replicated objects

Reviewed-by: Shilpa Jagannath <smanjara@redhat.com>

Merge pull request #55625 from rzarzynski/wip-end2end-tracing-squid

squid: tracer/osd/librados/build/rgw: rgw and osd end2end tracing using open…

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

Merge pull request #56358 from Svelar/wip-65027-squid

squid: mgr/dashboard/frontend:Ceph dashboard supports multiple languages

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

Merge pull request #56386 from rhcs-dashboard/wip-65069-squid

squid: mgr/dashboard: fix subvolume group edit size

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

Merge pull request #56473 from rhcs-dashboard/wip-65138-squid

squid: mgr/dashboard: add system users to rgw user form

Reviewed-by: Pedro Gonzalez Gomez <pegonzal@redhat.com>
Reviewed-by: afreen23 <NOT@FOUND>

Merge pull request #56465 from zdover23/wip-doc-2024-03-26-backport-56436-to-squid

squid: doc/start: improve MDS explanation

Merge pull request #56258 from guits/wip-64945-squid

squid: ceph-volume: add missing import

Merge pull request #56035 from vshankar/wip-64779-squid

squid: mds,cephfs_mirror: add labelled per-client and replication metrics

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

mgr/dashboard: add system users to rgw user form

Fixes: https://tracker.ceph.com/issues/65074
Signed-off-by: Pedro Gonzalez Gomez <pegonzal@redhat.com>
(cherry picked from commit b658bb9c32ad31608921e77603003c004ab0f15b)

doc/start: improve MDS explanation

Improve the explanation of MDS as requested by Anthony D'Atri here:

https://github.com/ceph/ceph/pull/56367#discussion_r1536667249

Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit e7eff3cf8738ddd378c2fd70075ac61dbac3cf55)

Merge pull request #56434 from zdover23/wip-doc-2024-03-25-backport-56287-to-squid

squid: mon, doc: overriding ec profile requires --yes-i-really-mean-it

Reviewed-by: Anthony D'Atri <anthony.datri@gmail.com>

mgr/dashboard: replace piechart plugin charts with native pie chart
panel

Fixes: https://tracker.ceph.com/issues/64579
Signed-off-by: Aashish Sharma <aasharma@redhat.com>
(cherry picked from commit 6e5efb626f4b3503a12371a8f609fdb5175a3ebe)

Merge pull request #56451 from rhcs-dashboard/wip-65112-squid

squid: mgr/dashboard: sanitize dashboard user creation

Reviewed-by: Pedro Gonzalez Gomez <pegonzal@redhat.com>
Reviewed-by: afreen23 <NOT@FOUND>

Merge pull request #56123 from xhernandez/wip-64738-squid

squid: client: fix leak of file handles

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

Merge PR #56391 into squid

* refs/pull/56391/head:
qa/suites/fs/nfs: use standard health ignorelist

Reviewed-by: Leonid Usov <leonid.usov@ibm.com>

Merge PR #56395 into squid

* refs/pull/56395/head:
qa: change log-whitelist to log-ignorelist

Reviewed-by: Leonid Usov <leonid.usov@ibm.com>

Merge PR #56397 into squid

* refs/pull/56397/head:
mds: ensure snapclient is synced before corruption check

Reviewed-by: Leonid Usov <leonid.usov@ibm.com>

Merge PR #56405 into squid

* refs/pull/56405/head:
qa: test fuse/kclient for mds upgrade seq
qa: ignore OSD_DOWN during cephadm upgrades
qa: ignore warning "Replacing daemon"
qa: ignore MDS_INSUFFICIENT_STANDBY
qa: remove ignorelist error parenthesis
mds: update comment on kclient decoding of MDSMap
qa: do upgrades from quincy and older reef minor releases
mds: reverse MDSMap encoding of max_xattr_size/bal_rank_mask

Reviewed-by: Leonid Usov <leonid.usov@ibm.com>

Merge PR #56406 into squid

* refs/pull/56406/head:
doc/dev: update quiesce developer document
qa: wrap quiesce verification to dump debugging on error
qa: update quiesce tests for control via locallock
qa: set archive path in vstart_runner
qa: refactor CephFSMount.kill_background to optionally kill all background jobs
qa: use kwarg for rank parameter
qa: simplify calls to (rank|mds)_(tell|asok)
Revert "pybind/mgr/volumes: block quiesce for critical .meta file"
mds: remove is_root indication on quiesce_inode op
mds: prevent new lock cache cons when invalidating an existing one
mds: use XLOCK_WAIT For local lock xlockers
mds: prevent new wrlocks on LocalLock if there exists any xlock waiter
mds: block import discover when parent directory inode is quiesced
mds: avoid issuing exclusive caps to clients lacking w caps
mds: print lock cache during invalidation
mds: use inodeno_t to track quiesce requests
mds: dispatch quiesce_inode ops after dir traversal
mds: remove quiescelock handling for SimpleLock type
mds: quiescelock as local lock + cap masking
qa: run quiesce unit tests in fs:functional
qa: add quiesce protocol unit tests
qa: detect partial migrations during large config of dist epin
qa: use stdin-killer to timeout run_shell_payload
qa: simplify run_shell argument processing
doc: add dev docs for quiesce protocol
pybind/mgr/volumes: block quiesce for critical .meta file
mds: add vxattr to block quiesce on an inode
mds: convert encoded ephemeral dist pin to flags
mds: add counter to throttle quiesce
mds: add quiesce set feature flag
mds: skip non-head inodes for quiesce
mds: add quiesce op
mds: print all SimpleLock flags in debug output
mds: pretty print mutation when dumping lock
mds: add new inode quiescelock
mds: use 128 bits for waiters on MDSCacheObject
mds: provide mechanism to authpin while freezing
mds: add command to get specific op
mds: finish request before completing internal req
mds: complete internal op if killed
mds: avoid killing dead requests
mds: add command to kill request
mds: add path argument to `ops` and `dump tree` to stream result to local file
mds: print internal_request filepaths if present
mds: add more information to debug message
mds: remove redundant parenthesis
mds: implement Mutation::dump method
mds: make LockType fields const
mds: annotate mdr with try_rdlock_snap_layout failure
mds: refactor if into switch
mds: call Locker method using this
mds: simplify assert
mds: dump locks passed to Locker::acquire_locks
mds: add LockOp::print method for debugging
mds: use new insert template via print
mds: add request result to mutation for analysis by tests
mds: add comment on locking order rules
mds: allow specifying rdlock position
mds: remove dead method
common: provide a template for object dumps
common: support long running ops without slow warnings
common: simplify loop
common: add JSONFormatterFile class
common: use more efficient vector for stack
include: use larger int for large gathers

Reviewed-by: Leonid Usov <leonid.usov@ibm.com>

Merge PR #56407 into squid

* refs/pull/56407/head:
qa/cephfs: stop ignoring MON_DOWN globally
qa: extend mon timeout coming up after mondb creation
qa: update dashboard schema for mon_status
mon: do not log MON_DOWN if monitor uptime is less than threshold

Reviewed-by: Leonid Usov <leonid.usov@ibm.com>

Merge PR #56416 into squid

* refs/pull/56416/head:
qa/cephfs: add mgr debugging

Reviewed-by: Leonid Usov <leonid.usov@ibm.com>

mgr/dashboard: sanitize dashboard user creation

Fixes: https://tracker.ceph.com/issues/65030
Signed-off-by: Pedro Gonzalez Gomez <pegonzal@redhat.com>
(cherry picked from commit 9951a3a7f314c7cd1eb147fb1b398ac6c5212806)

Merge pull request #56440 from rhcs-dashboard/wip-65109-squid

squid: mgr/dashboard: fix empty tags

Reviewed-by: Pedro Gonzalez Gomez <pegonzal@redhat.com>
Reviewed-by: afreen23 <NOT@FOUND>

mgr/dashboard: rm warning/error threshold for cpu usage

for multi-core cpu's the value can be more than 100% so it doesn't make
sense to show warning/error when the usage is at or more than 100%.
hence removing it

Fixes: https://tracker.ceph.com/issues/58838
Signed-off-by: Nizamudeen A <nia@redhat.com>
(cherry picked from commit 131a70754a6bfbbb2d1481ba74436953c6ecc226)

mgr/dashboard: fix empty tags

Fixes: https://tracker.ceph.com/issues/64911
Signed-off-by: Pedro Gonzalez Gomez <pegonzal@redhat.com>
(cherry picked from commit a60f82566551e731987b4a02d1bb2bf4cfb9a9ce)

Merge pull request #56018 from vshankar/wip-64755-squid

squid: qa: ignore container checkpoint/restore related selinux denials for centos9

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

Merge pull request #56418 from rhcs-dashboard/wip-65077-squid

squid: Snapshot schedule show subvolume path

Reviewed-by: Ankush Behl <cloudbehl@gmail.com>

mon, doc: overriding ec profile requires --yes-i-really-mean-it

This is per https://tracker.ceph.com/issues/64333#note-17 describing
driving factors of a catastrophic cluster failure.

Signed-off-by: Radoslaw Zarzynski <rzarzyns@redhat.com>
(cherry picked from commit 629ba7bd349d48cdaa6d094751e7cfce651ba2bc)

Merge pull request #56425 from zdover23/wip-doc-2024-08-24-backport-56367-to-squid

squid: doc/start: improve MDS explanation

Merge pull request #56409 from zdover23/wip-doc-2024-03-23-backport-56377-to-squid

squid: doc/start: link to mon map command

Reviewed-by: Anthony D'Atri <anthony.datri@gmail.com>

doc/start: improve MDS explanation

Remove a prolix "on behalf of" and streamline some sentences explaining
the benefits of MDS.

Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit 2da1812a1de7e32395cb95c2e0ebeea689c1b810)

mgr/dashboard: show full subvol path snap schedule

Show full subvolume path when adding snap schedule

Fixes: https://tracker.ceph.com/issues/65000
Signed-off-by: Ivo Almeida <ialmeida@redhat.com>
(cherry picked from commit b6dbf0f24d189d5b75148f07f1dc7f332d23b64a)

Merge pull request #56388 from rhcs-dashboard/wip-64960-squid

squid: mgr/dashboard: fixed nfs attach command

Reviewed-by: Ankush Behl <cloudbehl@gmail.com>

qa/cephfs: add mgr debugging

Fixes: https://tracker.ceph.com/issues/64985
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit d2d844b60be5b0881844cecf7b0a54124061152b)

doc/start: link to mon map command

Link to the "ceph mon stat" command when "Intro to Ceph" document first
mentions Monitor Maps.

Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit 5d300db825316db25706111bbe577d41d41ab4b3)

Merge pull request #56214 from rhcs-dashboard/applitools-disable-squid

squid: mgr/dashboard: disable applitools e2e

Reviewed-by: ivoalmeida <NOT@FOUND>

qa/cephfs: stop ignoring MON_DOWN globally

This should hopefully no longer be necessary with the addition of
mon_down_uptime_grace.

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit ae96ed4afb288fbc5a7ed818ba556c21418c631c)

Conflicts:
qa/cephfs/overrides/ignorelist_wrongly_marked_down.yaml: already ignored

qa: extend mon timeout coming up after mondb creation

Fixes: https://tracker.ceph.com/issues/64968
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit 57b9e161f44d2c685b4a0ab68aff1b43bfd63e66)

qa: update dashboard schema for mon_status

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit 23de8e318f56b3052873fdd3744d313423e871bf)

mon: do not log MON_DOWN if monitor uptime is less than threshold

Fixes: https://tracker.ceph.com/issues/64968
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit 4182362c1195f3f322ae50dc0054518e62fea73a)

qa: test fuse/kclient for mds upgrade seq

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit 182f4c0f54385ab15ba13e1f06c1146ba16fd64b)

qa: ignore OSD_DOWN during cephadm upgrades

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit 4bcaaa45eb45f12df2d1835831c01fe0eb6ca595)

qa: ignore warning "Replacing daemon"

This is expected for cephadm deployments where join_fs is configured, causing
affinity replacements.

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit 75d76f97b0d7c842bd52cb609b00ffe3508c01b5)

qa: ignore MDS_INSUFFICIENT_STANDBY

This is expected when bringing a volume and its mds up initially.

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit 560300f1c5ea44a3c81d87f3bf9d10727622901c)

qa: remove ignorelist error parenthesis

Some messages are duplicated to the cluster log lookign like:

    2024-02-15T22:54:31.244 INFO:teuthology.orchestra.run.smithi033.stdout:2024-02-15T22:50:00.000263+0000 mon.smithi033 (mon.0) 558 : cluster 4 [ERR] MDS_ALL_DOWN: 1 filesystem is offline

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit 0e5e847f087cde893cc49ed3434c78006055d0b5)

mds: update comment on kclient decoding of MDSMap

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit 427ad7c0f93d374db45b86073f2c7787d2e8cb9a)

qa: do upgrades from quincy and older reef minor releases

Fixes: https://tracker.ceph.com/issues/64441
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit 52c09aa1e50e3d9f36909354c43da89dec8f9bb5)

mds: reverse MDSMap encoding of max_xattr_size/bal_rank_mask

Commit e134c890 adds the bal_rank_mask with encoded (ev) version 17.  This was
merged into main Oct 2022 and made it into the reef release normally.

Commit 7b8def5c adds the max_xattr_size also with encoded (ev) version 17 but
places it before bal_rank_mask. This is problematic as there were no plans to
backport e134c890 to quincy or pacific so piggybacking on the ev 17 bump would
not work and otherwise would require the backports to be done as a set to
ensure consistency (including with the kernel client).

However, the real issue is that 7b8def5c was not merged until after reef was
already cut. This required 7b8def5c to be backported separately in [1] which
was not merged until after v18.2.1 (current reef HEAD as of this commit).
Ultimately, this means that there are reef versions (v18.2.[01]) in the wild
which expect bal_rank_mask to be encoded at ev17 and not (max_xattr_size,
bal_rank_mask). Adding to the complications, the kernel client has already
merged code [2] expecting max_xattr_size for ev17.

It was decided in a github discussion [3] to move bal_rank_mask to ev18 to
avoid updating the kernel client which was done in the main branch via 36ee8e7e
and update the reef max_xattr_size backport with the same change (d8cebd67).

Unfortunately, this breaks upgrades v18.2.[01] to newer reef versions or to
main.  The reason is that monitors will encode v17 with bal_rank_mask
(max_xattr_size is not merged yet) and send that to upgraded mgrs (which are
upgraded first). The mgr will attempt to decode bal_rank_mask as a uint64_t
(max_xattr_size) but fail because an empty (by default) bal_rank_mask is simply
encoded as a signed 32-bit integer. Consequently, the mgr will fail decoding
with:

    failed to decode message of type 45 v1: End of buffer [buffer:2]

Of course the problem does not stop there, even if the mgr were able to handle
this, the monitors/mds/clients would fail in similar fashion.

So the only choice left is to fix max_xattr_size to be encoded at ev18.
Fortunately, v18.2.2 has not been released nor has any max_xattr_size backport
to quincy/pacific been merged. The main downside will be that kernels will
wrongly decode ev17 (which is already true for ceph clusters running
v18.2.[01]). A follow-up kernel fix will be required.

[1] https://tracker.ceph.com/issues/59405
[2] linux.git d93231a6bc8a452323d5fef16cca7107ce483a27
[3] https://github.com/ceph/ceph/pull/53340#discussion_r1399255031

Fixes: https://tracker.ceph.com/issues/64440
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit 78abfeaff27fee343fb664db633de5b221699a73)

doc/dev: update quiesce developer document

To include changes relating to it now being a local lock that prevents mutable
caps.

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit 719d30d2774ab05bd9f92b7902487aec859c5d99)

qa: wrap quiesce verification to dump debugging on error

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit 48766b336d6d21051221f3cd33a5dd9b97df7740)

qa: update quiesce tests for control via locallock

- Verify multirank quiesce incl. caps issued.
- Unset splitauth experimental

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit c0f0997e17e5639499bde09a82e8fbeef95ce006)

qa: set archive path in vstart_runner

For storing misc. test artifacts.

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit 561965131eefb092163fa000fda7da9afcdcfcd4)