ceph-daemon: fchmod before writing to keyring file

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

test_ceph_daemon.sh: skip ssh until container image has remoto

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

ceph-daemon: decode utf-8 in run() helper

Less repetition

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

mgr/ssh: clean up debug cruft

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

mgr/ssh: clean up bare except: block

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

ceph-daemon: clean up bare except: blocks

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

ceph-daemon: all imports to top

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

ceph-volume: no_tmpfs -> tmpfs

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

doc/bootstrap: add new bootstrap documentation

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

ceph-daemon: add --output-pub-ssh-key for bootstrap

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

ceph-daemon: make 'shell' easier to use

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

ceph-daemon: support docker; prefer podman

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

qa: add ceph-daemon

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

debian: ceph-daemon package, required by ceph-mgr-ssh

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

ceph.spec.in: ceph-daemon package, required by ceph-mgr

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

common/options: cleanup whitespace

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

mgr/ssh: simplify getting the cluster fsid

This never changes, so just stash it once.

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

mgr/ssh: pipe ceph-daemon script to stdin of python3

This avoids any need for the script to be present on the remote host.

We introduce a config option to indicate where the script should be
read from, since the location varies between a vstart environment (source
dir) and a real install (/usr/sbin).

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

ceph-daemon: add support for args and/or stdin from top of script

Allow someone to run this script by prepending injected_{args,stdin} to
the top and then piping it all to a python3 binary.

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

ceph-daemon: make ceph-volume use get_config_and_keyring

This will always take the first config+keyring json branch.

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

ceph-daemon: ls: behave if /var/log/ceph doesn't exist

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

ceph-daemon: implement 'adopt' for legacy style daemons

Caveats:

 - this doesn't clean out /etc/ceph/*, since we don't know which is the
   last daemon to go, and/or whether the user wants to keep it around
   for using the ceph CLI on this host
 - leaves behind /var/lib/ceph/bootstrap-* keys, even after all daemons
   have been converted.

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

ceph-daemon: fix fsid detection for legacy osds

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

ceph-daemon: make rm-cluster clean up system-ceph*.slice too

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

ceph-daemon: configure ssh orchestrator

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

ceph-daemon: be more restrictive with file permissions

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

mgr/ssh: create osd with ceph-daemon

Three basic steps:

1- ceph-volume lvm prepare
2- ceph-volume lvm list
3- for each osd, ceph-daemon deploy (which calls c-v activate inside the
   new container)

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

mgr/ssh: pass daemon id separately to _create_daemon

Don't assume it is the hostname (with osds, it's not!).

Also, just pass arbitrary args down, instead of special-casing the
network option.

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

ceph-daemon: add --config-and-keyring to ceph-volume command

...to provide a ceph.conf and the bootstrap-osd key.

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

ceph-daemon: create log path for shell (if needed)

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

mgr/ssh: use _run_ceph_daemon for _create_daemon

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

mgr/ssh: factor _run_ceph_daemon out of _get_device_inventory

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

mon/ConfigMonitor: allow entity type only for 'config get'

e.g., 'ceph config get osd debug_osd' to return the config value that
would apply to a generic OSD (either from the osd or global section of
the config).

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

ceph-daemon: add ceph-volume subcommand

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

ceph-daemon: remove unused CephContainer dname property

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

ceph-daemon: drop useless uid/gid checks

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

mgr/ssh: deploy new mgrs with ceph-daemon

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

mgr/ssh: factor _create_daemon out of create_mon

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

mon/MonCap: allow mgr to create new auth keys

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

mgr/ssh: run c-v with podman when getting inventory

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

mgr/ssh: simplify ssh connection management

- Use a single instance of the config and identity files for the whole
module.  There's no need to create these for *every* connection--it just
pollutes /tmp.
- Drop the SSHConnection wrapper, since the temp files are tied to the
daemon lifecycle now.
- Prefix the tmp files so I can tell wtf is going on.
- Always connect to root@host, to avoid remoto's localhost detection
feature.  This ensures we have a consistent connection model and user.
(The daemon might be running as user ceph and try to connect to localhost,
but end up running the command as the wrong user and/or inside the
ceph-mgr container.)

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

mgr/ssh: use ceph-daemon for deploying mon

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

ceph-daemon: allow --mon-network for deploying new mon (vs specifying IP)

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

ceph-daemon: --config-and-keyring (not key)

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

common/options: add 'image' config option

Defaults to 'ceph/daemon-base'.

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

test_ceph_daemon: specify image name

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

vstart.sh: add --ssh to enable+configure ssh orchestrator

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

mgr/ssh: use ssh identity from config-key, if present

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

mgr/ssh: hardcode default ssh_config

Use this unless either ssh_config or ssh_config_file are present.

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

ceph-daemon: store ssh identity in mon config-key store

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

ceph-daemon: --privileged arg for 'exec'

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

ceph-daemon: make deploy work for osd (do a c-v prepare)

This is sufficient to deploy an OSD that is based on ceph-volume lvm.
YMMV if it's not an lvm-based OSD.

Run the OSD container privileged so we can open the raw block device.

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

ceph-daemon: make shell privileged

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

ceph-daemon: move get_container_mounts to a helper

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

ceph-daemon: pass full path for entrypoint

Sometimes we run things in /usr/sbin...

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

ceph-daemon: make id portion of 'shell' optional

This lets you start up a 'generic' container of a particular class,
without a data mount, but with the appropriate other mounts and privilege
levels.

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

ceph-volume: accept --no-tmpfs argument for bluestore

When activating a bluestore inside a container, we want to (be able to)
make the osd dir metadata persistent inside the container.

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

ceph-daemon: 'unit' command

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

ceph-daemon: fix run command to use call(), not check_output()

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

src/ceph-daemon: whitespace

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

ceph-daemon: add 'enter', 'exec' commands

'enter' is really just 'exec' + bash

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

ceph-daemon: bind config to default location

This is one less thing on the command line, and an easier time when inside
the container.

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

test_ceph_daemon.sh: test deploy mds too

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

ceph-daemon: generate ssh keys

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

ceph-daemon: --config, not --conf

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

ceph-daemon: long lines

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

ceph-daemon: add --config to bootstrap

Settle on --config|-c too, not --conf.

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

ceph-daemon: add 'shell' command

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

ceph-daemon: do not import subprocess symbols directly

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

ceph-daemon: add mons with 'deploy mon.x ...'

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

ceph-daemon: add 'ls'

Eventually we should teach this to understand rook mons+osds and
ceph-ansible containerized daemons.

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

ceph-daemon: simplify uid/gid a bit

Assume user ceph and group ceph exist inside the container.  We still need
the uid/gid for creating daemon directories, though, and for the bootstrap,
where we are creating files on the host and passing them through.

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

ceph-daemon: fix libudev

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

ceph-daemon: autodetect uid/gid from container image

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

ceph-daemon: default to empty log files, log to stderr (systemd journal)

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

ceph-daemon: rm-{daemon,cluster}

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

ceph-daemon: fix bootstrap config

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

ceph-daemon: fix args.fsid usage

Pass it explicitly, since it only sometimes comes from args.

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

ceph-daemon: be careful overwriting live files

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

ceph-daemon: slurp some options over from the standard systemd unit

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

ceph-daemon: add ceph.target and ceph-$fsid.target units

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

test_ceph_daemon.sh: stupid test script

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

ceph-daemon: bootstrap and deploy (mgr) work

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

ceph-daemon: initial checkin

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

ceph-mon: fix debug print of public_addr

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

Merge PR #30542 into master

* refs/pull/30542/head:
doc: cleanup CephFS Landing Page

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

mgr/dashboard: Show smartctl summary in OSD tab (#30378)

mgr/dashboard: Show smartctl summary in OSD tab

Reviewed-by: Laura Paduano <lpaduano@suse.com>
Reviewed-by: Volker Theile <vtheile@suse.com>

Merge pull request #29922 from smithfarm/wip-41524

cmake: set empty RPATH for some test executables

Reviewed-by: Kefu Chai <kchai@redhat.com>
Reviewed-by: Ricardo Dias <rdias@suse.com>

Merge PR #30416 into master

* refs/pull/30416/head:
client: add procession of SEEK_HOLE and SEEK_DATA in lseek.

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

Merge pull request #30653 from tchaikov/wip-os/drop-kinetic

os: remove KineticStore

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

Merge pull request #28569 from alimaredia/wip-rgw-ioctx-operate-replace

rgw: replace direct calls to ioctx.operate()

Merge pull request #30655 from tchaikov/wip-cmake/crimson/pass-args

cmake: pass arguments to crimson tests

Reviewed-by: Radoslaw Zarzynski <rzarzyns@redhat.com>

client: add procession of SEEK_HOLE and SEEK_DATA in lseek.

Fixes: https://tracker.ceph.com/issues/42107
Signed-off-by: Shen Hang <harryshen18@gmail.com>

Merge pull request #30613 from joke-lee/rgw-ceph-dencoder-add-RGWPeriodLatestEpochInfo

ceph-dencoder: add RGWPeriodLatestEpochInfo support.

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

cmake: pass arguments to crimson tests

otherwise they will take all CPU cores and all memory

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

build/ops: set empty RPATH for some test executables

Fixes: https://tracker.ceph.com/issues/41524
Signed-off-by: Nathan Cutler <ncutler@suse.com>

Merge pull request #30587 from sebastian-philipp/python-common-readme

python-common: Add small Readme

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

Merge PR #30341 into master

* refs/pull/30341/head:
cephfs-shell: better complain info, when deleting non-empty directory

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

Merge PR #30559 into master

* refs/pull/30559/head:
mds: Reorganize class members in MDBalancer header

Reviewed-by: Jos Collin <jcollin@redhat.com>

Merge PR #30580 into master

* refs/pull/30580/head:
client: don't ceph_abort on bad llseek whence value
client: remove Inode dir_contacts field

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