diff --git a/changelog.d/18964.misc b/changelog.d/18964.misc
new file mode 100644
index 00000000000..69be53ad27b
--- /dev/null
+++ b/changelog.d/18964.misc
@@ -0,0 +1 @@
+Fix `run_coroutine_in_background(...)` incorrectly handling logcontext.
diff --git a/synapse/logging/context.py b/synapse/logging/context.py
index aa4b98e7c76..b5b434f3a8d 100644
--- a/synapse/logging/context.py
+++ b/synapse/logging/context.py
@@ -802,13 +802,15 @@ def run_in_background(
     deferred returned by the function completes.
 
     To explain how the log contexts work here:
-     - When `run_in_background` is called, the current context is stored ("original"),
-       we kick off the background task in the current context, and we restore that
-       original context before returning
-     - When the background task finishes, we don't want to leak our context into the
-       reactor which would erroneously get attached to the next operation picked up by
-       the event loop. We add a callback to the deferred which will clear the logging
-       context after it finishes and yields control back to the reactor.
+     - When `run_in_background` is called, the calling logcontext is stored
+       ("original"), we kick off the background task in the current context, and we
+       restore that original context before returning.
+     - For a completed deferred, that's the end of the story.
+     - For an incomplete deferred, when the background task finishes, we don't want to
+       leak our context into the reactor which would erroneously get attached to the
+       next operation picked up by the event loop. We add a callback to the deferred
+       which will clear the logging context after it finishes and yields control back to
+       the reactor.
 
     Useful for wrapping functions that return a deferred or coroutine, which you don't
     yield or await on (for instance because you want to pass it to
@@ -857,22 +859,36 @@ def run_in_background(
 
     # The deferred has already completed
     if d.called and not d.paused:
-        # The function should have maintained the logcontext, so we can
-        # optimise out the messing about
+        # If the function messes with logcontexts, we can assume it follows the Synapse
+        # logcontext rules (Rules for functions returning awaitables: "If the awaitable
+        # is already complete, the function returns with the same logcontext it started
+        # with."). If it function doesn't touch logcontexts at all, we can also assume
+        # the logcontext is unchanged.
+        #
+        # Either way, the function should have maintained the calling logcontext, so we
+        # can avoid messing with it further. Additionally, if the deferred has already
+        # completed, then it would be a mistake to then add a deferred callback (below)
+        # to reset the logcontext to the sentinel logcontext as that would run
+        # immediately (remember our goal is to maintain the calling logcontext when we
+        # return).
         return d
 
-    # The function may have reset the context before returning, so we need to restore it
-    # now.
+    # Since the function we called may follow the Synapse logcontext rules (Rules for
+    # functions returning awaitables: "If the awaitable is incomplete, the function
+    # clears the logcontext before returning"), the function may have reset the
+    # logcontext before returning, so we need to restore the calling logcontext now
+    # before we return ourselves.
     #
     # Our goal is to have the caller logcontext unchanged after firing off the
     # background task and returning.
     set_current_context(calling_context)
 
-    # The original logcontext will be restored when the deferred completes, but
-    # there is nothing waiting for it, so it will get leaked into the reactor (which
-    # would then get picked up by the next thing the reactor does). We therefore
-    # need to reset the logcontext here (set the `sentinel` logcontext) before
-    # yielding control back to the reactor.
+    # If the function we called is playing nice and following the Synapse logcontext
+    # rules, it will restore original calling logcontext when the deferred completes;
+    # but there is nothing waiting for it, so it will get leaked into the reactor (which
+    # would then get picked up by the next thing the reactor does). We therefore need to
+    # reset the logcontext here (set the `sentinel` logcontext) before yielding control
+    # back to the reactor.
     #
     # (If this feels asymmetric, consider it this way: we are
     # effectively forking a new thread of execution. We are
@@ -894,10 +910,9 @@ def run_coroutine_in_background(
     Useful for wrapping coroutines that you don't yield or await on (for
     instance because you want to pass it to deferred.gatherResults()).
 
-    This is a special case of `run_in_background` where we can accept a
-    coroutine directly rather than a function. We can do this because coroutines
-    do not run until called, and so calling an async function without awaiting
-    cannot change the log contexts.
+    This is a special case of `run_in_background` where we can accept a coroutine
+    directly rather than a function. We can do this because coroutines do not continue
+    running once they have yielded.
 
     This is an ergonomic helper so we can do this:
     ```python
@@ -908,33 +923,7 @@ def run_coroutine_in_background(
     run_in_background(lambda: func1(arg1))
     ```
     """
-    calling_context = current_context()
-
-    # Wrap the coroutine in a deferred, which will have the side effect of executing the
-    # coroutine in the background.
-    d = defer.ensureDeferred(coroutine)
-
-    # The function may have reset the context before returning, so we need to restore it
-    # now.
-    #
-    # Our goal is to have the caller logcontext unchanged after firing off the
-    # background task and returning.
-    set_current_context(calling_context)
-
-    # The original logcontext will be restored when the deferred completes, but
-    # there is nothing waiting for it, so it will get leaked into the reactor (which
-    # would then get picked up by the next thing the reactor does). We therefore
-    # need to reset the logcontext here (set the `sentinel` logcontext) before
-    # yielding control back to the reactor.
-    #
-    # (If this feels asymmetric, consider it this way: we are
-    # effectively forking a new thread of execution. We are
-    # probably currently within a ``with LoggingContext()`` block,
-    # which is supposed to have a single entry and exit point. But
-    # by spawning off another deferred, we are effectively
-    # adding a new exit point.)
-    d.addBoth(_set_context_cb, SENTINEL_CONTEXT)
-    return d
+    return run_in_background(lambda: coroutine)
 
 
 T = TypeVar("T")
diff --git a/tests/util/test_logcontext.py b/tests/util/test_logcontext.py
index 0ecf712bab7..651dd844835 100644
--- a/tests/util/test_logcontext.py
+++ b/tests/util/test_logcontext.py
@@ -22,7 +22,6 @@
 import logging
 from typing import Callable, Generator, cast
 
-import twisted.python.failure
 from twisted.internet import defer, reactor as _reactor
 
 from synapse.logging.context import (
@@ -33,6 +32,7 @@
     current_context,
     make_deferred_yieldable,
     nested_logging_context,
+    run_coroutine_in_background,
     run_in_background,
 )
 from synapse.types import ISynapseReactor
@@ -249,73 +249,80 @@ async def competing_callback() -> None:
         # Back to the sentinel context
         self._check_test_key("sentinel")
 
-    def _test_run_in_background(self, function: Callable[[], object]) -> defer.Deferred:
-        sentinel_context = current_context()
+    async def _test_run_in_background(self, function: Callable[[], object]) -> None:
+        clock = Clock(reactor)
+
+        # Sanity check that we start in the sentinel context
+        self._check_test_key("sentinel")
 
-        callback_completed = False
+        callback_finished = False
 
         with LoggingContext("foo"):
-            # fire off function, but don't wait on it.
-            d2 = run_in_background(function)
+            # Fire off the function, but don't wait on it.
+            deferred = run_in_background(function)
+            self._check_test_key("foo")
 
-            def cb(res: object) -> object:
-                nonlocal callback_completed
-                callback_completed = True
-                return res
+            def callback(result: object) -> object:
+                nonlocal callback_finished
+                callback_finished = True
+                # Pass through the result
+                return result
 
-            d2.addCallback(cb)
+            # We `addBoth` because when exceptions happen, we still want to mark the
+            # callback as finished so that the test can complete and we see the
+            # underlying error.
+            deferred.addBoth(callback)
 
             self._check_test_key("foo")
 
-        # now wait for the function under test to have run, and check that
-        # the logcontext is left in a sane state.
-        d2 = defer.Deferred()
-
-        def check_logcontext() -> None:
-            if not callback_completed:
-                reactor.callLater(0.01, check_logcontext)
-                return
+            # Now wait for the function under test to have run, and check that
+            # the logcontext is left in a sane state.
+            while not callback_finished:
+                await clock.sleep(0)
+                self._check_test_key("foo")
 
-            # make sure that the context was reset before it got thrown back
-            # into the reactor
-            try:
-                self.assertIs(current_context(), sentinel_context)
-                d2.callback(None)
-            except BaseException:
-                d2.errback(twisted.python.failure.Failure())
-
-        reactor.callLater(0.01, check_logcontext)
+        self.assertTrue(
+            callback_finished,
+            "Callback never finished which means the test probably didn't wait long enough",
+        )
 
-        # test is done once d2 finishes
-        return d2
+        # Back to the sentinel context
+        self._check_test_key("sentinel")
 
     @logcontext_clean
-    def test_run_in_background_with_blocking_fn(self) -> defer.Deferred:
+    async def test_run_in_background_with_blocking_fn(self) -> None:
         async def blocking_function() -> None:
             await Clock(reactor).sleep(0)
 
-        return self._test_run_in_background(blocking_function)
+        await self._test_run_in_background(blocking_function)
 
     @logcontext_clean
-    def test_run_in_background_with_non_blocking_fn(self) -> defer.Deferred:
+    async def test_run_in_background_with_non_blocking_fn(self) -> None:
         @defer.inlineCallbacks
         def nonblocking_function() -> Generator["defer.Deferred[object]", object, None]:
             with PreserveLoggingContext():
                 yield defer.succeed(None)
 
-        return self._test_run_in_background(nonblocking_function)
+        await self._test_run_in_background(nonblocking_function)
 
     @logcontext_clean
-    def test_run_in_background_with_chained_deferred(self) -> defer.Deferred:
+    async def test_run_in_background_with_chained_deferred(self) -> None:
         # a function which returns a deferred which looks like it has been
         # called, but is actually paused
         def testfunc() -> defer.Deferred:
             return make_deferred_yieldable(_chained_deferred_function())
 
-        return self._test_run_in_background(testfunc)
+        await self._test_run_in_background(testfunc)
 
     @logcontext_clean
-    def test_run_in_background_with_coroutine(self) -> defer.Deferred:
+    async def test_run_in_background_with_coroutine(self) -> None:
+        """
+        Test `run_in_background` with a coroutine that yields control back to the
+        reactor.
+
+        This will stress the logic around incomplete deferreds in `run_in_background`.
+        """
+
         async def testfunc() -> None:
             self._check_test_key("foo")
             d = defer.ensureDeferred(Clock(reactor).sleep(0))
@@ -323,14 +330,111 @@ async def testfunc() -> None:
             await d
             self._check_test_key("foo")
 
-        return self._test_run_in_background(testfunc)
+        await self._test_run_in_background(testfunc)
 
     @logcontext_clean
-    def test_run_in_background_with_nonblocking_coroutine(self) -> defer.Deferred:
+    async def test_run_in_background_with_nonblocking_coroutine(self) -> None:
+        """
+        Test `run_in_background` with a "nonblocking" coroutine (never yields control
+        back to the reactor).
+
+        This will stress the logic around completed deferreds in `run_in_background`.
+        """
+
         async def testfunc() -> None:
             self._check_test_key("foo")
 
-        return self._test_run_in_background(testfunc)
+        await self._test_run_in_background(testfunc)
+
+    @logcontext_clean
+    async def test_run_coroutine_in_background(self) -> None:
+        """
+        Test `run_coroutine_in_background` with a coroutine that yields control back to the
+        reactor.
+
+        This will stress the logic around incomplete deferreds in `run_coroutine_in_background`.
+        """
+        clock = Clock(reactor)
+
+        # Sanity check that we start in the sentinel context
+        self._check_test_key("sentinel")
+
+        callback_finished = False
+
+        async def competing_callback() -> None:
+            nonlocal callback_finished
+            try:
+                # The callback should have the same logcontext as the caller
+                self._check_test_key("foo")
+
+                with LoggingContext("competing"):
+                    await clock.sleep(0)
+                    self._check_test_key("competing")
+
+                self._check_test_key("foo")
+            finally:
+                # When exceptions happen, we still want to mark the callback as finished
+                # so that the test can complete and we see the underlying error.
+                callback_finished = True
+
+        with LoggingContext("foo"):
+            run_coroutine_in_background(competing_callback())
+            self._check_test_key("foo")
+            await clock.sleep(0)
+            self._check_test_key("foo")
+
+        self.assertTrue(
+            callback_finished,
+            "Callback never finished which means the test probably didn't wait long enough",
+        )
+
+        # Back to the sentinel context
+        self._check_test_key("sentinel")
+
+    @logcontext_clean
+    async def test_run_coroutine_in_background_with_nonblocking_coroutine(self) -> None:
+        """
+        Test `run_coroutine_in_background` with a "nonblocking" coroutine (never yields control
+        back to the reactor).
+
+        This will stress the logic around completed deferreds in `run_coroutine_in_background`.
+        """
+        # Sanity check that we start in the sentinel context
+        self._check_test_key("sentinel")
+
+        callback_finished = False
+
+        async def competing_callback() -> None:
+            nonlocal callback_finished
+            try:
+                # The callback should have the same logcontext as the caller
+                self._check_test_key("foo")
+
+                with LoggingContext("competing"):
+                    # We `await` here but there is nothing to wait for here since the
+                    # deferred is already complete so we should immediately continue
+                    # executing in the same context.
+                    await defer.succeed(None)
+
+                    self._check_test_key("competing")
+
+                self._check_test_key("foo")
+            finally:
+                # When exceptions happen, we still want to mark the callback as finished
+                # so that the test can complete and we see the underlying error.
+                callback_finished = True
+
+        with LoggingContext("foo"):
+            run_coroutine_in_background(competing_callback())
+            self._check_test_key("foo")
+
+        self.assertTrue(
+            callback_finished,
+            "Callback never finished which means the test probably didn't wait long enough",
+        )
+
+        # Back to the sentinel context
+        self._check_test_key("sentinel")
 
     @logcontext_clean
     @defer.inlineCallbacks