Lucian Petrut [Wed, 3 Feb 2021 08:59:24 +0000 (08:59 +0000)]
win32*.sh: move debug symbols to separate files
This patch simplifies releasing Windows binaries along with debug
symbols.
By default, we're going to provide minimum debug information (-g1).
The symbols are extracted from the binaries and placed in separate
files in the ".debug" folder, which is used by gdb implicitly.
This is more convenient than having separate versions of the binaries,
with or without debug symbols.
Lucian Petrut [Fri, 29 Jan 2021 11:03:20 +0000 (11:03 +0000)]
rbd: propagate WNBD start errors
This change will propagate the errors that WNBD may return when
spinning up the IO workers.
Also, we'll avoid removing the registry record for failed
non-persistent mappings. Those will be cleaned up when the service
restarts or when explicitly unmapped.
Lucian Petrut [Fri, 29 Jan 2021 09:54:10 +0000 (09:54 +0000)]
rbd: improve Windows remap failure handling
At the moment, if an image can't be remapped when the centralized
RBD service starts, the service will stop and already started
daemons will continue running.
This change adds a new option: "--remap-failure-fatal". If set,
when an image can't be remmaped, the service stops AND cleans up
the running daemons. By default, an error will be logged and the
service will continue running.
Lucian Petrut [Thu, 28 Jan 2021 14:08:29 +0000 (14:08 +0000)]
rbd: add image map timeouts on Windows
This change adds configurable timeouts used when starting the centralized RBD
service and mapping images.
This change also fixes an issue where the service would wait indefinitely for
a failed mapping. This issue was caused by the fact that multiple child
processes were inheriting pipe handles.
Note that we can't use timeouts with Windows anonymous pipes, which is why
we're going to use unique named pipes.
Lucian Petrut [Wed, 27 Jan 2021 12:58:48 +0000 (12:58 +0000)]
rbd: restrict Windows service exec calls
The centralized Ceph Windows service is responsible of managing
rbd-wnbd daemons. When starting, it's respawns the daemons using the
command line saved in the Windows registry. Also, for new mappings,
the command line is passed through a named pipe.
While writing to the according named pipe and windows registry entries
requires admin privileges, it's better to avoid running arbitrary
commands.
This patch will drop the executable from the commands that the
Ceph service accepts. Instead, it will only accept arguments that
are passed to the the binary that was used to start the service
(rbd-wnbd.exe).
Lucian Petrut [Wed, 27 Jan 2021 09:06:19 +0000 (09:06 +0000)]
rbd: allow non persistent Windows mappings
At the moment, all Windows RBD mappings are persistent, being
recreated when the Ceph service starts.
This change adds the "--non-persistent" flag to allow skipping
certain images.
Note that even for non-persistent mappings, we're still storing
data in the Windows registry, allowing us to retrieve image
details such as the rbd pool as well as the admin socket path.
When the daemons stop or the Ceph service starts, non-persistent
entries are cleaned up from the Windows registry.
Lucian Petrut [Tue, 26 Jan 2021 14:13:22 +0000 (14:13 +0000)]
rbd: delegate map requests to the Windows service
At the moment, rbd-wnbd daemons are tied to the Windows
session. When the user logs off or the WinRM remote session
terminates, the daemons are killed.
In order to avoid such issues, we'll delegate the daemon
initialization to the centralized Ceph Windows service.
We're using a named pipe for the Ceph service communication.
We're not using Ceph admin sockets for now since unix sockets
aren't available on Windows Server 2016.
Worth mentioning that the Ceph Windows service will restart
registered mappings when the host reboots based on the command
lines saved in the Windows registries. Writing to the registry
key as well as the service named pipe requires admin privileges.
`ShardServices::start_operation<T>()` spawns an operation,
calls `start()` on it and returns a future plus a smart
pointer controlling its life-time. Callers are responsible
to ensure the pointer doesn't go out-of-scope before
the entire execution is finished. This is error-prone.
`OSD::handle_osd_op()` forgets about its responsibility
which results in dangling `this` pointer of `ClientRequest`.
I believe the problem is much wider spread and the class
is just the tip of the iceberg.
In this commit `start_operation<T>()` is altered to extend
the life-time a bit. However, this isn't an ultimate solution
as callers are still able to e.g. put extra `this`-catching
lambdas on the returned future. A new `with_operation<T>()`-
like interface is expected as a follow-up.
also use updated build with sha1 of 1fadde94b08fab574b17637c2bebd2b1e7f9127b, the older version has an issue
where both libzbd and libzbd-dev packaged .pc file with the same path,
so they conflicted with each other. the new version addressed this
issue.
and install libzbd-dev instead of libzbd and libzbd-dev, as the package
name of the former could change if its so version is bumped, so use the
*-dev package is a safer choice from the perspective of testing the
dev build.
Kefu Chai [Tue, 19 Jan 2021 17:56:02 +0000 (01:56 +0800)]
pybind/mgr/dashboard: use _get_localized_key for composing store key
Just for keeping this 'encapsulated', there's _get_localized_key helper
in mgr_module (there's also a set_localized_store but it only deals with
the active manager)
As suggested by Ernesto Puerta <epuertat@redhat.com>
Kefu Chai [Tue, 19 Jan 2021 17:24:01 +0000 (01:24 +0800)]
pybind/mgr/cephadm/inventory: cast variable to expected type
to silence mypy warnings like:
cephadm/module.py:55: note: In module imported here,
cephadm/__init__.py:6: note: ... from here:
cephadm/inventory.py: note: In member "load" of class "SpecStore":
cephadm/inventory.py:130: error: Invalid index type "str" for "str"; expected type "Union[int, slice]"
cephadm/inventory.py:131: error: Invalid index type "str" for "str"; expected type "Union[int, slice]"
Kefu Chai [Tue, 19 Jan 2021 06:49:07 +0000 (14:49 +0800)]
pybind/mgr: correct annotation for BasePyOSDMap._apply_incremental()
it's incorrect. as the underlying C implementation,
"osdmap_apply_incremental()", actually expects an instance of
`BasePyOSDMapIncrementalType`. which is mapped to
`BasePyOSDMapIncremental` in Python. and this class is wrapped using
`OSDMapIncremental` in mgr_module.py.
Kefu Chai [Sun, 17 Jan 2021 08:36:35 +0000 (16:36 +0800)]
mgr/cephadm: use dict instead of defaultdict to silence mypy warning
otherwise we have:
1: cephadm/module.py:30: note: In module imported here:
1: cephadm/serve.py: note: In member "_check_for_strays" of class "CephadmServe":
1: cephadm/serve.py:395: error: Argument "default" to "get_metadata" of "MgrModule" has incompatible type "defaultdict[str, None]"; expected "Optional[Dict[str, str]]"
Kefu Chai [Sun, 17 Jan 2021 07:29:33 +0000 (15:29 +0800)]
mgr/progress: pass id to constructor of Event
to assure that the self.id is always valid, it helps to silence the
warnings from mypy like:
progress/module.py: note: In member "_refresh" of class "Event":
progress/module.py:46: error: Argument 1 to "update_progress_event" of "MgrModule" has incompatible type "Optional[str]"; expected "str"
Kefu Chai [Sun, 17 Jan 2021 01:35:36 +0000 (09:35 +0800)]
pybind/mgr/iostat: use correct types
for better readability and for the sake of correctness.
* do not cast denominator to float before diving by it
* use '//' for divding an int
* cast int to str as it will be used as an element in a list of str
Lucian Petrut [Thu, 7 Jan 2021 13:14:00 +0000 (13:14 +0000)]
cmake,win32*.sh: Windows DLL support
At the moment, the Windows binaries can only be linked statically.
This is less than ideal since:
* the resulting binaries can be quite large, especially when
including debug symbols
* librados and librbd cannot be used directly by 3rd party apps
* the build duration is increased
In order to do a dynamic build, we'll:
* add an option to win32_build.sh
* fix link order
* dynamically link boost
* disable the "-pie" flag when using Mingw, which generates incorrect
executable entry points
* by default, cmake generates import libs for executables. The issue
is that for rados.exe, it generates librados.dll.a, thus overriding
the librados.dll import library, which breaks the build process.
We'll configure cmake to skip generating import libs for executables.
The current bionice version triggers a podman/conmon bug that
truncates output, affecting both cephadm bootstrap when 'mgr dump' is
large, and teuthology 'pg dump' when it is large.
Ilya Dryomov [Wed, 20 Jan 2021 15:00:18 +0000 (16:00 +0100)]
qa/suites/krbd: add msgr2 modes to most subsuites
basic, rbd and rbd-nomount subsuites are expanded to run with each
of ms_mode=legacy, ms_mode=crc and ms_mode=secure. This increases
the total number of jobs in the suite from 100 to 220.
fsx, singleton and thrash subsuites choose ms_mode at random (from
the above plus ms_mode=prefer-crc).
Ilya Dryomov [Mon, 18 Jan 2021 12:49:49 +0000 (13:49 +0100)]
krbd: add support for msgr2
Recognize ms_mode map option and filter initial monitor addresses
accordingly: if ms_mode is not given or ms_mode=legacy, discard v2
addresses, otherwise discard v1 addresses.
Note that nothing was discarded (i.e. v2 addresses were passed to
the kernel) previously. The intent was to preserve that behaviour
in case ms_mode is not given, allowing to change the kernel default
in the future. However, it turns out that mount.ceph helper has
been misguidedly discarding v2 addresses since commit eae01275134e
("mount.ceph: fork a child to get info from local configuration"),
so that ship has sailed.
Or Ozeri [Sun, 24 Jan 2021 07:32:44 +0000 (09:32 +0200)]
librbd: do not share crypto image layers with ancestors
This commit creates a unique crypto image dispatch per image,
instead of sharing a single object with all ancestors.
This should fix a seg-fault caused when closing an encrypted cloned image.
projects / ceph.git / log