RTD does not support installing system packages, the only ways to install
dependencies are setuptools and pip. while ditaa is a tool written in
Java. so we need to find a native python tool allowing us to render ditaa
images. plantweb is able to the web service for rendering the ditaa
diagram. so let's use it as a fallback if "ditaa" is not around.
also start a new line after the directive, otherwise planweb server will
return 500 at seeing the diagram.
doc/conf.py: exclude pybindings docs from build for RTD
because it'd difficult to prepare (dummy) librados,libcephfs and librbd for
their python bindings in the building environment offered by Read the Docs.
Introduced in 4d42b4c5a0ed ("common/TextTable: default to 2 spaces
separating columns") and 41f003518a07 ("common/TextTable: only pad
between columns").
qa/tests: pass "-t 2" to teuthology-suite when testing non-master branches
-t is an alias of "--teuthology-branch" which defaults to master. we
have a "py2" branch in teuthology. that branch is compatible with
python2, and the teuthology-worker is executed in a Python2 environment,
so all qa suites tested using "py2" branch of teuthology should work
fine as long as they are Python2 compatible. currently, all LTS branches
are compatible with Python2, and only master is also compatible with
Python3. but we plan to migrate master and octopus' qa suites to
python3 to ease the pain of maintainance in long term.
to achieve this goal, in this change, all non-master branches are now
tested using the "py2" teuthology branch.
once the python3 compatible changes are backported to octopus, "-t 2"
options will be removed from the command lines testing octopus.
Michael Fritch [Wed, 8 Apr 2020 20:42:12 +0000 (14:42 -0600)]
cephadm: show error when no command is specified
Traceback (most recent call last):
File "./cephadm", line 4452, in <module>
if args.func != command_check_host:
AttributeError: 'Namespace' object has no attribute 'func'
Jan Fajerski [Wed, 8 Apr 2020 09:55:57 +0000 (11:55 +0200)]
ceph-volume/batch: return success when all devices are filtered
batch should only return an error if some (but not all) devices are
filtered. When only some devices are filtered the resulting osd layout
could look very different from what a user expects. If all devies are
filtered just return success.
Fixes: https://tracker.ceph.com/issues/44994 Signed-off-by: Jan Fajerski <jfajerski@suse.com>
Sage Weil [Tue, 31 Mar 2020 21:47:54 +0000 (16:47 -0500)]
mon/MgrMap: use always_on_modules from most recent release
If our release isn't in the always_on_map, say because you just upgraded,
then start with the modules from the last/most recent release. This
should be sufficient to allow an upgrade when the mon is still the old
release and the mgrmap doesn't tell us what modules we should be enabling.
(Arguably, the always_on_modules should be enshrined in the MgrMap itself
and not in the mon?)
in tasks/module_selftest.yaml, `TestModuleSelftest.test_telegraf()` is
called. but we fail to prepare a unix domain socket to which the telegraf
module can send stats. and telegraf module does not catch
FileNotFoundError exception, so the exception is populated to ceph-mgr
and is found by the test, hence the test is marked a failure whenever
telegraf is tested.
in this change,
* catch this exception, so it won't be caught by ceph-mgr
* whitelist the error message, so the test can pass
qa/tasks/openssl_keys.py: sort cert configs before creating certs
we cannot rely on the order in which items are arranged in a dict, the
order varies from version to another. in Python2, it happens to work,
and we can always have the self-signed cert added first. but in Python3,
it does not. and an exception is thrown
```
teuthology.exceptions.ConfigError: ssl: ca root not found for
certificate rgw.client.0
```
in this change, before creating certs, the settings are reordered so
that the self-signed ones are created first.
in tasks/module_selftest.yaml, `TestModuleSelftest.test_telegraf()` is
called. but we fail to prepare a unix domain socket to which the telegraf
module can send stats. and telegraf module does not catch
FileNotFoundError exception, so the exception is populated to ceph-mgr
and is found by the test, hence the test is marked a failure whenever
telegraf is tested.
in this change,
* catch this exception, so it won't be caught by ceph-mgr
* whitelist the error message, so the test can pass
Varsha Rao [Tue, 31 Mar 2020 11:48:03 +0000 (17:18 +0530)]
mgr/volumes: Create multiple CephFS exports
Using the following fs nfs interface multiple exports can be created:
ceph fs nfs export create <fsname> <binding> --readonly --path=<pathname> --attach=<clusterid>
Varsha Rao [Mon, 30 Mar 2020 10:17:15 +0000 (15:47 +0530)]
mgr/volumes: Move ganesha common config to vstart
This is a preparatoy patch before calling orchestrator for nfs cluster
deployment. All the ganesha config is moved to vstart. Keyring creation is also
taken care by vstart.
The volumes fs nfs cluster create interface does the following things:
1) Create a common recovery pool for all ganesha clusters. Each cluster will
have their own namespace.
2) Create an empty rados conf object named 'conf-nfs' for saving all export
urls.
Call to orch interface will be done in future patch.
Varsha Rao [Tue, 24 Mar 2020 18:56:56 +0000 (00:26 +0530)]
mgr/volumes/nfs: Fix mypy errors
This patch fixes the following mypy errors:
volumes/module.py:9: note: In module imported here,
volumes/__init__.py:2: note: ... from here:
volumes/fs/nfs.py: note: In member "validate_path" of class "CephFSFSal":
volumes/fs/nfs.py:80: error: Name 're' is not defined
volumes/fs/nfs.py: note: In member "create_path" of class "CephFSFSal":
volumes/fs/nfs.py:83: error: Name 'CephFS' is not defined
volumes/fs/nfs.py: note: In member "_persist_daemon_configuration" of class "GaneshaConf":
volumes/fs/nfs.py:259: error: Need type annotation for 'daemon_map' (hint: "daemon_map: Dict[<type>, <type>] = ...")
volumes/fs/nfs.py: note: In member "create_instance" of class "NFSConfig":
volumes/fs/nfs.py:386: error: Incompatible types in assignment (expression has type "GaneshaConf", variable has type "str")
volumes/fs/nfs.py: note: In member "create_export" of class "NFSConfig":
volumes/fs/nfs.py:389: error: "str" has no attribute "create_export"
volumes/fs/nfs.py: note: In member "delete_export" of class "NFSConfig":
volumes/fs/nfs.py:404: error: "str" has no attribute "has_export"
volumes/fs/nfs.py:407: error: "str" has no attribute "remove_export"
volumes/fs/nfs.py:408: error: "str" has no attribute "reload_daemons"
volumes/__init__.py:2: note: In module imported here:
volumes/module.py: note: In member "_cmd_fs_nfs_export_create" of class "Module":
volumes/module.py:406: error: "str" has no attribute "check_fsal_valid"
volumes/module.py:407: error: "str" has no attribute "create_instance"
volumes/module.py:408: error: "str" has no attribute "create_export"
volumes/module.py: note: In member "_cmd_fs_nfs_export_delete" of class "Module":
volumes/module.py:412: error: "str" has no attribute "delete_export"
volumes/module.py: note: In member "_cmd_fs_nfs_cluster_create" of class "Module":
volumes/module.py:415: error: Incompatible types in assignment (expression has type "NFSConfig", variable has type "str")
volumes/module.py:416: error: "str" has no attribute "create_nfs_cluster"
projects / ceph.git / log