rbd-mirror: ImageReplayer: pass registered client ID as external param

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

Merge pull request #7687 from yangdongsheng/ceph-disk_nargs

ceph-disk: make some arguments as required if necessary

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

Merge pull request #7640 from batrick/vstart-btrfs-missing

vstart.sh: silence a harmless msg where btrfs is not found

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

Merge pull request #7737 from trociny/wip-librados_test_stub-debug

test: more debug logging for TestWatchNotify

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

Merge pull request #7714 from SUSE/wip-qa-typo

doc: standardize @param (not @parma, @parmam, @params)

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

Merge pull request #7738 from trociny/test-librbd-poolname

test: create pools for rbd tests with different prefix

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

Merge pull request #7727 from jmunhoz/ceph-fix-14829

vstart.sh: avoid race condition starting rgw via vstart.sh

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

Merge pull request #7614 from trociny/wip-rbd-mirroring-image-reply

rbd-mirror: implement ImageReplayer

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

Merge pull request #7744 from dachary/wip-14838-ceph-disk-dmcrypt

ceph-disk: s/dmcrpyt/dmcrypt/

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

ceph-disk: s/dmcrpyt/dmcrypt/

http://tracker.ceph.com/issues/14838 Fixes: #14838

Signed-off-by: Frode Sandholtbraaten <fsa@basefarm.no>
Signed-off-by: Loic Dachary <loic@dachary.org>

Merge pull request #7688 from tchaikov/wip-14788

buffer: hide iterator_impl symbols

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

test: create pools for rbd tests with different prefix

This makes easier to find out what test fails to cleanup.

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

Merge pull request #7723 from tchaikov/wip-rgw-test

test/radosgw-admin: update the expected usage outputs

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

Merge pull request #7647 from yuyuyu101/wip-spdk

NVMEDevice: refactor probe/attach codes and support zero command

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

test: more debug for TestWatchNotify

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

NVMEDevice: accept spdk api name changes

Signed-off-by: Haomai Wang <haomai@xsky.com>

spdk: update spdk submodule to accept new interface changes

Signed-off-by: Haomai Wang <haomai@xsky.com>

NVMEDevice: add default constructor for Task and clean up

Signed-off-by: Haomai Wang <haomai@xsky.com>

NVMEDevice: only activate zero command when backend supported

Signed-off-by: Haomai Wang <haomai@xsky.com>

BlockDevice: detect symbol file basename

Signed-off-by: Haomai Wang <haomai@xsky.com>

BlueStore: fix spdk link file create failed

Signed-off-by: Haomai Wang <haomai@xsky.com>

NVMEDevice: fix atomic and lock changes

Signed-off-by: Haomai Wang <haomai@xsky.com>

NVMEDevice: use nvme_probe to simply init

Signed-off-by: Haomai Wang <haomai@xsky.com>

spdk: update spdk submodule

Signed-off-by: Haomai Wang <haomai@xsky.com>

NVMEDevice: use nvme zero command instead of writing zero

Signed-off-by: Haomai Wang <haomai@xsky.com>

NVMEDevice: cleanup task iterator process

Signed-off-by: Haomai Wang <haomai@xsky.com>

vstart.sh: avoid race condition starting rgw via vstart.sh

In slow environments running vstart.sh hits a race condition setting up
the rgw server. Creating the S3 and Swift users before booting rgw
avoids the issue.

Fixes: #14829
Signed-off-by: Javier M. Mellid <jmunhoz@igalia.com>

test/radosgw-admin: update the expected usage outputs

to match with the latest radosgw-admin changes.

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

test: rbd-mirror: add tool to test ImageReplayer

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

test: rbd-mirror: add ImageReplayer tests

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

rbd-mirror: implement ImageReplayer

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

journal: wrappers to get journal metadata

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

librbd: allow to use journal replay externally

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

cls::journal: don't print client data

It makes `rbd journal status` output look ugly.

It is a temporary solution and is going to be improved in
the future (it could be useful to see data but decoded).

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

Merge pull request #7627 from dillaman/wip-librbd-read-replay

librbd: block read requests until journal replayed

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

Merge pull request #7718 from dillaman/wip-librbd-mirror-peer

librbd: update of mirror pool mode and mirror peer handling

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

Merge pull request #7698 from dillaman/wip-librbd-replay-flush

librbd: differentiate journal replay flush vs shut down

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

Merge pull request #7720 from ceph/wip-rgw-aws4

rgw support for aws authentication v4 (Javier M. Mellid)

Merge remote-tracking branch 'jmunhoz/wip-aws4' into wip-aws4

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>
Conflicts:
src/rgw/rgw_auth_s3.h
src/rgw/rgw_client_io.cc
src/rgw/rgw_client_io.h
src/rgw/rgw_common.h
src/rgw/rgw_main.cc
src/rgw/rgw_op.cc
src/rgw/rgw_op.h
src/rgw/rgw_rest.cc
src/rgw/rgw_rest_metadata.cc
src/rgw/rgw_rest_s3.cc

Merge pull request #7719 from ceph/wip-rgw-keystone-v3

rgw keystone v3 (Mark Barnes, Radoslaw Zarzynski)

support for keystone v3

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

librbd: update mirror peer data structure

Replace cluster uuid with mirror peer uuid and support an
optional pool id within a peer (when using internal APIs
for testing).

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

librbd: support image and pool-level mirroring modes

If pool-level mirroring is enabled, all journaled images within
the pool are eligible for mirroring.  If image-level mirroring
is enabled, mirroring needs to be explicitly enabled on each
image.

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

rgw: Keystone token parsing doesn't need to know API version.

This patch targets an issue with S3Extension in Keystone:
requested tokens always conform to Keystone API v2 - regardless
of the version used to make the request.

Previous implementation of KeystoneToken::parse() as well
as KeystoneToken::decode_json() had to know the API version
explicitly. Thus, they might be affected in the future by
changes in S3-compatibility middleware of Keystone.

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

rgw: S3 always must use token format of Keystone v2.

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

rgw: Keystone token parsing should fail on misformed JSONs.

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

rgw: improve debugs in RGWPostHTTPData.

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

rgw: ONLY formatting improvements for Keystone-related code.

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

rgw: handle Keystone API version with dedicated enum.

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

rgw: fix wrong format of admin token request in Keystone v3.

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

rgw: rework serialization in token retrieval for Keystone's admin.

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

rgw: fix Keystone v2 regression in role's JSON parsing

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

rgw: accommodate Keystone v3 in S3-related auth mechanisms.

This patch is based on changes proposed by Mark Barnes
<mark.barnes@ocado.com> (lperiquito) in commit
af71f6a518529ea6cccb25bd46da2b6d1458c1da.

Signed-off-by: Mark Barnes <mark.barnes@ocado.com>
Signed-off-by: Radoslaw Zarzynski <rzarzynski@mirantis.com>

rgw: accommodate Keystone v3 in RGWSwift class.

This patch is based on changes proposed by Mark Barnes
<mark.barnes@ocado.com> (lperiquito) in commit
af71f6a518529ea6cccb25bd46da2b6d1458c1da.

Signed-off-by: Mark Barnes <mark.barnes@ocado.com>
Signed-off-by: Radoslaw Zarzynski <rzarzynski@mirantis.com>

rgw: basic data structures for Keystone v3.

The change has been extracted from changeset proposed
in commit af71f6a518529ea6cccb25bd46da2b6d1458c1da by
Mark Barnes <mark.barnes@ocado.com> (lperiquito).

Signed-off-by: Mark Barnes <mark.barnes@ocado.com>
Signed-off-by: Radoslaw Zarzynski <rzarzynski@mirantis.com>

rgw: unify RGWValidateKeystoneToken with RGWPostHTTPData.

The change has been extracted from changeset proposed
in commit af71f6a518529ea6cccb25bd46da2b6d1458c1da by
Mark Barnes <mark.barnes@ocado.com> (lperiquito).

Signed-off-by: Mark Barnes <mark.barnes@ocado.com>
Signed-off-by: Radoslaw Zarzynski <rzarzynski@mirantis.com>

rgw: make RGWPostHTTPData able to extract X-Subject-Token.

The change has been extracted from changeset proposed
in commit af71f6a518529ea6cccb25bd46da2b6d1458c1da by
Mark Barnes <mark.barnes@ocado.com> (lperiquito).

Signed-off-by: Mark Barnes <mark.barnes@ocado.com>
Signed-off-by: Radoslaw Zarzynski <rzarzynski@mirantis.com>

PendingReleaseNotes: document the dropped buffer symbols

the symbols of buffer::list::iterator_impl<> were wrongly exposed
in previous infernalis release, and the clients linked against
librados are very likely using them. so we need to document this
change.

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

buffer: hide iterator_impl symbols

buffer::list::iterator_impl symbols are referenced by const_iterator
and iterator, and are exposed as weak symbols. if a source file is
compiled using the buffer.h, the produced object file will reference
these symbols as well, so we'd better hiding them and avoid using
them in the header file.
as a side-effect, buffer::list::const_iterator is also hidden, but
currently we don't have any librados client using this class, so
we can just leave it as an internal class at this moment.

Fixes: #14788
Signed-off-by: Kefu Chai <kchai@redhat.com>

librbd: read ops require journal replay

When journaling is enabled, a read op should acquire the exclusive
lock and replay the journal in case uncommitted writes are stored
in the journal.

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

Merge pull request #7717 from cbodley/wip-cmake-rbd-image-watcher

cmake: add missing librbd image_watcher sources

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

cmake: add missing librbd image_watcher sources

Fixes: #14823
Signed-off-by: Casey Bodley <cbodley@redhat.com>

Merge pull request #7709 from ceph/wip-rgw-new-multisite-merge

rgw multisite v2 (Yehuda Sadeh, Orit Wasserman, Casey Bodley)
 - rename region to zonegroup
 - new zone and zonegroup configuration
 - new realm and period structures
 - rgw handles synchronization (no need for sync agent)
 - active-active data

Signed-off-by: Yehuda Sadeh

doc: standardize @param (not @parma, @parmam, @params)

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

Merge pull request #7371 from theanalyst/mon/quota-msg-cleanup

mon: cleanup set-quota error msg

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

Merge pull request #7708 from jdurgin/wip-cython-cleanup

pybind: replace __del__ with __dealloc__ for rbd

Reviewed-by: Hector Martin <marcan@marcan.st>
Reviewed-by: Mehdi Abaakouk <sileht@sileht.net>

mon: cleanup set-quota error msg

When we send in an unrecognized field for set-quota, specify that the
options should be `max_bytes` or `max_objects`. This doesn't affect the
ceph-cli as the choices are caught at an earlier stage, however the
error message is seen while reaching the cluster via rados api for eg.

Signed-off-by: Abhishek Lekshmanan <abhishek@suse.com>

Merge pull request #7668 from dillaman/wip-librbd-replay-locks

librbd: use async librados notifications

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

rgw: avoid empty object names

Don't try to sync an object with empty name (shouldn't happen, but
bugs may happen). Also, don't try to write_meta() of an empty object
name, in case we get there through a different path somehow.

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

.gitignore: include recently added rgw programs

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

rgw: fix s3 list bucket (affects format=json)

another merge casualty. The section wasn't opened properly.

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

pybind: replace __del__ with __dealloc__ for rbd

Cython extension types like these don't call __del__, but use
__dealloc__ instead:

http://docs.cython.org/src/userguide/special_methods.html#finalization-method-dealloc

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

Merge pull request #7706 from sileht/sileht/rados-cython

pybind: remove next() on iterators

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

rgw: only validate bucket name if not empty

This got broken due to code movement

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

rgw: resurrect lost code

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

pybind: remove next() on iterator

Cython recommends to implement only __next__() methods.

http://docs.cython.org/src/userguide/special_methods.html#the-next-method

Signed-off-by: Mehdi Abaakouk <sileht@redhat.com>

Merge pull request #7641 from batrick/vstart-usage

Clarify usage on starting single osd/mds/mon.

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

test/Makefile-client.am: adjustment following merge

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

librbd: helper method for flushing journal commit positions

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

journal: added interface to flush pending commit position updates

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

Merge pull request #7692 from sileht/sileht/rados-cython

pybind/rados: use __dealloc__ since __del__ is ignored by cython

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

Merge remote-tracking branch 'origin/master' into wip-rgw-new-multisite

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>
Conflicts:
src/CMakeLists.txt
src/rgw/Makefile.am
src/rgw/rgw_admin.cc
src/rgw/rgw_common.h
src/rgw/rgw_main.cc
src/rgw/rgw_op.cc
src/rgw/rgw_rados.h
src/rgw/rgw_rest_s3.cc
src/test/Makefile-client.am

librbd: differentiate journal replay flush vs shut down

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

rgw: add drain_all() before exiting a cr

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

rgw: silence compilation #warnings

these are now FIXMEs

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

rgw: more error reporting

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

rgw: data sync error handling (for sync init)

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

rgw: a bit better hashing

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

librbd: don't hold owner lock while refreshing image synchronously

The write lock will be taken when the new state is applied.

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

librbd: need owner write lock when disabling exclusive lock

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

librbd: correct memory leaks discovered via valgrind

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

librbd: notifications should be flushed between exclusive lock states

Avoid leaving in-flight notification messages when transitioning lock
states.

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

rgw: RGWMetaSyncCR loops through period history

RGWMetaSyncCR uses a period history Cursor to track its position. it
uses this to get the max sync markers for each shard from the following
period, so that RGWMetaSyncShardCR knows when to stop syncing and return
control to RGWMetaSyncCR

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

rgw: get period history cursor when starting sync

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

rgw: remove RGWMetadataManager::store_md_log_entries

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

rgw: rados coroutines take bucket by const ref

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

rgw: add first log period to mdlog info response

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

rgw: factor get_log_shard_id() out of RGWMetadataLog

by using RGWRados::key_to_shard_id(), we can avoid the dependency on
RGWMetadataLog's prefix (and period)

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

rgw: add period id to rgw_rest_log operations

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

rgw-admin: update callers of get_log

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

rgw: add map of period_id -> RGWMetadataLog

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