Threads hanging due to errors in OkHttp TaskRunner thread

[GeorgeFitzpatrick]

After sending a lot of consecutive requests, around 5/okhttp client instance/second, from anywhere between half an hour to 3 hours I start to get the following exception:

Exception in thread "OkHttp TaskRunner" java.lang.RuntimeException: invalid key or spec in GCM mode
	at java.base/sun.security.ssl.SSLCipher$T13GcmWriteCipherGenerator$GcmWriteCipher.encrypt(SSLCipher.java:2023)
	at java.base/sun.security.ssl.OutputRecord.t13Encrypt(OutputRecord.java:500)
	at java.base/sun.security.ssl.OutputRecord.encrypt(OutputRecord.java:463)
	at java.base/sun.security.ssl.SSLSocketOutputRecord.encodeAlert(SSLSocketOutputRecord.java:80)
	at java.base/sun.security.ssl.TransportContext.warning(TransportContext.java:248)
	at java.base/sun.security.ssl.SSLSocketImpl.deliverClosedNotify(SSLSocketImpl.java:639)
	at java.base/sun.security.ssl.SSLSocketImpl.closeNotify(SSLSocketImpl.java:627)
	at java.base/sun.security.ssl.SSLSocketImpl.duplexCloseOutput(SSLSocketImpl.java:550)
	at java.base/sun.security.ssl.SSLSocketImpl.close(SSLSocketImpl.java:485)
	at okhttp3.internal.Util.closeQuietly(Util.kt:498)
	at okhttp3.internal.connection.RealConnectionPool.cleanup(RealConnectionPool.kt:184)
	at okhttp3.internal.connection.RealConnectionPool$cleanupTask$1.runOnce(RealConnectionPool.kt:44)
	at okhttp3.internal.concurrent.TaskRunner.runTask(TaskRunner.kt:116)
	at okhttp3.internal.concurrent.TaskRunner.access$runTask(TaskRunner.kt:42)
	at okhttp3.internal.concurrent.TaskRunner$runnable$1.run(TaskRunner.kt:65)
	at java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
	at java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628)
	at java.base/java.lang.Thread.run(Thread.java:834)
Caused by: java.security.InvalidAlgorithmParameterException: Cannot reuse iv for GCM encryption
	at java.base/com.sun.crypto.provider.CipherCore.init(CipherCore.java:582)
	at java.base/com.sun.crypto.provider.AESCipher.engineInit(AESCipher.java:346)
	at java.base/javax.crypto.Cipher.init(Cipher.java:1431)
	at java.base/sun.security.ssl.SSLCipher$T13GcmWriteCipherGenerator$GcmWriteCipher.encrypt(SSLCipher.java:2019)
	... 17 more

The error is not propagated to the callee causing the thread of the callee to hang indefinitely, this is causing me large issues as a workaround I'm using the guava TimeLimiter but this is far from an ideal solution. I'm assuming "OkHttp TaskRunner" threads are shared between OkHttpClient instances as well as when the error occurs other threads hang on their next request.

My Environment is as follows:

• Windows 10
• Oracle JDK SE 11.0.9
• OkHttp 4.9.0

I don't see why it would, but If it makes a difference I am also using the following network interceptor

private Interceptor createNetworkInterceptorInterceptor() {
        return chain -> {
           try {
               HttpUrl url = chain.request().url();
               Request.Builder builder = chain.request().newBuilder();

               builder.removeHeader("User-Agent");
               builder.addHeader("User-Agent", useragent);

               builder.removeHeader("cookie");
               builder.addHeader("cookie", getCookiesAsHeader(url));

               Response response = chain.proceed(builder.build());
               Cookie.parseAll(url, response.headers()).forEach(cookie ->
                       addCookie(url, cookie));
               return response;
           } catch (Exception e){
               throw new IOException(e.getMessage());
           }
        };
    }

[swankjesse]

This looks a lot like a JDK race condition:
https://bugs.openjdk.java.net/browse/JDK-8206968

But what’s weird is that was fixed in JDK 11.0.2.

Could you please report to the JDK issue tracker?

[yschimke]

Closing out here - needs to be reported to JDK issue tracker.