rpm: bump up required GCC version to 7.3.1

so we can include the fix of
https://gcc.gnu.org/ml/libstdc++/2017-07/msg00077.html and the fix at
https://github.com/gcc-mirror/gcc/commit/7c78f2e67270187dd755632eafe3a526f4e6df30
. otherwise, if a class offers implicit conversion operator to
std::string and std::string_view, seastar will fail to compile. like

/opt/rh/devtoolset-7/root/usr/include/c++/7/bits/basic_string.h:3946:7:
note: candidate: std::basic_string<_CharT, _Traits, _Alloc>&
std::basic_string<_CharT, _Traits, _Alloc>::operator+=(const
std::basic_string<_CharT, _Traits, _Alloc>&) [with _CharT = char;
_Traits = std::char_traits<char>; _Alloc = std::allocator<char>]
       operator+=(const basic_string& __str)
       ^~~~~~~~
/opt/rh/devtoolset-7/root/usr/include/c++/7/bits/basic_string.h:3988:7:
note: candidate: std::basic_string<_CharT, _Traits, _Alloc>&
std::basic_string<_CharT, _Traits,
_Alloc>::operator+=(std::basic_string<_CharT, _Traits,
_Alloc>::__sv_type) [with _CharT = char; _Traits =
std::char_traits<char>; _Alloc = std::allocator<char>;
std::basic_string<_CharT, _Traits, _Alloc>::__sv_type =
std::basic_string_view<char>]
       operator+=(__sv_type __sv)
       ^~~~~~~~

Signed-off-by: Kefu Chai <kchai@redhat.com>
(cherry picked from commit 5bcc179bc4ac756e0a6c81cb19fffcc508ecbcd0)

Merge pull request #26585 from capri1989/doc-fix-mimic

mimic: doc: Fix default value of 'mon_osd_max_split_count'

Reviewed-by: Nathan Cutler <ncutler@suse.com>

doc: Fix default value of 'mon_osd_max_split_count'

This commit is not cherry-picked from master because the config
parameter mon_osd_max_split_count has been dropped in master

Signed-off-by: Kai Wagner <kwagner@suse.com>

Merge pull request #26442 from xiexingguo/wip-mimic

mimic: Revert "msg/async: bump global_seq when retrying connection"

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

Merge pull request #26438 from pdvian/wip-38317-mimic

mimic: os/filestore: ceph_abort() on fsync(2) or fdatasync(2) failure

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

Merge pull request #26573 from neha-ojha/wip-38083-mimic

mimic: qa/workunits/cephtool/test_kvstore_tool.sh: run test in ., not /tmp

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

qa/workunits/cephtool/test_kvstore_tool.sh: run test in ., not /tmp

Notably, we can't make a bluestore store on a tmpfs.

Signed-off-by: Sage Weil <sage@redhat.com>
(cherry picked from commit 5940487be3c848cf45d251f0a00a5d85a5b584e4)

Merge pull request #26493 from dzafman/wip-alternate-backport

mimic: core: scrub warning check incorrectly uses mon scrub interval

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

Merge pull request #26285 from smithfarm/wip-38185-mimic

mimic: tests: krbd discard no longer guarantees zeroing

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

Merge pull request #26424 from smithfarm/wip-38103-mimic

mimic: client: session flush does not cause cap release message flush

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

Merge pull request #26472 from pdvian/wip-38334-mimic

mimic: cephfs: MDCache::finish_snaprealm_reconnect() create and drop MClientSnap message

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

Merge pull request #26474 from pdvian/wip-38335-mimic

mimic: mds: fix potential re-evaluate stray dentry in _unlink_local_finish

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

Merge pull request #26496 from batrick/i35975-2

mimic: qa: whitelist cap revoke warning

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

qa: whitelist cap revoke warning

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit 4367de377e68102f3c17c8dd85321c221d06d9dd)

Merge pull request #26309 from pdvian/wip-38192-mimic

mimic: rgw: Object can still be deleted even if s3:DeleteObject policy is set

Reviewed-by: Matt Benjamin <mbenjami@redhat.com>

Merge pull request #25782 from pdvian/wip-37693-mimic

mimic: mon: mark REMOVE_SNAPS messages as no_reply

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

Merge pull request #25785 from ukernel/mimic-37740

mimic: mds: extend reconnect period when mds is busy

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

Merge pull request #26252 from ukernel/mimic-37480

mimic: mds: handle fragment notify race

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

Merge pull request #26292 from pdvian/wip-38189-mimic

mimic: mds: broadcast quota message to client when disable quota

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

osd, test: Add test case with osd support for overdue PG scrubs and deep scrubs

Add trigger_deep_scrub osd command for testing
Publish stats when trigger_scrub/trigger_deep_scrub is used for testing
Add optional argument to trigger_scrub/trigger_deep_scrub
for amount of extra time to change last scrub stamps

Signed-off-by: David Zafman <dzafman@redhat.com>
(cherry picked from commit ef2dc05de00fbc5db8d864f83ee8d318286b6d6e)

Conflicts:
        src/osd/OSD.cc : Resolved in test_ops

Modify osd-scrub-repair.sh from master branch version to use mimic
mon_warn_not_scrubbed and mon_warn_not_deep_scrubbed config values.

mon: Fix scrub health warning handling

Make this mon_warn code clearer since it involves 2 values
Code used mon scrub interval instead of pg scrub interval
Fix scrub warniing handling use per-pool intervals when specified

Fixes: http://tracker.ceph.com/issues/37264
Signed-off-by: David Zafman <dzafman@redhat.com>
(cherry picked from commit 6a9895b97a2b2fb533092f294e45fede154a7f82)

Modified from master to leave config name alone and not to use ratio

common: Fix some configuration descriptions

osd_scrub_backoff_ratio
osd_deep_scrub_randomize_ratio

Signed-off-by: David Zafman <dzafman@redhat.com>
(cherry picked from commit aa4fc197858bfd94924c09aada1c408c8ef01273)

osd: Increase log level in PG::_update_calc_stats()

Signed-off-by: David Zafman <dzafman@redhat.com>
(cherry picked from commit c660ef1b1fe0812bd502ea0e7f446a88f2b08431)

test: Correct typo trying to call flush_pg_stats

Signed-off-by: David Zafman <dzafman@redhat.com>
(cherry picked from commit 879d89aace47bff63ddd08bd470ded7f1904eeaf)

Merge pull request #26129 from pdvian/wip-38039-mimic

mimic: rgw: fails to start on Fedora 28 from default configuration

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

Merge pull request #26266 from pdvian/wip-38147-mimic

mimic: rgw: "radosgw-admin bucket rm ... --purge-objects" can hang

Reviewed-by: Nathan Cutler <ncutler@suse.com>
Reviewed-by: J. Eric Ivancich <ivancich@redhat.com>

Merge pull request #26277 from ceph/wip-wjw-boost_noncopyable-mimic

mimic: common: fix missing include boost/noncopyable.hpp

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

mds: fix potential re-evaluate stray dentry in _unlink_local_finish

If MDS handles cap release before _unlink_local_finish, then when MDS
processes _unlink_local_finish, eval_stray might be re-entered twice,
so this will cause crashing.

Fixes: http://tracker.ceph.com/issues/38263
Signed-off-by: Zhi Zhang <zhangz.david@outlook.com>
(cherry picked from commit 72c93baac64211e9f02fa01e743451bf5d63d7a6)

Conflicts:
src/mds/MDCache.h : Resolved in notify_stray

mds: remove superfluous cap sequence increment

"cap->issue_norevoke(cap->issued())" was introduced by commit
10b1a5663f5 "mds: share inode max size after MDS recovers".
It is no longer required since commit 77515b7a3c7 "mds: increase
cap sequence when sharing max size"

Signed-off-by: "Yan, Zheng" <zyan@redhat.com>
(cherry picked from commit d86287fa3cfaf897041fb179e8f7bc7a8818592b)

mds: fix MDCache::finish_snaprealm_reconnect()

It creates MClientSnap message, but does not send it to client

Signed-off-by: "Yan, Zheng" <zyan@redhat.com>
(cherry picked from commit b4e7a66235450b9c30ebcaba97f03d371fac2c02)

Merge pull request #26304 from dzafman/wip-38208

mimic: mon: A PG with PG_STATE_REPAIR doesn't mean damaged data, PG_STATE_IN…

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

Merge pull request #26342 from ifed01/wip-ifed-fix-compress-mimic2

mimic: os/bluestore: do not assert on non-zero err codes from compress() call

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

Merge pull request #26354 from pdvian/wip-38239-mimic

mimic: qa/tasks/radosbench: default to 64k writes

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

Revert "msg/async: bump global_seq when retrying connection"

This reverts commit ed65a7a53aedfd56343fdaefdeb59c22256f01b7.

See: https://github.com/ceph/ceph/pull/25956

Seems we don't want [39e8de2](https://github.com/ceph/ceph/commit/39e8de2d72e2541323f8ed9d8b68075fcbd02701)
to be backported.
It turns out the old async-connection implementation will
automatically bump the global_seq each time it switches to
the STATE_CONNECTING state and restarts the connecting process.

The same applies to mimic.

Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>

os/filestore: ceph_abort() on fsync(2) or fdatasync(2) failure

If we get an error from f[data]sync that is always a fatal error.

Fixes: http://tracker.ceph.com/issues/38258
Signed-off-by: Sage Weil <sage@redhat.com>
(cherry picked from commit 3892b81303bcc35056ca06371651e13b7a4c4f2a)

client: flush cap release on session flush

MDS expects the completed cap release prior to responding to the session flush
for cache drop.

Fixes: http://tracker.ceph.com/issues/38009
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit 59eca7a17816317d5064840745ea5e28d3349d1a)

Conflicts:
src/client/Client.cc
- use m.detach() instead of std::move(m) as argument to send_message()

Merge pull request #26344 from liewegas/wip-centos-notmcalloc-only-mimic

mimic: qa: don't use notcmalloc flavor unless we need it

Merge pull request #25944 from wido/mimic-zabbix-pg

mimic: mgr/zabbix: Send more PG information to Zabbix

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

Merge pull request #26376 from tchaikov/mimic-38266

mimic: tests: Allow multi instances of "make tests" on the same machine

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

test: listen on random port in tests which start ceph-mon

See-also: http://tracker.ceph.com/issues/36737
Signed-off-by: Kefu Chai <kchai@redhat.com>
(cherry picked from commit 94a84b6f5adad75b3cbf956447ee5d9caf1018d2)

test: write log file to current directory

instead of writing log files to '/tmp', write them to current directory.

by default, ctest runs the tests in ${CMAKE_CURRENT_BINARY_DIR}, so, in
this case, the log files are located in build/src/test/. this ensure
that current user has enough previledge to remove the log files.

this also allows multiple instances of 'unittest_log' to run in parallel.

See-also: http://tracker.ceph.com/issues/36737
Signed-off-by: Kefu Chai <kchai@redhat.com>
(cherry picked from commit a13a6f1d0b39a541896591cd04fc1980bb1acfee)

Conflicts:
src/log/test.cc: trivial resolution

Merge pull request #26373 from pdvian/wip-38096-mimic

mimic: doc: rados/configuration: refresh osdmap section

Reviewed-by: Nathan Cutler <ncutler@suse.com>

Merge pull request #25721 from ashishkumsingh/wip-37689-mimic

mimic: ceph-objectstore-tool: Dump hashinfo

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

Merge pull request #26077 from ashishkumsingh/wip-37992-mimic

mimic: osd: keep using cache even if op will invalid cache

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

Merge pull request #26235 from pdvian/wip-38109-mimic

mimic: mgr will refuse connection from the monitor who starts behind it

Reviewed-by: Tim Serong <tserong@suse.com>

Merge pull request #26239 from pdvian/wip-38106-mimic

mimic: osd/PG.cc: account for missing set irrespective of last_complete

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

Merge pull request #26260 from pdvian/wip-38143-mimic

mimic: os/bluestore: fixup access a destroy cond cause deadlock or undefine

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

doc/rados/configuration: refresh osdmap section

"osd map cache size" and "osd map message max" were reduced in commit
855955e58e63 ("osd: reduce size of osdmap cache, messages").

"osd map cache bl size" and "osd map cache bl inc size" were removed
six years ago.

Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
(cherry picked from commit ae88c63d0441b1dffcc1dc469f4693dc909650bf)

Merge pull request #26288 from neha-ojha/wip-wl-slow-mimic

mimic: qa/tasks/thrashosds-health.yaml: whitelist "slow request"

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

Merge pull request #26291 from pdvian/wip-38187-mimic

mimic: deep fsck fails on inspecting very large onodes

Reviewed-by: xie xingguo <xie.xingguo@zte.com.cn>
Reviewed-by: Igor Fedotov <ifedotov@suse.com>

Merge pull request #26324 from smithfarm/wip-38111-mimic

mimic: osd: force-backfill sets forced_recovery instead of forced_backfill in 13.2.1

Reviewed-by: David Zafman <dzafman@redhat.com>
Reviewed-by: xie xingguo <xie.xingguo@zte.com.cn>

qa/tasks/radosbench: default to 64k writes

The default rados bench is 4MB writes, but this tends to fill up our
tiny test clusters.  Use 64KB instead.

Signed-off-by: Sage Weil <sage@redhat.com>
(cherry picked from commit 78070b96bf72c0724caad0afdb3d2ae0197cc95b)

qa/tasks/keystone.py: no need for notcmalloc in example

Signed-off-by: Sage Weil <sage@redhat.com>
(cherry picked from commit 9c729a78be29b7ad017dc652039057a0861c48ba)

qa/suites/rgw/tempest/tasks/rgw_tempest: no need for notcmalloc

Signed-off-by: Sage Weil <sage@redhat.com>
(cherry picked from commit b8b398e1abfdfc9ad7b290378b5698c31675413b)

os/bluestore: do not assert on non-zero err codes from  compress() call

Fixes: https://tracker.ceph.com/issues/37839
Signed-off-by: Igor Fedotov <ifedotov@suse.com>
(cherry picked from commit 4083bc29b6b4191a5ef3c17976eeac091626f1b0)

Merge pull request #26329 from dzafman/wip-37393

mimic: osd-backfill-stats.sh fails in rados/standalone/osd.yaml

Fixes: http://tracker.ceph.com/issues/37393
Reviewed-by: Neha Ojha <nojha@redhat.com>

Merge pull request #26333 from alfredodeza/mimic-rm38229

mimic: doc: pin the version for "breathe" to 4.1.11

Reviewed-by: Andrew Schoen <aschoen@redhat.com>

doc: pin the version for `breathe` to 4.1.11

It has the compatibility needed for Sphinx version 1.6.3

Signed-off-by: Alfredo Deza <adeza@redhat.com>

test: osd-backfill-stats.sh Fix check of multi backfill OSDs, skip remapped test

Signed-off-by: David Zafman <dzafman@redhat.com>
(cherry picked from commit 6d59fdd9710b79685325e8c445948745e7ec0002)

test: Fix test to detect a test setup failure

Signed-off-by: David Zafman <dzafman@redhat.com>
(cherry picked from commit c1b2bd7f16fe4ffb63baa48e8a154f275bd20126)

test: osd-backfill-stats.sh: Fix subtests to get primary which can change

Fixes: http://tracker.ceph.com/issues/35982
Signed-off-by: David Zafman <dzafman@redhat.com>
(cherry picked from commit ef6940fbb6c084b7f9abe893a18faef75d3a5680)

qa/standalone/osd/osd-backfill-stats: fixes

Grep from the primary's log, not every osd's log.

For the backfill_remapped task in particular, after the pg_temp change it
just so happens that the primary changes across the pool size change and
thus two different primaries do (some) backfill.  Fix that test to pass
the correct primary.

Other tests are unaffected as they do not (happen to) trigger a primary
change and already satisfied the (removed) check that only one OSD does
backfill.

Signed-off-by: Sage Weil <sage@redhat.com>
(cherry picked from commit f47921f29397257f77c946eef7ba4346bf6481a4)

Merge pull request #26294 from jan--f/25390-mimic

[mimic] ceph-volume: revert partition as disk

Reviewed-by: Andrew Schoen <aschoen@redhat.com>

osd/PG: fix misused FORCE_RECOVERY[BACKFILL] flags

__set_force_backfill__ should set PG_STATE_FORCED_BACKFILL
instead of PG_STATE_FORCED_RECOVERY.

Fixes: http://tracker.ceph.com/issues/27985
Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>
(cherry picked from commit 60ef742a84d5201b2a6e6c1780741da0b31c49a8)

rgw: cleanup for RGWDeleteObj::verify_permission()

- if the 'Effect' field of user policy and bucket policy
  is set to 'Allow', RGWDeleteObj::verify_permission()
  should return 0 instead of true;

- as the same with bucket policy, if the 'Effect' field
  of user policy is set to 'Deny', RGWDeleteObj::verify_permission()
  should return -EACCES instead of false;

Signed-off-by: Enming.Zhang <zvampirem77@gmail.com>
(cherry picked from commit 86334645cf96c03515d30804e2df0bdf22916fb5)

Conflicts:
src/rgw/rgw_op.cc : Resolve in verify_permission

rgw: fix obj can still be deleted even if deleteobject policy is set

Fixes: http://tracker.ceph.com/issues/37403
Signed-off-by: Enming.Zhang <zvampirem77@gmail.com>
(cherry picked from commit 84879e915eb3f01b453aa2138cd44c09925e7cd1)

Conflicts:
src/rgw/rgw_op.cc : Resolved in verify_permission

mon: A PG with PG_STATE_REPAIR doesn't mean damaged data, PG_STATE_INCONSISTENT does

Fixes: http://tracker.ceph.com/issues/38070
Signed-off-by: David Zafman <dzafman@redhat.com>
(cherry picked from commit a4b1574532ee4f789b30634f0eca03cc30ddcf8d)

Merge pull request #25856 from ashishkumsingh/wip-37821-mimic

mimic: qa/rados/upgrade: align thrashing with upgrade suite, don't import/export pgs

Reviewed-by: Nathan Cutler <ncutler@suse.com>
Reviewed-by: Neha Ojha <nojha@redhat.com>

Merge pull request #26074 from wido/mimic-mgr-perf

mimic: mgr: add get_latest_counter() to C++ -> Python interface.

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

Merge pull request #25957 from pdvian/wip-37900-mimic

mimic: [journal] max journal order is incorrectly set at 64

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

Merge pull request #25987 from ashishkumsingh/wip-37925-mimic

mimic: common: fix for broken rbdmap parameter parsing

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

ceph-volume: add inventory call to functional tests

Signed-off-by: Jan Fajerski <jfajerski@suse.com>
(cherry picked from commit a183a28c5e423c5658a37f7a1d3add9d64c6d683)

ceph-volume: add partition support to Device

Signed-off-by: Jan Fajerski <jfajerski@suse.com>
(cherry picked from commit 05ac680cb37fd15fdeadaea86780bf1b4f2d566a)

Revert "ceph-volume tests.util ensure partitions are listed as top-level keys"

This reverts commit 6dc01772245e557b73551111f7c48cc60aec0893.

Signed-off-by: Jan Fajerski <jfajerski@suse.com>
(cherry picked from commit 3c1c32cbc2736d72295f1bd7921c4a3cd2f4727e)

Revert "ceph-volume util.disk include partitions as top level devices in get_devices"

This reverts commit 7f3c3598491e556ec1b93ea0a54f52144202fab6.
Fixes: http://tracker.ceph.com/issues/37506
Signed-off-by: Jan Fajerski <jfajerski@suse.com>
(cherry picked from commit 50bc44f1c2327ad38504950fe62820a44ead9b14)

Merge pull request #25385 from ifed01/wip-ifed-osd-mkfs-fix

mimic: osd/OSD: OSD::mkfs asserts when reusing disk with existing superblock.

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

Merge pull request #26201 from smithfarm/wip-37688-mimic

mimic: tests: Command failed on smithi191 with status 1: 'sudo yum -y install ceph-radosgw'

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

Merge pull request #26206 from neha-ojha/wip-37902-mimic

mimic: osd/mon: fix upgrades for pg log hard limit

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

doc: pglog_hardlimit flag recommendations

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

mds: broadcast quota message to client when disable quota

When disable quota by setting quota.max_files or quota.max_bytes
to zero, client does not receive the quota broadcast message.

Signed-off-by: Junhui Tang <tangjunhui@sangfor.com.cn>
Fixes: https://tracker.ceph.com/issues/38054
(cherry picked from commit 061ace2eb7c08d271db4d5d08ffa91ae832715f5)

os/bluestore: cap reads issued during deep fsck

Fixes: https://tracker.ceph.com/issues/38065
Signed-off-by: Igor Fedotov <ifedotov@suse.com>
(cherry picked from commit ced308000aef6050ca1fdbdfe3b4b63a887a1706)

os/bluestore: do not buffer read data on deep fsck.

Signed-off-by: Igor Fedotov <ifedotov@suse.com>
(cherry picked from commit a4e55dd352a28c48029a0f74cb253673e1aca109)

Merge pull request #25348 from ifed01/wip-ifed-fix-expand-mimic

mimic: core: os/bluestore_tool: fix bluefs expand

Reviewed-by: Nathan Cutler <ncutler@suse.com>

Merge pull request #25525 from smithfarm/wip-36751-mimic

mimic: cmake: do not pass -B{symbolic,symbolic-functions} to linker on FreeBSD

Reviewed-by: Kefu Chai <kchai@redhat.com>
Reviewed-by: Willem Jan Withagen <wjw@digiware.nl>

Merge pull request #25531 from smithfarm/wip-37642-mimic

mimic: ceph-create-keys: fix octal notation for Python 3 without losing compatibility with Python 2

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

qa/tasks/thrashosds-health.yaml: whitelist "slow request"

https://github.com/ceph/ceph/pull/25824 brought back slow requests in luminous
OSD logs. This is resulting in the upgrade/luminous-x/stress-split* suites in
mimic to show these warnings.

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

qa: update krbd tests for zeroout

Discard no longer guarantees zeroing, use BLKZEROOUT and "fallocate -z"
instead (blkdiscard(8) in xenial doesn't support -z).

Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
(cherry picked from commit 04f5b343f9987f0cf67509c2a76ddf7ef46288e5)

Merge pull request #25857 from ashishkumsingh/wip-37822-mimic

mimic: mds: log evicted clients to clog/dbg

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

Merge pull request #25885 from pdvian/wip-37818-mimic

mimic: mds: remove wrong assertion in Locker::snapflush_nudge

Reviewed-by: Yan, Zheng <zyan@redhat.com>
Reviewed-by: Venky Shankar <vshankar@redhat.com>

Merge pull request #26010 from smithfarm/wip-37819-mimic

mimic: mds: create separate config for heartbeat timeout

Reviewed-by:  Venky Shankar <vshankar@redhat.com>
Reviewed-by: Yan, Zheng <zyan@redhat.com>

Merge pull request #26055 from smithfarm/wip-37898-mimic

mimic: mds: purge queue recovery hangs during boot if PQ journal is damaged

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

Merge pull request #26214 from pdvian/wip-38101-mimic

mimic: mds: trim cache after journal flush

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

common: fix missing include boost/noncopyable.hpp

Clang complains:
```
In file included from /home/jenkins/workspace/ceph-master/src/osd/ReplicatedBackend.cc:15:
In file included from /home/jenkins/workspace/ceph-master/src/osd/ReplicatedBackend.h:18:
In file included from /home/jenkins/workspace/ceph-master/src/osd/PGBackend.h:27:
In file included from /home/jenkins/workspace/ceph-master/src/osd/PGTransaction.h:25:
/home/jenkins/workspace/ceph-master/src/common/inline_variant.h:88:64: error: expected class name
struct generic_visitor : boost::static_visitor<Result>, boost::noncopyable
                                                               ^
1 error generated.
gmake[2]: *** [src/osd/CMakeFiles/osd.dir/build.make:141: src/osd/CMakeFiles/osd.dir/ReplicatedBackend.cc.o] Error 1
gmake[2]: *** Waiting for unfinished jobs....
```

Signed-off-by: Willem Jan Withagen <wjw@digiware.nl>
(cherry picked from commit 0654099d4b2f9f9423d7e8cf00e9e3351a24ec67)

Merge pull request #26009 from sebastian-philipp/dashboard-mirror_mode

mimic: mgr/dashboard: Set mirror_mode to None

Reviewed-by: Lenz Grimmer <lgrimmer@suse.com>
Reviewed-by:  Volker Theile <vtheile@suse.com>

Merge pull request #26208 from vumrao/wip-vumrao-38107

mimic: mgr/DaemonServer: log pgmap usage to cluster log

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

rgw: `radosgw-admin bucket rm ... --purge-objects` can hang...

This command can hang (i.e., enter an infinite loop) due to
problematic bucket index entries left as a result of bug
https://tracker.ceph.com/issues/38007 .

The fix is to ignore the false bucket index entries -- since they do
not represent actual objects -- and remove all actual objects in the
bucket, so that bucket itself can be removed.

This fixes the both code paths whether `--bypass-gc` is specified or
not.

Furthermore, to made these operations more efficient, the internal
listing of the bucket is done unordered. This would improve behavior
when removing buckets with a large number of objects.

Fixes: http://tracker.ceph.com/issues/38134
Signed-off-by: J. Eric Ivancich <ivancich@redhat.com>
(cherry picked from commit 5b3dc38c628728651b26415ab3d2190cf06b9e39)

os/bluestore: fixup access a destroy cond cause deadlock or undefine
              behavior.

Fixes: https://tracker.ceph.com/issues/37733
Signed-off-by: linbing <linbing@t2cloud.net>
(cherry picked from commit dbbde306484276251cbd7a7b7a6fde0093f98ba2)

Conflicts:
src/os/bluestore/BlockDevice.h : Resolved in try_aio_wake

mds: bump mds protocol version

Signed-off-by: "Yan, Zheng" <zyan@redhat.com>

mds: handle fragment notify race

In the nornal case, mds does not trim dir inode whose child dirfrags
are likely being fragmented (see trim_inode()). But when fragmenting
subtree roots, following race can happen:

- mds.a (auth mds of dirfrag) sends fragment_notify message to
  mds.c and drops wrlock on dirfragtreelock.
- mds.b (auth mds of dir inode) changes dirfragtreelock state to
  SYNC and send lock message mds.c
- mds.c receives the lock message and changes dirfragtreelock state
  to SYNC
- mds.c trim dirfrag and dir inode from its cache
- mds.c receives the fragment_notify message

The fix is asking replicas to ack fragment_notify message, unlocking
dirfragtreelock after mds gets all acks.

Fixes: http://tracker.ceph.com/issues/36035
Signed-off-by: "Yan, Zheng" <zyan@redhat.com>
(cherry picked from commit 72887980a7d6517740a2841d6014a92a2d1c3063)

 Conflicts:
src/mds/MDCache.cc
src/mds/MDCache.h
src/mds/Locker.cc
src/messages/MMDSFragmentNotify.h
src/msg/Message.h
src/msg/Message.cc

mds: don't wait for snapclient sync when handling reconnect

It's no longer required since commit 93e72677575 "mds: send snap
related messages centrally during mds recovery".

Signed-off-by: "Yan, Zheng" <zyan@redhat.com>
(cherry picked from commit 0565fb3e646a9c133b621dd6ed4f26bed9a9edbc)