qa: run install task only once

Fixes: https://tracker.ceph.com/issues/45612
Fixes: 8d77d134c73d3471515f113320e9db1f0e59eb1f
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>

Merge pull request #35113 from athanatos/sjust/wip-crimson-clang-cleanup

crimson clang cleanups

Reviewed-by: Ronen Friedman <rfriedma@redhat.com>
Reviewed-by: Radoslaw Zarzynski <rzarzyns@redhat.com>
Reviewed-by: Kefu Chai <kchai@redhat.com>

crimson/osd: remove unnecessary captures

Signed-off-by: Samuel Just <sjust@redhat.com>

crimson/os: remove some unnecessary lambda captures

Signed-off-by: Samuel Just <sjust@redhat.com>

crimson/os/seastore/journal: avoid capturing destructured labels

clang at least doesn't allow those labels to be captured.

Signed-off-by: Samuel Just <sjust@redhat.com>

crimson: remove a few unnecessary lambda captures

Signed-off-by: Samuel Just <sjust@redhat.com>

errorator: don't rely on non-evaluation of static_assert in constexpr branch

At least in clang, we appear to fail that static assert even with another
branch taken.  Instead, assert that one of the other conditions is true.

Signed-off-by: Samuel Just <sjust@redhat.com>

Merge pull request #35127 from ronen-fr/scrub4

test/librbd, Crimson: remove some unused captures

Reviewed-by: Samuel Just <sjust@redhat.com>
Reviewed-by: Radoslaw Zarzynski <rzarzyns@redhat.com>
Reviewed-by: Jason Dillaman <dillaman@redhat.com>

Merge pull request #32927 from linuxbox2/wip-rgw-lc-threads

rgwlc: enable thread-parallelism in RGWLC

Crimson: cleanup Clang issues in Crimson code

Mostly: remove redundant captures

Signed-off-by: Ronen Friedman <rfriedma@redhat.com>

test/librbd: remove some unused captures

...silencing Clang warnings

Signed-off-by: Ronen Friedman <rfriedma@redhat.com>

Merge pull request #34980 from sebastian-philipp/cephadm-workunits-rm-ubuntu_2020

qa/cephadm: rm ubuntu_latest for now

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

Merge pull request #35101 from sebastian-philipp/cephadm-ls-systemd-unit

cephadm: `ls`: also return systemd_unit

Reviewed-by: Michael Fritch <mfritch@suse.com>

Merge pull request #35074 from dillaman/wip-45544

librbd: avoid completing mirror:DisableRequest while holding its lock

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

Merge PR #35062 into master

* refs/pull/35062/head:
qa/tasks/cephfs/fuse_mount.py: retry when the admin socket is not ready

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

Merge PR #34571 into master

* refs/pull/34571/head:
mds: don't shallow copy when decoding xattr map

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

Merge PR #34144 into master

* refs/pull/34144/head:
Filer: Add error handler in _do_purge_range()
mds: Handle blacklisted error in purge queue

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

Merge PR #34596 into master

* refs/pull/34596/head:
test: add a new program for testing ino_release_cb
client: add a new inode release request callback
client: expose ceph_ll_register_callbacks via libcephfs
client: move callback typedefs and arg struct into ceph_ll_client.h
client: rename ceph_statx.h to ceph_ll_client.h
client: make client_dentry_callback_t more friendly for C
client: only override umask_cb with non-NULL values

Reviewed-by: Xiubo Li <xiubli@redhat.com>
Reviewed-by: Patrick Donnelly <pdonnell@redhat.com>
Reviewed-by: Zheng Yan <zyan@redhat.com>
Reviewed-by: Greg Farnum <gfarnum@redhat.com>

Merge PR #34890 into master

* refs/pull/34890/head:
cephfs-shell: Remove get_error_code() call on OSError type exception objects

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

Merge PR #34951 into master

* refs/pull/34951/head:
qa/cephfs: run() cleanup whether FS was mounted or not

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

Merge PR #35058 into master

* refs/pull/35058/head:
qa/tasks: logger.warn() -> logger.warning()

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

Merge pull request #35108 from trociny/wip-45589

mgr/rbd_support: rename "rbd_trash_trash_purge_schedule" oid

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

Merge pull request #34900 from gekios/wip-cephadm-bootstrapped

qa/tasks/cephadm.py: Support already bootstrapped clusters

Reviewed-by: Sebastian Wagner <sebastian.wagner@suse.com>

Merge pull request #35071 from sebastian-philipp/cephadm-upgrade-check-mon-command

mgr/cephadm: upgrade: use check_mon_command

Merge pull request #35068 from bk201/wip-45163

mgr/cephadm: config iSCSI gateways in Dashboard

Merge pull request #35084 from jschmid1/caching_osdspec_previews

mgr/cephadm: implement cached osdspec previews

Reviewed-by: Sebastian Wagner <sebastian.wagner@suse.com>

mgr/rbd_support: rename "rbd_trash_trash_purge_schedule" oid

to "rbd_trash_purge_schedule"

Fixes: https://tracker.ceph.com/issues/45589
Signed-off-by: Mykola Golub <mgolub@suse.com>

Merge branch 'wip-joshd-fix'

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

mgr/orchestrator: adapting orch interface to preview changes

Signed-off-by: Joshua Schmid <jschmid@suse.de>

Merge pull request #35015 from smithfarm/wip-45514

doc: drop obsolete cache tier options

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

Merge pull request #35083 from p-se/wip-pse-cephadm-prom-scrape-interval

mgr/cephadm: increase global scrape interval

Reviewed-by: Sebastian Wagner <sebastian.wagner@suse.com>

Merge pull request #35092 from varshar16/wip-cephadm-mdcache-ganesha

 mgr/cephadm/nfs: Instead of CACHEINODE use MDCACHE

Reviewed-by: Jeff Layton <jlayton@redhat.com>
Reviewed-by: Michael Fritch <mfritch@suse.com>
Reviewed-by: Patrick Donnelly <pdonnell@redhat.com>
Reviewed-by: Sebastian Wagner <sebastian.wagner@suse.com>

librbd: avoid completing mirror:DisableRequest while holding its lock

Ensure that the lock is released before another thread has the chance
to complete the state machine and attempt to destruct the in-use lock.

Fixes: https://tracker.ceph.com/issues/45544
Signed-off-by: Jason Dillaman <dillaman@redhat.com>

mgr/cephadm: upgrade: use check_mon_command

Signed-off-by: Sebastian Wagner <sebastian.wagner@suse.com>

Merge pull request #34964 from dillaman/wip-45137

librbd: replace ImageRequestWQ with pluggable ImageDispatcher

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

doc/cephadm: Synopsis: add `cephadm unit`

Signed-off-by: Sebastian Wagner <sebastian.wagner@suse.com>

cephadm: `ls`: also return systemd_unit

Let users directly know the systemd unit name.

Enable things like

```
cephadm ls | jq -r '.[].systemd_unit' | xargs -n 1 systemctl status | cat
```

Signed-off-by: Sebastian Wagner <sebastian.wagner@suse.com>

mgr/cephadm: increase global scrape interval

Signed-off-by: Patrick Seidensal <pseidensal@suse.com>

mgr/dashboard: support setting iSCSI gateways from orchestrator

- Dashboard doesn't actively query iSCSI gateways from orchestrator anymore.
  Instead, Orchestrator backend invokes iscsi-gateway-add CLI after any
  iSCSI daemons are deployed.
- Add an optional argument `name` to `iscsi-gateway-add` command. If
  `name` is provided, Dashboard won't issue API request to iSCSI gateway
  to get the hostname. This is required because when orchestrator sets
  the gateway via CLI, the gateway is just spawned and might be not
  ready yet.

Fixes: https://tracker.ceph.com/issues/45163
Signed-off-by: Kiefer Chang <kiefer.chang@suse.com>

mgr/cephadm: config iSCSI gateways in Dashboard

- Add `daemon_check_post` fuction to allow services implement actions
  needed to be done after daemons are checked.
- After iSCSI daemons are spawned, call dashboard CLI to
  set the API URL of the daemons.

Fixes: https://tracker.ceph.com/issues/45163
Signed-off-by: Kiefer Chang <kiefer.chang@suse.com>

Merge pull request #32682 from lorenzen-b1/patch-1

doc/architecture.rst: clarify scrubbing

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

Merge pull request #35091 from tchaikov/wip-crimson-cleanup

crimson/osd: avoid using variadic future

Reviewed-by: Samuel Just <sjust@redhat.com>

Merge PR #35081 into master

* refs/pull/35081/head:
qa: do not save/restore ip_forward sysctl

Reviewed-by: Xiubo Li <xiubli@redhat.com>

Merge pull request #34764 from ianwatsonrh/master

Docs: vault kv and transit engine updates

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

 mgr/cephadm/nfs: Instead of CACHEINODE use MDCACHE

As CACHEINODE will be deprecated soon.

Signed-off-by: Varsha Rao <varao@redhat.com>

crimson/osd: avoid using variadic future

it is deprecated by seastar. let's use future<tuple<...>> instead

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

Merge pull request #35025 from dillaman/wip-45518

librbd: copy API should not inherit v1 image format by default

Merge pull request #34789 from trociny/wip-rbd-nbd-quiesce

rbd-nbd: add quiesce/unquiesce hooks

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

Merge pull request #35023 from yuvalif/fix_issue_45513

cls/2pc_queue: fix timing issue in test

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

Merge pull request #34959 from yison/rwl-image-writeback-cache-seq10

rbd/cache: Replicated Write Log core codes - load existing cache

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

Merge pull request #34930 from trociny/wip-45409

rbd-mirror: wait for events to replay before shut down journal replay

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

Merge pull request #35042 from tchaikov/wip-cmake-remove-iwyu-cppcheck

cmake: remove cppcheck and iwyu targets

Reviewed-by: Yuval Lifshitz <ylifshit@redhat.com>
Reviewed-by: Brad Hubbard <bhubbard@redhat.com>

Merge pull request #35070 from Devp00l/wip-45545-2

mgr/dashboard: Updates jest to fix cache failure

Reviewed-by: Tiago Melo <tmelo@suse.com>

Merge pull request #34913 from athanatos/sjust/wip-seastore-journal

crimson/os/seastore: initial PR

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

mgr/cephadm: move iSCSI service class to its own file

Signed-off-by: Kiefer Chang <kiefer.chang@suse.com>

mgr/cephadm: adapt cephadm to osdspec preview caching

Signed-off-by: Joshua Schmid <jschmid@suse.de>

mgr/cephadm: add preview caching to inventory.py

Signed-off-by: Joshua Schmid <jschmid@suse.de>

Merge pull request #35052 from tspmelo/wip-unittest-warnings

Reviewed-by: Alfonso Martínez <almartin@redhat.com>
Reviewed-by: Kiefer Chang <kiefer.chang@suse.com>

Merge pull request #35013 from matthewoliver/cephadm-iscsi-log-to-stderr

cephadm: enable logging to stderr in ceph-iscsi

Reviewed-by: Michael Fritch <mfritch@suse.com>

dec/dev/seastore.rst: add more detailed design information and status

Signed-off-by: Samuel Just <sjust@redhat.com>

crimson/os/seastore: add initial cache support with tests

Signed-off-by: Samuel Just <sjust@redhat.com>

crimson/os/seastore: introduce initial journal implementation and tests

Signed-off-by: Samuel Just <sjust@redhat.com>

crimson/gtest_seastar: add helpers for building seastar gtest unit tests

Also adds unsafe_get methods for errorator for use within unit tests.

Signed-off-by: Samuel Just <sjust@redhat.com>

qa: do not save/restore ip_forward sysctl

We don't really care if this is left on during the course of the mount.
These settings don't persist across tests anyway.

Fixes: https://tracker.ceph.com/issues/45525
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>

qa/tasks/cephfs/fuse_mount.py: retry when the admin socket is not ready

Fixes: https://tracker.ceph.com/issues/45552
Signed-off-by: Xiubo Li <xiubli@redhat.com>

Merge pull request #35077 from yuriw/wip-yuriw-crontab-master

qa/tests: reduced powercycle jobs amount

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

qa/tests: reduced powercycle jobs amount

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

Merge pull request #32829 from Devp00l/wip-43765

Reviewed-by: Alfonso Martínez <almartin@redhat.com>
Reviewed-by: Tatjana Dehler <tdehler@suse.com>
Reviewed-by: Volker Theile <vtheile@suse.com>
Signed-off-by: Yuri Weinstein <yweinste@redhat.com>

Merge pull request #35076 from yuriw/wip-yuriw-distro-master

qa/tests: run `nautilus-x` on `pacific` on `ubuntu 18.04`

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

qa/tests: run `nautilus-x` on `pacific` on `ubuntu 18.04`

...as we don't build `nauiltus` on `20.04`

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

librbd: QoS delayed dispatch needs to use op work queue

The QoS throttle will invoke its throttle complete callback within
the scope of the ImageCtx timer lock singleton. This would cause
the potential for deadlock if a future dispatch layer attempted
to acquire the same timer lock.

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

librbd: corrected naming of singleton timer locker

It incorrectly included "Journal" in the name.

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

librbd: correct misnamed parameters in io::ImageDispatcher

It was incorrectly using to object dispatch names instead of
image.

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

librbd: api::Io now properly passing the parent trace information

A copy-and-paste error from moving the original implentation from
ImageRequestWQ resulted in the parent trace being incorrectly
uninitialized.

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

Merge pull request #32829 from Devp00l/wip-43765

Reviewed-by: Alfonso Martínez <almartin@redhat.com>
Reviewed-by: Tatjana Dehler <tdehler@suse.com>
Reviewed-by: Volker Theile <vtheile@suse.com>

librbd: ensure in-flight read-ahead + copy-on-read ops are finalized

Now that the image dispatcher isn't flushing AsyncOperations with each
flush request, we need to ensure that we flush all background ops prior
to fully closing the image.

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

librbd: remove io::ImageRequestWQ

This has been fully deprecated in favor of the new plugable image
dispatcher system.

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

librbd: switch to new api::Io dispatch helper methods

This is another step in the removal of ImageRequestWQ in favor
of the new ImageDispatcher.

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

librbd: created new api::Io helper methods

These will be invoked by the user API and wrap various user API
specific logic.

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

librbd: exclusive-lock dispatch flush pseudo-handling

Ensure that the exclusive-lock layer tracks all in-flight IOs so that
it can flush the IOs from lower layers when the lock becomes required.

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

librbd: qos dispatch flush pseudo-handling

Flush requests will be properly queued behind write operations to
ensure data consistency.

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

librbd: generic helper for tracking in-flight IO and managing flush requests

Layers that potentially queue IOs but not all IOs will need to track all
in-flight IOs to properly ensure that a flush cannot complete while
older IO is still stuck in a queue.

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

librbd: temporarily redirect ImageRequestWQ IO through ImageDispatcher

The next series of commits will work to remove ImageRequestWQ from librbd.

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

librbd: complete ImageDispatchSpec upon AioCompletion finalization

This will allow ImageDispatchSpec to invoke any necessary finalizers
for the different dispatch layers and allows in-flight IO to be tracked
since the image dispatch layer finalizers won't be invoked until after
the AioCompletion fires.

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

librbd: switched IO (un)block_write calls to new queue image dispatcher

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

librbd: queue image IO dispatch layer

This layer replaces the queue implemented by ImageRequestWQ and handles
blocking all write requests through the dispatch pipeline.

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

librbd: exclusive lock image IO dispatch layer

This layer will handle acquiring the exclusive lock if required upon
incoming IO requests.

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

librbd: refresh image IO dispatch layer

This layer will handle refreshing the image if required upon incoming
IO requests.

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

librbd: QoS image io dispatch layer

Re-implement the existing QoS throttling behavior from ImageRequestWQ
as a new image IO dispatch layer. The existing QoS hooks in
ImageRequestWQ are temporarily hooked to the new layer.

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

librbd: support a finish handler for image IO dispatch layers

This will be faster than allocating per-IO Context callbacks for
layers that need to track the in-flight state of particular IOs.

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

librbd: init and shut down image IO dispatch layer

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

librbd: initial implementation for image IO dispatcher

This is similar to the object IO dispatcher but it expects image extents
for IO operations.

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

librbd: add image dispatch layer to image dispatch spec

This will allow image extent-based IOs to skip layers. Also added helpful
enums for the start of user-based API requests and internal requests.

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

librbd: added io::ImageDispatchSpec::C_Dispatcher

This will assist with dispatching image extent-based IO requests
through the image IO dispatcher system (to be added in a later commit).

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

librbd: move generic IO dispatcher logic to base class

This will allow re-use between the existing ObjectDispatcher and
the ImageDispatcher that will be added in a future commit.

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

common: TokenBucketThrottle should use perfect forwarding for item

This will allow more than just a pointer to an item to be passed to
the callback.

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

mgr/dashboard: Updates jest to fix cache failure

Fixes: https://tracker.ceph.com/issues/45545
Signed-off-by: Stephan Müller <smueller@suse.com>

rgwlc: wait for WorkQ threads on shutdown

Fixes: https://tracker.ceph.com/issues/43841
Signed-off-by: Mark Kogan <mkogan@redhat.com>

Merge pull request #31889 from ofriedma/wip-ofriedma-reshard-debug

rgw: Fix reshard log so we can know why the reshard is returning error

qa/tasks/cephadm: Add support for raw disks

Currently tests on raw disks are failing due to stripping '/dev/' from
device name. (ie https://tracker.ceph.com/issues/45383#change-165721)

Signed-off-by: Georgios Kyratsas <gkyratsas@suse.com>