osd/PrimaryLogPG: do not expect FULL_TRY ops to get resent

The objecter will not resend FULL_TRY requests that were sent prior to
becoming full, so we should not discard them.

Fixes: http://tracker.ceph.com/issues/19430
Signed-off-by: Sage Weil <sage@redhat.com>
(cherry picked from commit 3f7acdbc9a942fd18937dbcf07fbc7b752c50ba3)

Merge pull request #16069 from smithfarm/wip-20345-kraken

kraken: make check fails with Error EIO: load dlopen(build/lib/libec_FAKE.so): build/lib/libec_FAKE.so: cannot open shared object file: No such file or directory

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

ceph.spec.in, debian/control: Add bc to build dependencies

The bc is missing for ceph-helpers.sh

Fixes: http://tracker.ceph.com/issues/18876
Signed-off-by: Kyr Shatskyy <kyrylo.shatskyy@gmail.com>
(cherry picked from commit 3ff4be6d6896346b7b9ec7f158fcde8866faeb38)

qa/workunits/ceph-helpers: do not error out if is_clean

it would be a race otherwise, because we cannot be sure that the cluster
pgs are not all clean or not when run_osd() returns, but we can be sure
that they are expected to active+clean after a while. that's what
wait_for_clean() does.

Signed-off-by: Kefu Chai <kchai@redhat.com>
(cherry picked from commit 6cb4503a40ae4ebee5690fda024cae8d1a506bce)

qa/workunits/ceph-helpers: display rejected string

Signed-off-by: Kefu Chai <kchai@redhat.com>
(cherry picked from commit 0196e154ed2e164cf55b0d7ed9f9cdd1f4f50100)

Merge pull request #15869 from smithfarm/wip-swift-task-move-kraken

tests: move swift.py task from teuthology to ceph, phase one (kraken)

tests: swift.py: tweak imports

The ".." form only works within the teuthology repo. With swift.py now in the
Ceph repo, we have to be explicit.

Error message was: "ValueError: Attempted relative import beyond toplevel
package

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

Merge branch 'master' of /home/smithfarm/src/ceph/upstream/teuthology into wip-swift-task-move-kraken

tests: move swift.py task to qa/tasks

In preparation for moving this task from ceph/teuthology.git into ceph/ceph.git

The move is necessary because jewel-specific changes are needed, yet teuthology
does not maintain a separate branch for jewel. Also, swift.py is a
Ceph-specific task so it makes more sense to have it in Ceph.

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

Merge pull request #13181 from smithfarm/wip-18548-kraken

kraken: rgw: multisite: segfault after changing value of rgw_data_log_num_shards

Reviewed-by: Orit Wasserman <owasserm@redhat.com>

Merge pull request #13838 from smithfarm/wip-19049-kraken

kraken: rgw: multisite: some yields in RGWMetaSyncShardCR::full_sync() resume in incremental_sync()

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

Merge pull request #13224 from smithfarm/wip-18780-kraken

kraken: rgw: Swift API: spurious newline after http body causes weird errors

Reviewed-by: Orit Wasserman <owasserm@redhat.com>

Merge pull request #13174 from smithfarm/wip-18711-kraken

kraken: rgw: slave zonegroup cannot enable the bucket versioning

Reviewed-by: Orit Wasserman <owasserm@redhat.com>

rgw: fix off-by-one in RGWDataChangesLog::get_info

Fixes: http://tracker.ceph.com/issues/18488
Signed-off-by: Casey Bodley <cbodley@redhat.com>
(cherry picked from commit a0974fdcf62e60cf31bc15588e7b718da6f6ade3)

radosgw/swift: clean up flush / newline behavior.

The current code emits a newline after swift errors, but fails
to account for it when it calculates 'content-length'.  This results in
some clients (go github.com/ncw/swift) producing complaints about the
unsolicited newline such as this,
Unsolicited response received on idle HTTP channel starting with "\n"; err=<nil>

This logic eliminates the newline on flush.  This makes the content length
calculation correct and eliminates the stray newline.

There was already existing separator logic in the rgw plain formatter
that can emit a newline at the correct point.  It had been checking
"len" to decide if previous data had been emitted, but that's reset to 0
by flush().  So, this logic adds a new per-instance variable to separately
track state that it emitted a previous item (and should emit a newline).

Fixes: http://tracker.ceph.com/issues/18473
Signed-off-by: Marcus Watts <mwatts@redhat.com>
Signed-off-by: Matt Benjamin <mbenjamin@redhat.com>
(cherry picked from commit 5f229d6a33eae4906f22cdb90941835e47ee9f02)

rgw: fix for broken yields in RGWMetaSyncShardCR

Fixes: http://tracker.ceph.com/issues/18076
Signed-off-by: Casey Bodley <cbodley@redhat.com>
(cherry picked from commit e62d48a9bf2e309eab1a863f167af5267ebcc371)

rgw: complete versioning enablement after sending it to meta master

Fixes: http://tracker.ceph.com/issues/18003
Signed-off-by: Orit Wasserman <owasserm@redhat.com>
(cherry picked from commit 2d8aafb9dbe64bd9dd2b7d5ed50c6e9550cbe1ab)

Merge pull request #15421 from dzafman/wip-20125

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

osd: Object level shard errors are tracked and used if no auth available

Shards with object mismatch are tracked to mark them inconsistent
Fix test because storing omap_digest in object_info not behaving as before

Fixes: http://tracker.ceph.com/issues/20089
Signed-off-by: David Zafman <dzafman@redhat.com>
(cherry picked from commit 1cacbea763c7aabfeaaf4bd5e878301044184117)

Merge pull request #14345 from ceph/wip-ceph-disk-fix-kraken

ceph-disk: Add fix subcommand kraken back-port

Reviewed-by: Loic Dachary <ldachary@redhat.com>
Reviewed-by: Nathan Cutler <ncutler@suse.com>

ceph-disk: Add --system option for fix command

This adds the ability to restore the labels of the underlying system
data in addition to ceph data.

Signed-off-by: Boris Ranto <branto@redhat.com>
(cherry picked from commit 8d81af42fd507c7b92c8279eb114b0a733ac1da6)

ceph-disk: Add more fix targets

It turns out I forgot several more directories that needs to be fixed by
this script. We need to fix /var/log/ceph, /var/run/ceph and /etc/ceph
as well.

Signed-off-by: Boris Ranto <branto@redhat.com>
(cherry picked from commit ae139307d6b2bfba47e21d29d6dbd3c8dc01b5b5)

ceph-disk: Add unit test for fix command

This will simulate the command* functions to not actually run anything
thus excercising the python code directly. It also checks that the
proper (sub-strings) are in the output.

Signed-off-by: Boris Ranto <branto@redhat.com>
(cherry picked from commit 1ec53dee9a690134936bdc3a09c9a02fecf13a9d)

ceph-disk: Add fix subcommand

This subcommand will fix the SELinux labels and/or file permissions on
ceph data (/var/lib/ceph).

The command is also optimized to run the commands in parallel (per
sub-dir in /var/lib/ceph) and do restorecon and chown at the same time
to take advantage of the caching mechanisms.

Signed-off-by: Boris Ranto <branto@redhat.com>
(cherry picked from commit 6d5d30f6ed7538271579cc2ef4e2e364f01a4a6f)

Merge pull request #14983 from liewegas/wip-denc-dump-kraken

ceph-object-corpus: kraken objects (kraken)

test/encoding/readable.sh: join children at finish

Signed-off-by: Sage Weil <sage@redhat.com>
(cherry picked from commit c8a349eb4603be295a6f0e1683105cb93bfcd066)

test/encoding/readable: fix passing of results to parent

Signed-off-by: Sage Weil <sage@redhat.com>
(cherry picked from commit 1325ba672d5b5c8a12722a395ad029c98d1ff236)

messages/MCommand: fix type on decode

Wow, this has been broken since v0.38, but apparently
the message never made it into the object corpus so
we never noticed!

In reality the bug is harmless: decode_message() will
set_header which clobbers whatever version the default
ctor fills in, so this only affects ceph-dencoder's
test.

Signed-off-by: Sage Weil <sage@redhat.com>
(cherry picked from commit 3018b907c1d0375b5ae20802811591c0a46b61be)

ceph-object-corpus: kraken objects

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

test/encoding/generate-corpus-objects: fix typo

Signed-off-by: Sage Weil <sage@redhat.com>
(cherry picked from commit b89dff472c86c435fcee767ec9d992b308c4d816)

test/encoding/generate-corpus-objects.sh: simplify object corpus population

Script that generates corpus objects.

Signed-off-by: Sage Weil <sage@redhat.com>
(cherry picked from commit b38b663a71fdc37a2d5c3341836b6716d576cecc)

qa/workunits/rgw/run-s3tests.sh: run s3-tests

This works out of the box with a vstart environment and

 RGW=1 ../src/vstart.sh -n -l
 PATH=bin:$PATH ../qa/workunits/rgw/run-s3tests.sh

Signed-off-by: Sage Weil <sage@redhat.com>
(cherry picked from commit 40fe1d181cf8d31b80db6f8f99394bd66d2affd5)

include/denc: support ENCODE_DUMP

Signed-off-by: Sage Weil <sage@redhat.com>
(cherry picked from commit 97987b4ace737b86a4dc2dce94ce9d330fbe903c)

mgr/DaemonServer: do not crash on bad authorizer

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

Merge pull request #14506 from smithfarm/wip-19119-kraken

kraken: doc: PendingReleaseNotes: warning about 'osd rm ...' and #13733

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

Merge pull request #14692 from smithfarm/wip-fix-hammer-jewel-x

qa/suites/upgrade: add tiering test to hammer-jewel-x

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

Merge pull request #14425 from smithfarm/wip-19564-kraken

kraken: build/ops: ceph-base missing dependency for psmisc in Ubuntu Xenial

Reviewed-by: Ken Dreyer <kdreyer@redhat.com>

Merge pull request #13494 from shinobu-x/wip-18516-kraken

kraken: build/ops: systemd: Start OSDs after MONs

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

Merge pull request #14588 from asheplyakov/19618-kraken

kraken: mon/MonClient: make get_mon_log_message() atomic

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

Merge pull request #13543 from shinobu-x/wip-18954-kraken

kraken: ceph-disk prepare get wrong group name in bluestore

Reviewed-by: Loic Dachary <ldachary@redhat.com>

Merge pull request #12746 from SUSE/wip-18387-kraken

kraken: tests: use ceph-kraken branch for s3tests

Reviewed-by: Orit Wasserman <owasserm@redhat.com>

qa/suites/upgrade: add tiering test to hammer-jewel-x

Related: http://tracker.ceph.com/issues/19185
Signed-off-by: Kefu Chai <kchai@redhat.com>
(cherry picked from commit e8c04a027b5d108b30e63ab759760add57d43198)

Merge pull request #14620 from smithfarm/wip-19659-kraken

kraken: librbd: corrected resize RPC message backwards compatibility

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

Merge pull request #14574 from smithfarm/wip-19620-kraken

kraken: cephfs: MDS server crashes due to inconsistent metadata.

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

Merge pull request #14573 from smithfarm/wip-19483-kraken

kraken: cephfs: No output for ceph mds rmfailed 0 --yes-i-really-mean-it command

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

Merge pull request #14572 from smithfarm/wip-19335-kraken

kraken: cephfs: MDS heartbeat timeout during rejoin, when working with large amount of caps/inodes

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

Merge pull request #14571 from smithfarm/wip-19045-kraken

kraken: cephfs: buffer overflow in test LibCephFS.DirLs

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

Merge pull request #14570 from smithfarm/wip-18950-kraken

kraken: cephfs: mds/StrayManager: avoid reusing deleted inode in StrayManager::_purge_stray_logged

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

Merge pull request #14569 from smithfarm/wip-18899-kraken

kraken: cephfs: test_open_inode fails

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

Merge pull request #14568 from smithfarm/wip-18706-kraken

kraken: cephfs: fragment space check can cause replayed request fail

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

Merge pull request #14567 from smithfarm/wip-18700-kraken

kraken: cephfs: client: fix the cross-quota rename boundary check conditions

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

Merge pull request #14566 from smithfarm/wip-18616-kraken

kraken: cephfs: segfault in handle_client_caps

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

Merge pull request #14565 from smithfarm/wip-18566-kraken

kraken: cephfs: MDS crashes on missing metadata object

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

Merge pull request #14564 from smithfarm/wip-18562-kraken

kraken: cephfs: test_client_recovery.TestClientRecovery fails

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

Merge pull request #14563 from smithfarm/wip-18552-kraken

kraken: cephfs: ceph-fuse crash during snapshot tests

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

Merge pull request #14521 from smithfarm/wip-19462-kraken

kraken: doc: rgw: admin ops: fix the quota section

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

Merge pull request #14641 from smithfarm/wip-19693-kraken

kraken: [test] test_notify.py: rbd.InvalidArgument: error updating features for image test_notify_clone2

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

Merge pull request #14533 from smithfarm/wip-18501-kraken

kraken: rbd-mirror: potential race mirroring cloned image

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

Merge pull request #14534 from smithfarm/wip-18549-kraken

kraken: rbd: 'metadata_set' API operation should not change global config setting

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

Merge pull request #14536 from smithfarm/wip-18557-kraken

kraken: rbd: 'rbd bench-write' will crash if --io-size is 4G

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

Merge pull request #14537 from smithfarm/wip-18601-kraken

kraken: rbd: Add missing parameter feedback to 'rbd snap limit'

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

Merge pull request #14538 from smithfarm/wip-18632-kraken

kraken: rbd: [qa] crash in journal-enabled fsx run

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

Merge pull request #14622 from smithfarm/wip-19037-kraken

kraken: rbd-mirror: deleting a snapshot during sync can result in read errors

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

Merge pull request #14545 from smithfarm/wip-19324-kraken

kraken: rbd: [api] temporarily restrict (rbd_)mirror_peer_add from adding multiple peers

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

test: rbd master/slave notify test should test active features

Fixes: http://tracker.ceph.com/issues/19692
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit 0dcba41cba96566d0b8da54cf0316d523b88ded2)

qa/suites: client-upgrade/jewel-client-x should test multiple features

Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit 7ab93b7779b293ceda325a6d1cdb446cf624a6e3)

Merge pull request #14487 from zmc/wip-kraken-workunit

qa/tasks/workunit: Backport repo fixes from master

Merge pull request #14531 from smithfarm/wip-18493-kraken

kraken: rbd-mirror: sporadic image replayer shut down failure

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

Merge pull request #14532 from smithfarm/wip-18495-kraken

kraken: rbd: Possible deadlock performing a synchronous API action while refresh in-progress

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

librbd: corrected resize RPC message backwards compatibility

Commit d1f2c557 incorrectly changed the order of variables within
the payload. This resulted in breaking the resize RPC message
with older versions of Ceph.

Fixes: http://tracker.ceph.com/issues/19636
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit 9db305a99fd945059a173f5ae8ae61744dd28615)

Merge pull request #14095 from shinobu-x/wip-19319-kraken

kraken: RadosImport::import should return an error if Rados::connect fails

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

mon/MonClient: make get_mon_log_message() atomic

* LogClient: move reset_session() into get_mon_log_message() and add a
  "flush" param to the latter. so it can get_mon_log_message()
  atomically. otherwise another call changing the log queue could sneak
  in between reset_session() and get_mon_log_message().
* MonClient: add a "flush" param to do_send() so we can reset the
  LogClient session once we are connected to a monitor.

Fixes: http://tracker.ceph.com/issues/19427
Signed-off-by: Kefu Chai <kchai@redhat.com>
(cherry picked from commit 5215e291da2b527d85e129eda86043490843178e)

Conflicts:
src/mon/MonClient.cc: handle_auth: replaced 'log_client->reset_session();
send_log();' sequence with newly introduced 'send_log(true);' like
the original patch does

Merge pull request #14323 from shinobu-x/wip-19371-kraken

kraken: common: monitor creation with IPv6 public network segfaults

Reviewed-by: Kefu Chai <kchai@redhat.com>
Reviewed-by: Nathan Cutler <ncutler@suse.com>

Merge pull request #13284 from shinobu-x/wip-18599-kraken

kraken: osd: os/bluestore: fix statfs to not include DB partition in free space

Reviewed-by: Sage Weil <sage@redhat.com>
Reviewed-by: Nathan Cutler <ncutler@suse.com>

Merge pull request #13046 from SUSE/wip-18554-kraken

kraken: mon: peon wrongly delete routed pg stats op before receive pg stats ack

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

Merge pull request #14480 from dillaman/wip-19467-kraken

kraken: librbd: is_exclusive_lock_owner API should ping OSD

Reviewed-by: Jason Dillaman <dillaman@redhat.com>
Reviewed-by: Mykola Golub <mgolub@mirantis.com>

Merge pull request #13877 from smithfarm/wip-19178-kraken

kraken: rgw: anonymous user error code of getting object is not consistent with SWIFT

Reviewed-by: Radoslaw Zarzynski <rzarzynski@mirantis.com>

Merge pull request #13866 from smithfarm/wip-19157-kraken

kraken: rgw: health check errors out incorrectly

Reviewed-by: Radoslaw Zarzynski <rzarzynski@mirantis.com>

Merge pull request #13843 from smithfarm/wip-19146-kraken

kraken: rgw: a few cases where rgw_obj is incorrectly initialized

Reviewed-by: Orit Wasserman <owasserm@redhat.com>

Merge pull request #13829 from smithfarm/wip-18898-kraken

kraken: no http referer info in container metadata dump in swift API

Reviewed-by: Radoslaw Zarzynski <rzarzynski@mirantis.com>

Merge pull request #13780 from smithfarm/wip-18896-kraken

kraken: should parse the url to http host to compare with the container referer acl

Reviewed-by: Radoslaw Zarzynski <rzarzynski@mirantis.com>

Merge pull request #12986 from Werkov/rgw-fix-tempurl-url-encoding

kraken: rgw: Use decoded URI when verifying TempURL

Reviewed-by: Radoslaw Zarzynski <rzarzynski@mirantis.com>

tools/cephfs: set dir_layout when injecting inodes

When we left this as zero, the MDS would interpret it was HASH_LINUX
rather than the default HASH_RJENKINS.  Potentially that
could cause problems if there perhaps were already dirfrags in
the metadata pool that were set up using rjenkins.  Mainly
it just seems more appropriate to explicitly set this field
rather than hit the fallback behaviour.

Related: http://tracker.ceph.com/issues/19406
Signed-off-by: John Spray <john.spray@redhat.com>
(cherry picked from commit 7d6d542885bd29b71214f9ca52bd26e9183c5d01)

mon: fix hiding mdsmonitor informative strings

Local `stringstream ss` declarations were hiding
the real variable used to feed back to the user.

Fixes: http://tracker.ceph.com/issues/16709
Signed-off-by: John Spray <john.spray@redhat.com>
(cherry picked from commit 00404ae9bd4cce0518a44d36d2d6a5612f4f9d04)

mds: reset heartbeat in export_remaining_imported_caps

This loop can be very long.

Fixes: http://tracker.ceph.com/issues/19118
Signed-off-by: John Spray <john.spray@redhat.com>
(cherry picked from commit 85071f1509beba4a390730e6a3a4332484646d63)

mds: heartbeat_reset in dispatch

Previously we only heartbeated in tick.  However, our locking is
not guaranteed to be fair, so on a super-busy dispatch queue it may be
possible for the heartbeat to time out while the tick() function
is waiting for mds_lock.

Fixes: http://tracker.ceph.com/issues/19118
Signed-off-by: John Spray <john.spray@redhat.com>
(cherry picked from commit 819394549af10532419d88742fae3a69d2ea487d)

test/libcephfs: avoid buffer overflow when testing ceph_getdents()

The buffer size should be at least "2 * sizeof(struct dirent)".
Otherwise, the code that checks dentry '..' overflow.

Fixes: http://tracker.ceph.com/issues/18941
Signed-off-by: "Yan, Zheng" <zyan@redhat.com>
(cherry picked from commit fa6671345b8f3a82dcd232f99e55a982b0a641f1)

mds/StrayManager: aviod reusing deleted inode in StrayManager::_purge_stray_logged

Signed-off-by: Zhi Zhang <zhangz.david@outlook.com>
(cherry picked from commit 4978e57419482384279d7e784a625f5e5c10961a)

tasks/cephfs: switch open vs. write in test_open_inode

Do the write after opening the file, so that we get good
behaviour wrt the change in Mount.open_background that uses
file existence to confirm that the open happened.

Signed-off-by: John Spray <john.spray@redhat.com>
(cherry picked from commit a027dba78fc8bc84ae39d7998b386ce21c01e1bf)

qa: fix race in Mount.open_background

Previously a later remote call could end up executing
before the remote python program in open_background
had actually got as far as opening the file.

Fixes: http://tracker.ceph.com/issues/18661
Signed-off-by: John Spray <john.spray@redhat.com>
(cherry picked from commit c6d91dd91252e703d08b8ac62ac6a47ee82c0bed)

mds: don't purge strays when mds is in clientreplay state

MDS does not trim log when it's in clientreplay state. If mds hang
at clientreplay state (due to bug), purging strays can submit lots
of log events and create very large mds log.

Signed-off-by: "Yan, Zheng" <zyan@redhat.com>
(cherry picked from commit 86bbc7fff02668077f27d0924ba3efe6544b77f6)

mds: skip fragment space check for replayed request

when handling replayed request, stray directory can be different
from the stray directory used by the original request. The fragment
space check for stray directory can fail.

Fixes: http://tracker.ceph.com/issues/18660
Signed-off-by: "Yan, Zheng" <zyan@redhat.com>
(cherry picked from commit afe889cbc5baab196567c2aad01f49fe90901fda)

client: fix the cross-quota rename boundary check conditions

We were previously rejecting a rename if either of the involved directories
was a quota root, even if the other directory was part of the same quota
"tree". What we really want to do is identify the correct quota root
(whether local or ancestral) for each directory and compare them. So
now we do.

Signed-off-by: Greg Farnum <gfarnum@redhat.com>
(cherry picked from commit 8e8892aa46accb519faa4bb9fecf66618f1b11b2)

mds: fix null pointer dereference in Locker::handle_client_caps

Locker::handle_client_caps delays processing cap message if the
corresponding inode is freezing or frozen. When the message gets
processed, client can have already closed the session.

Fixes: http://tracker.ceph.com/issues/18306
Signed-off-by: Yan, Zheng <zyan@redhat.com>
(cherry picked from commit e281a0b9c1fdeaf09f1b01f34cecd62e4f49d02e)

qa/tasks: add test_open_ino_errors

Validate that errors encountered during opening inos are properly
propagated

Signed-off-by: Yan, Zheng <zyan@redhat.com>
(cherry picked from commit 6526ecc084733b34129aa1f21085fa41fb53b785)

mds: propagate error encountered during opening inode by number

Fixes: http://tracker.ceph.com/issues/18179
Signed-off-by: Yan, Zheng <zyan@redhat.com>
(cherry picked from commit 2213cc2dcc0e8fb01bcae3863d0d8a4a1fd8873f)

qa/tasks/cephfs: fix kernel force umount

Fixes: http://tracker.ceph.com/issues/18396
Signed-off-by: Yan, Zheng <zyan@redhat.com>
(cherry picked from commit 4cdeeaac105d7824452a8b578632b7371275b45c)

client: fix Client::handle_cap_flushsnap_ack() crash

Struct CapSnap holds a reference to its parent inode. So erasing
struct CapSnap from Inode::cap_snaps may drop inode's last reference.
The inode gets freed in the middle of erasing struct CapSnap

Fixes: http://tracker.ceph.com/issues/18460
Signed-off-by: Yan, Zheng <zyan@redhat.com>
(cherry picked from commit 525c52fd491ed1ced385c8047872e3f557f8423f)

rbd: prevent adding multiple mirror peers to a single pool

The rbd-mirror daemon does not currently support replication
from multiple peers. Until that is supported, add a temporary
restriction to prevent confusion.

Fixes: http://tracker.ceph.com/issues/19256
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit c0c9d1014d57b3d5b95e7513fcc38d04b9ea5165)

rbd-mirror: retry object copy after -ENOENT error

Fixes: http://tracker.ceph.com/issues/18990
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit b4f36d5dc3f4f3cbb23f61cbb945b222248a50df)