This mostly adds the api described in
https://github.com/goodboy/tractor/issues/53#issuecomment-806258798
The first draft summary:
- formalize bidir steaming using the `trio.Channel` style interface
which we derive as a `MsgStream` type.
- add `Portal.open_context()` which provides a `trio.Nursery.start()`
remote task invocation style for setting up and tearing down tasks
contexts in remote actors.
- add a distinct `'started'` message to the ipc protocol to facilitate
`Context.start()` with a first return value.
- for our `ReceiveMsgStream` type, don't cancel the remote task in
`.aclose()`; this is now done explicitly by the surrounding `Context`
usage: `Context.cancel()`.
- streams in either direction still use a `'yield'` message keeping the
proto mostly symmetric without having to worry about which side is the
caller / portal opener.
- subtlety: only allow sending a `'stop'` message during a 2-way
streaming context from `ReceiveStream.aclose()`, detailed comment
with explanation is included.
Relates to #53
Since we currently have no real "discovery protocol" between process
trees, the current naive approach is to check via a connect and drop to
see if a TCP server is bound to a particular address during root actor
startup. This was a historical decision and had no real grounding beyond
taking a simple approach to get something working when the project
was first started.
This is obviously problematic from an error handling perspective since
we need to be able to avoid such quick connect-and-drops from cancelling
an "arbiter"'s (registry actor's) channel-msg loop machinery (which
would propagate and cancel the actor).
For now we map this particular TCP error, which gets remapped by `trio`
as a `trio.BrokenResourceError` to our own internal `TransportClosed`
which is swallowed by channel message loop processing and indicates
a graceful teardown of the far end actor.
This change some super old (and bad) code from the project's very early
days. For some redic reason i must have thought masking `trio`'s
internal stream / transport errors and a TCP EOF as `StopAsyncIteration`
somehow a good idea. The reality is you probably
want to know the difference between an unexpected transport error
and a simple EOF lol. This begins to resolve that by adding our own
special `TransportClosed` error to signal the "graceful" termination of
a channel's underlying transport. Oh, and this builds on the `msgspec`
integration which helped shed light on the core issues here B)
It's clear now that special attention is needed to handle the case where
a spawned `multiprocessing` proc is started but then the parent is
cancelled before the child can connect back; in this case we need to be
sure to kill the near-zombie child asap. This may end up being the
solution to other resiliency issues seen around mp with nested process
trees too. More testing is needed to be sure.
Relates to #84#89#134#146
NB: this is a breaking change removing support for `Portal.run()` being
able to invoke remote streaming functions and instead replacing the
method call with an async context manager api `Portal.open_stream_from()`
This style explicitly defines stream teardown at the call site instead
of expecting the user to handle tricky things correctly themselves: eg.
`async_geneartor.aclosing()`. Going forward `Portal.run()` can be used
only for invoking async functions.
Move receive stream into streaming modules and rebrand as a "message
stream". Factor out cancellation mechanics in `.aclose()` into the
`Context` type which will soon provide the api for for cancelling portal
invocations. Comment-stage a few methods on both types in anticipation
of a new bi-directional streaming api. Add a `MsgStream` bidirectional
channel type which will be the eventual type yielded from
`Context.open_stream()`. Adjust the response/dialog types to be the set
`{'asyncfun', 'asyncgen', 'context'}`. OH, and add async func checking
in `Portal.run()` to catch and error on sync funcs early.
You can always wrap a sync function in an async one and there seems to
be no good reason to support invoking them directly especially since
cancellation won't work without some thread hackery. If it's requested
we'll point users to `trio-parallel`.
Resolves#77
Add a sync method that can be used to cancel the current actor from
a synchronous context. This is useful in debugging situations where
sync debugger code may need to kill the process tree.
Also, make the internal "lifetime stack" a global var; easier to manage
from client code that may was to add callbacks prior to the actor
runtime being fully setup.
Using `None` as the default key for a `@msg.pub` can cause conflicts if
there is more then one "taskless" (no tasks={,} passed) pub offered on
an actor... So instead use the first trio "task name" (usually just the
function name) instead thus avoiding this very hard to debug and
understand problem.
Probably should throw in a test but I'm super lazy today.
This begins the move to dropping support for `tractor.run()` which we
don't really need since the runtime is started (as it always has been)
from a new sub-task / nursery. Instead this introduces starting the
actor tree through a `open_root_actor()` async context manager which
we'll likely implicitly call (from the root) on the first use of an
actor nursery.
Drop `_actor._start_actor()` and factor its contents into this new api.
Make `run()` and `run_daemon()` use `open_root_actor()` until we decide
to remove them.
Relates to #168 and #177
It turns out in order to maintain our sneaky little "call an `Actor`
method in this remote process" we still need the ability to invoke
functions from a namespace. We're currently using a "self" namespace as
a way to do this for internal inter-process method calling. Either way,
I see no reason not to keep a public method for this invoke style (we
just won't market it) since it is still how the machinery works
underneath.
This resolves and completes #69 allowing all RPC invocation APIs to pass
function references directly instead of explicit `str` names for the
target namespace and function (this is still done implicitly
underneath). This brings us closer to `trio`'s task running API as well
as acknowledges that any inter-host RPC system (and API) will likely
need to be implemented on top of local RPC primitives anyway. Even if
this ends up **not** being true we can always go to "function stubs" as
part of our IAC protocol or, add a new method to do explicit namespace
calls: `.run_from_module()` or whatever everyone votes on.
Resolves#69
Further, this commit drops `Actor.statespace` from the entire system
since a user can easily get this same functionality using module
level variables. Fix docs to match all these changes (luckily mostly
already done due to example scripts referencing).
Add a ``tractor._portal.StreamReceiveChannel.shield_channel()`` context
manager which allows for avoiding the closing of an IPC stream's
underlying channel for the purposes of task re-spawning. Sometimes you
might want to cancel a task consuming a stream but not tear down the IPC
between actors (the default). A common use can might be where the task's
"setup" work might need to be redone but you want to keep the
established portal / channel in tact despite the task restart.
Includes a test.
Turns out this is a lower level issue in terms of the stdlib's default
`pdb.Pdb` settings and how they conflict with `trio`s cancellation and
KBI handling. The details are hashed out more thoroughly in
python-trio/trio#1155. Maybe we can get a fix in trio so things are
solved under our feet :)
The channel server should be torn down *before* the rpc
task/service nursery. Do this explicitly even in the root's main task
to avoid a strange hang I found in the pubsub tests. Start dropping
the `warnings.warn()` usage.
Add `Actor._cancel_called` and `._cancel_complete` making it possible to
determine whether the actor has started the cancellation sequence and
whether that sequence has fully completed. This allows for blocking in
internal machinery tasks as necessary. Also, always trigger the end of
ongoing rpc tasks even if the last task errors; there's no guarantee the
trio cancellation semantics will guarantee us a nice internal "state"
without this.
For reliable remote cancellation we need to "report" `trio.Cancelled`s
(just like any other error) when exhausting a portal such that the
caller can make decisions about cancelling the respective actor if need
be.
Resolves#156
Every subactor in the tree now receives the socket (or whatever the
mailbox type ends up being) during startup and can call the new
`tractor._discovery.get_root()` function to get a portal to the current
root actor in their tree. The main reason for adding this atm is to
support nested child actors gaining access to the root's tty lock for
debugging.
Also, when a channel disconnects from a message loop, might as well kill
all its rpc tasks.
It's not like any of this code is really being used anyway since we
aren't indefinitely blocking for cancelled subactors to terminate (yet).
Drop the `do_hard_kill()` bit for now and just rely on the underlying
process api. Oh, and mark the nursery as cancelled asap.
Seems like the request task cancel scope is actually solving all the
deadlock issues and masking SIGINT isn't changing much behaviour at all.
I think let's keep it unmasked for now in case it does turn out useful
in cancelling from unrecoverable states while in debug.
This is needed in order to avoid the deadlock condition where
a child actor is waiting on the root actor's tty lock but it's parent
(possibly the root) is waiting on it to terminate after sending a cancel
request. The solution is simple: create a cancel scope around the
request in the child and always cancel it when a cancel request from the
parent arrives.
There seems to be no good reason not too since our cancellation
machinery/protocol should do this work when the root receives the
signal. This also (hopefully) helps with some debugging race condition
stuff.
This seems to prevent a certain class of bugs to do with the root actor
cancelling local tasks and getting into deadlock while children are
trying to acquire the tty lock. I'm not sure it's the best idea yet
since you're pretty much guaranteed to get "stuck" if a child activates
the debugger after the root has been cancelled (at least "stuck" in
terms of SIGINT being ignored). That kinda race condition seems to still
exist somehow: a child can "beat" the root to activating the tty lock
and the parent is stuck waiting on the child to terminate via its
nursery.
This aids with tearing down resources **after** the crash handling and
debugger have completed. Leaving this internal for now but should
eventually get a public convenience function like
`tractor.context_stack()`.
Keep an actor local (bool) flag which determines if there is already
a running debugger instance for the current process. If another task
tries to enter in this case, simply ignore it since allowing entry may
result in a deadlock where the new task will be sync waiting on the
parent stdio lock (a case that will never arrive due to the current
debugger's active use of it).
In the future we may want to allow FIFO queueing of local tasks where
instead of ignoring re-entrant breakpoints we allow tasks to async wait
for debugger release, though not sure the implications of that since
you'd likely want to support switching the debugger to the new task and
that could cause deadlocks where tasks are inter-dependent. It may be
more sane to just error on multiple breakpoint requests within an actor.
This is the first step in addressing #113 and the initial support
of #130. Basically this allows (sub)processes to engage the `pdbpp`
debug machinery which read/writes the root actor's tty but only in
a FIFO semaphored way such that no two processes are using it
simultaneously. That means you can have multiple actors enter a trace or
crash and run the debugger in a sensible way without clobbering each
other's access to stdio. It required adding some "tear down hooks" to
a custom `pdbpp.Pdb` type such that we release a child's lock on the
parent on debugger exit (in this case when either of the "continue" or
"quit" commands are issued to the debugger console).
There's some code left commented in anticipation of full support for
issue #130 where we're need to actually capture and feed stdin to the
target (remote) actor which won't necessarily being running on the same
host.
Allow entering and attaching to a `pdb` instance in a child process.
The current hackery is to have the child make an rpc to the parent and
ask it to hijack stdin, once complete the child enters a `pdb` blocking
method. The parent then relays all stdin input to the child thus
controlling the "remote" debugger.
A few things were added to accomplish this:
- tracking the mapping of subactors to their parent nurseries
- in the root actor, cancelling all nurseries under the root `trio` task
on cancellation (i.e. `Actor.cancel()`)
- pass a "runtime vars" map down the actor tree for propagating global state
In an effort acquire more deterministic actor cancellation,
this adds a clearer and more resilient (whilst possibly a bit
slower) internal nursery structure with explicit semantics for
clarifying the task-scope shutdown sequence.
Namely, on cancellation, the explicit steps are now:
- cancel all currently running rpc tasks and wait
for them to complete
- cancel the channel server and wait for it to complete
- cancel the msg loop for the channel with the immediate parent
- de-register with arbiter if possible
- wait on remaining connections to release
- exit process
To accomplish this add a new nursery called the "service nursery" which
spawns all rpc tasks **instead of using** the "root nursery". The root
is now used solely for async launching the msg loop for the primary
channel with the parent such that it is (nearly) the last thing torn
down on cancellation.
In the future it should also be possible to have `self.cancel()` return
a result to the parent once the runtime is sure that the rest of the
shutdown is atomic; this would allow for a true unbounded shield in
`Portal.cancel_actor()`. This will likely require that the error
handling blocks in `Actor._async_main()` are moved "inside" the root
nursery block such that the msg loop with the parent truly is the last
thing to terminate.
Always shield waiting for he process and always run
``trio.Process.__aexit__()`` on teardown. This enforces
that shutdown happens to due cancellation triggered inside
the sub-actor instead of the process being killed externally
by the parent.
Trio will kill subprocesses via `Process.__aexit__()` using a `finally:`
block (which, yes, will get triggered on cancellation) so we avoid that
until true process "tear down" since subactors do many things during
graceful shutdown (such as de-registering from the name discovery
system). Oddly this only seems to be an issue during cancellation of
infinite stream consumption.
Resolves#141
In order to have reliable subactor startup we need the following
sequence to take place:
- connect to the parent actor, handshake and receive runtime state
- load exposed modules into memory
- start the channel server up fully using the provided bind address
- finally, start processing new messages from the parent
Add a bunch more comments to clarify all this.
Using the context manager interface does some extra teardown beyond simply
calling `.wait()`. Pass the subactor's "uid" on the exec line for
debugging purposes when monitoring the process tree from the OS.
Hard code the child script module path to avoid a double import warning.
This is an edit to factor out changes needed for the `asyncio` in guest mode
integration (which currently isn't tested well) so that later more pertinent
changes (which are tested well) can be rebased off of this branch and
merged into mainline sooner. The *infect_asyncio* branch will need to be
rebased onto this branch as well before merge to mainline.
This is an initial solution for #120.
Allow spawning `asyncio` based actors which run `trio` in guest
mode. This enables spawning `tractor` actors on top of the `asyncio`
event loop whilst still leveraging the SC focused internal actor
supervision machinery. Add a `tractor.to_syncio.run()` api to allow
spawning tasks on the `asyncio` loop from an embedded (remote) `trio`
task and return or stream results all the way back through the `tractor`
IPC system using a very similar api to portals.
One outstanding problem is getting SC around calls to
`asyncio.create_task()`. Currently a task that crashes isn't able to
easily relay the error to the embedded `trio` task without us fully
enforcing the portals based message protocol (which seems superfluous
given the error ref is in process). Further experiments using `anyio`
task groups may alleviate this.
The logic in the `ActorNursery` block is critical to cancellation
semantics and in particular, understanding how supervisor strategies are
invoked. Stick in a bunch of explanatory comments to clear up these
details and also prepare to introduce more supervisor strats besides
the current one-cancels-all approach.
Instead of hackery trying to map modules manually from the filesystem
let Python do all the work by simply copying what ``multiprocessing``
does to "fixup the __main__ module" in spawned subprocesses. The new
private module ``_mp_fixup_main.py`` is simply cherry picked code from
``multiprocessing.spawn`` which does just that. We only need these
"fixups" when using a backend other then ``multiprocessing``; for
now just when using ``trio_run_in_process``.
Thanks to @salotz for pointing out that the first example in the docs
was broken. Though it's somewhat embarrassing this might also explain
the problem in #79 and certain issues in #59...
The solution here is to import the target RPC module using the its
unique basename and absolute filepath in the sub-actor that requires it.
Special handling for `__main__` and `__mp_main__` is needed since the
spawned subprocess will have no knowledge about these parent-
-state-specific module variables. Solution: map the modules name to the
respective module file basename in the child process since the module
variables will of course have different values in children.
Add a `--spawn-backend` option which can be set to one of {'mp',
'trio_run_in_process'} which will either run the test suite using the
`multiprocessing` or `trio-run-in-process` backend respectively.
Currently trying to run both in the same session can result in hangs
seemingly due to a lack of cleanup of forkservers / resource trackers
from `multiprocessing` which cause broken pipe errors on occasion (no
idea on the details).
For `test_cancellation.py::test_nested_multierrors`, use less nesting
when mp is used since it breaks if we push it too hard with the
whole recursive subprocess spawning thing...
Set `trio-run-in-process` as the default on *nix systems and
`multiprocessing`'s spawn method on Windows. Enable overriding the
default choice using `tractor._spawn.try_set_start_method()`. Allows
for easy runs of the test suite using a user chosen backend.
This took a ton of tinkering and a rework of the actor nursery tear down
logic. The main changes include:
- each subprocess is now spawned from inside a trio task
from one of two containing nurseries created in the body of
`tractor.open_nursery()`: one for `run_in_actor()` processes and one for
`start_actor()` "daemons". This is to address the need for
`trio-run-in_process.open_in_process()` opening a nursery which must
be closed from the same task that opened it. Using this same approach
for `multiprocessing` seems to work well. The nurseries are waited in
order (rip actors then daemon actors) during tear down which allows
for avoiding the recursive re-entry of `ActorNursery.wait()` handled
prior.
- pull out all the nested functions / closures that were in
`ActorNursery.wait()` and move into the `_spawn` module such that
that process shutdown logic takes place in each containing task's
code path. This allows for vastly simplifying `.wait()` to just contain an
event trigger which initiates process waiting / result collection.
Likely `.wait()` should just be removed since it can no longer be used
to synchronously wait on the actor nursery.
- drop `ActorNursery.__aenter__()` / `.__atexit__()` and move this
"supervisor" tear down logic into the closing block of `open_nursery()`.
This not only cleans makes the code more comprehensible it also
makes our nursery implementation look more like the one in `trio`.
Resolves#93
Get a few more things working:
- fail reliably when remote module loading goes awry
- do a real hacky job of module loading using `sys.path` stuffsies
- we're still totally borked when trying to spin up and quickly cancel
a bunch of subactors...
It's a small move forward I guess.
Prepend the actor and task names in each log emission. This makes
debugging much more sane since you can see from which process and
running task the log message originates from!
Resolves#13
If a nursery fails to cancel (some sub-actors presumably) then hard kill
the whole process tree to avoid hangs during a catastrophic failure.
This logic may get factored out (and changed) as we introduce custom
supervisor strategies.
`trio.MultiError` isn't an `Exception` (derived instead from
`BaseException`) so we have to specially catch it in the task
invocation machinery and ship it upwards (like regular errors)
since nurseries running in sub-actors can raise them.
Add `@tractor.stream` which must be used to denote non async generator
streaming functions which use the `tractor.Context` API to push values.
This enforces a more explicit denotation as well as allows enforcing the
declaration of the `ctx` argument in definitions.
This begins moving toward explicitly decorated "streaming functions"
instead of checking for a `ctx` arg in the signature.
- provide each context with its task's top level `trio.CancelScope`
such that tasks can cancel themselves explictly if needed via calling
`Context.cancel_scope()`
- make `Actor.cancel_task()` a private method (`_cancel_task()`) and
handle remote rpc calls specially such that the caller does not need
to provide the `chan` argument; non-primitive types can't be passed on
the wire and we don't want the client actor be require knowledge of
the channel instance the request is associated with. This also ties into
how we're tracking tasks right now (`Actor._rpc_tasks` is keyed by the
call id, a UUID, *plus* the channel).
- make `_do_handshake` a private actor method
- use UUID version 4
Add full support for using the "spawn" process starting method as per:
https://docs.python.org/3/library/multiprocessing.html#contexts-and-start-methods
Add a `spawn_method` argument to `tractor.run()` for specifying the
desired method explicitly. By default use the "fastest" method available.
On *nix systems this is the original "forkserver" method.
This should be the solution to getting windows support!
Resolves#60