Fix issue when running shards on Android (#1853)

[tokou]

Proposed changes

https://github.com/mobile-dev-inc/maestro/blob/53e31b47e43ff7482fe52e26c6d6cb05285bd2fa/maestro-client/src/main/java/maestro/drivers/AndroidDriver.kt#L180-L182

The issue seems to originate in the simultaneous call to blockingStubWithTimeout.deviceInfo in AndroidDriver#deviceInfo() from Maestro#deviceInfo() which causes the following error:

Stacktrace "java.net.ConnectException: Connection refused"

io.grpc.StatusRuntimeException: UNAVAILABLE: io exception
	at io.grpc.stub.ClientCalls.toStatusRuntimeException(ClientCalls.java:275)
	at io.grpc.stub.ClientCalls.getUnchecked(ClientCalls.java:256)
	at io.grpc.stub.ClientCalls.blockingUnaryCall(ClientCalls.java:169)
	at maestro_android.MaestroDriverGrpc$MaestroDriverBlockingStub.deviceInfo(MaestroDriverGrpc.java:634)
	at maestro.drivers.AndroidDriver$deviceInfo$1.invoke(AndroidDriver.kt:182)
	at maestro.drivers.AndroidDriver$deviceInfo$1.invoke(AndroidDriver.kt:181)
	at maestro.drivers.AndroidDriver.runDeviceCall(AndroidDriver.kt:1070)
	at maestro.drivers.AndroidDriver.deviceInfo(AndroidDriver.kt:181)
	at maestro.Maestro.fetchDeviceInfo(Maestro.kt:63)
	at maestro.Maestro.access$fetchDeviceInfo(Maestro.kt:39)
	at maestro.Maestro$cachedDeviceInfo$2.invoke(Maestro.kt:47)
	at maestro.Maestro$cachedDeviceInfo$2.invoke(Maestro.kt:46)
	at kotlin.SynchronizedLazyImpl.getValue(LazyJVM.kt:74)
	at maestro.Maestro.getCachedDeviceInfo(Maestro.kt:46)
	at maestro.Maestro.deviceInfo(Maestro.kt:57)
	at maestro.orchestra.Orchestra.initJsEngine(Orchestra.kt:217)
	at maestro.orchestra.Orchestra.runFlow(Orchestra.kt:107)
	at maestro.cli.runner.TestSuiteInteractor.runFlow(TestSuiteInteractor.kt:250)
	at maestro.cli.runner.TestSuiteInteractor.runTestSuite(TestSuiteInteractor.kt:82)
	at maestro.cli.command.TestCommand$handleSessions$1$1$results$1$1$1.invoke(TestCommand.kt:280)
	at maestro.cli.command.TestCommand$handleSessions$1$1$results$1$1$1.invoke(TestCommand.kt:258)
	at maestro.cli.session.MaestroSessionManager.newSession(MaestroSessionManager.kt:101)
	at maestro.cli.session.MaestroSessionManager.newSession$default(MaestroSessionManager.kt:54)
	at maestro.cli.command.TestCommand$handleSessions$1$1$results$1$1.invokeSuspend(TestCommand.kt:258)
	at kotlin.coroutines.jvm.internal.BaseContinuationImpl.resumeWith(ContinuationImpl.kt:33)
	at kotlinx.coroutines.DispatchedTask.run(DispatchedTask.kt:104)
	at kotlinx.coroutines.internal.LimitedDispatcher$Worker.run(LimitedDispatcher.kt:111)
	at kotlinx.coroutines.scheduling.TaskImpl.run(Tasks.kt:99)
	at kotlinx.coroutines.scheduling.CoroutineScheduler.runSafely(CoroutineScheduler.kt:585)
	at kotlinx.coroutines.scheduling.CoroutineScheduler$Worker.executeTask(CoroutineScheduler.kt:802)
	at kotlinx.coroutines.scheduling.CoroutineScheduler$Worker.runWorker(CoroutineScheduler.kt:706)
	at kotlinx.coroutines.scheduling.CoroutineScheduler$Worker.run(CoroutineScheduler.kt:693)
Caused by: io.netty.channel.AbstractChannel$AnnotatedConnectException: Connection refused: localhost/[0:0:0:0:0:0:0:1]:7078
Caused by: java.net.ConnectException: Connection refused
	at java.base/sun.nio.ch.Net.pollConnect(Native Method)
	at java.base/sun.nio.ch.Net.pollConnectNow(Net.java:672)
	at java.base/sun.nio.ch.SocketChannelImpl.finishConnect(SocketChannelImpl.java:973)
	at io.netty.channel.socket.nio.NioSocketChannel.doFinishConnect(NioSocketChannel.java:337)
	at io.netty.channel.nio.AbstractNioChannel$AbstractNioUnsafe.finishConnect(AbstractNioChannel.java:334)
	at io.netty.channel.nio.NioEventLoop.processSelectedKey(NioEventLoop.java:776)
	at io.netty.channel.nio.NioEventLoop.processSelectedKeysOptimized(NioEventLoop.java:724)
	at io.netty.channel.nio.NioEventLoop.processSelectedKeys(NioEventLoop.java:650)
	at io.netty.channel.nio.NioEventLoop.run(NioEventLoop.java:562)
	at io.netty.util.concurrent.SingleThreadEventExecutor$4.run(SingleThreadEventExecutor.java:997)
	at io.netty.util.internal.ThreadExecutorMap$2.run(ThreadExecutorMap.java:74)
	at io.netty.util.concurrent.FastThreadLocalRunnable.run(FastThreadLocalRunnable.java:30)
	at java.base/java.lang.Thread.run(Thread.java:1589)

Even though cachedDeviceConfig is lazy which means it's synchronized, it seems that the issue happens because all the calls come from on same thread.

https://github.com/mobile-dev-inc/maestro/blob/f9f4cc864116cc0e843fc64f795dbada2e7c42e2/maestro-client/src/main/java/maestro/Maestro.kt#L46-L48

I added a semaphore in AndroidDevice#runDeviceCall to ensure it can only be called once at a time.

https://github.com/mobile-dev-inc/maestro/blob/95ee50bb1109b4b8bfd61b4e11097a89eaae0c9a/maestro-client/src/main/java/maestro/drivers/AndroidDriver.kt#L1065-L1067

This seems to have fixed the issue. From now on, calls to runDeviceCall will not happen synchronously.

Testing

• I removed the workaround by deleting the Thread.sleep and rebuilding
• Launched 2 emulators
• Ran maestro test --shards 2 flows/ with a folder with around 7 successful flows
• I had the issue way more than half the time
• Detected the origin of the issue and added the semaphore
• Couldn't get the issue again after running multiple times

Issues fixed

Linked #1867
Fixes #1853

[bartekpacia]

Hey @tokou, it seems that this is never starting the test on CI. Any idea?

[tokou]

@bartekpacia No maybe just retrigger it?

[bartekpacia]

I did retrigger, unfortunately the result is the same

[tokou]

@bartekpacia I need some help here. I couldn't repro the issue and don't have access to more details from the CI. So I don't really know what to do next.

[bartekpacia]

Hey, unfortunately I don't have any ideas too (aside from capturing as much log output with adb logcat into some file and then exposing that file as a workflow artifact, and trying to go from there).