Publish fail: Request threads can only be created within the context of a running request

[gzc]

Hi, I am using appEngine flexible environment with Java8. I am using pubsub0.17.1-alpha. When publishing, I will have this exception. I tried to run publish in a background thread but things are still same.

Following is the exception stack

WARNING: [io.grpc.internal.ManagedChannelImpl-1] 
Failed to resolve name. status=Status{code=UNKNOWN, description=null, cause=java.lang.IllegalStateException: Request threads can only be created within the context of a running request. 
at com.google.appengine.repackaged.com.google.common.base.Preconditions.checkState(Preconditions.java:449) 
at com.google.apphosting.vmruntime.VmRequestThreadFactory.newThread(VmRequestThreadFactory.java:73)
at com.google.common.util.concurrent.ThreadFactoryBuilder$1.newThread(ThreadFactoryBuilder.java:162) 
at java.util.concurrent.ThreadPoolExecutor$Worker.<init>(ThreadPoolExecutor.java:612) 
at java.util.concurrent.ThreadPoolExecutor.addWorker(ThreadPoolExecutor.java:925) 
at java.util.concurrent.ThreadPoolExecutor.execute(ThreadPoolExecutor.java:1368) 
at io.grpc.internal.DnsNameResolver.resolve(DnsNameResolver.java:205) 
at io.grpc.internal.DnsNameResolver.start(DnsNameResolver.java:116) 
at io.grpc.internal.ManagedChannelImpl.exitIdleMode(ManagedChannelImpl.java:313) 
at io.grpc.internal.ManagedChannelImpl$4$1.run(ManagedChannelImpl.java:359) 
at io.grpc.internal.ChannelExecutor.drain(ChannelExecutor.java:87) 
at io.grpc.internal.ManagedChannelImpl$4.get(ManagedChannelImpl.java:361) 
at io.grpc.internal.ClientCallImpl.start(ClientCallImpl.java:221) 
at io.grpc.ForwardingClientCall.start(ForwardingClientCall.java:47) 
at com.google.api.gax.grpc.HeaderInterceptor$1.start(HeaderInterceptor.java:62) 
at io.grpc.stub.ClientCalls.startCall(ClientCalls.java:270) 
at io.grpc.stub.ClientCalls.asyncUnaryRequestCall(ClientCalls.java:249) 
at io.grpc.stub.ClientCalls.futureUnaryCall(ClientCalls.java:186) 
at com.google.pubsub.v1.PublisherGrpc$PublisherFutureStub.publish(PublisherGrpc.java:460) 
at com.google.cloud.pubsub.spi.v1.Publisher.publishOutstandingBatch(Publisher.java:333) 
at com.google.cloud.pubsub.spi.v1.Publisher.access$200(Publisher.java:76) 
at com.google.cloud.pubsub.spi.v1.Publisher$4.run(Publisher.java:271) 
at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511) 
at java.util.concurrent.FutureTask.run(FutureTask.java:266) 
at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$201(ScheduledThreadPoolExecutor.java:180) 
at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) 
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142) 
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617) 
at java.lang.Thread.run(Thread.java:745)

Thanks^

[tcoffee-google]

@gzc can you provide an example code snippet to help us reproduce the problem? Thanks!

[garrettjonesgoogle]

@neozwu could you take a look?

[neozwu]

Hi @gzc , I tested a GAE flex java 8 sample app with publishOutstandingBatch call (using google-cloud-pubsub-0.17.1-alpha), but cannot reproduce this error. Can you provide more information to help repro? I'd like to know in particular the dependency of your project, if possible, and how often you see this error. It is not clear to me why ThreadFactoryBuilder will end up using VmRequestThreadFactory.

[gzc]

@neozwu Hi,
I do something trick to bypass using appengine ThreadFactory

System.clearProperty("com.google.appengine.runtime.environment");

Then I will use java default thread factory and it works.

My repo has too many dependencies, I will give more info later.

Thanks ^

[garrettjonesgoogle]

I am going to close this issue because it appears that @gzc has a work around. Feel free to open if there are further issues.