OutOfMemoryError: unable to create new native thread #786

Activity

1. KevinLiu707

   added

   bug

   on Apr 30, 2025
2. KevinLiu707

   changed the title [-]OutOfMemoryError: unable toc reate new native thread[/-] [+]OutOfMemoryError: unable to create new native thread[/+] on May 7, 2025
3. ajayjapan commented on Jul 11, 2026

   

   ajayjapan

   on Jul 11, 2026

   More actions

   Hi @nkzawa @darrachequesne, thanks for maintaining this library.

   I may be wrong about this, but wanted to share some crash traces in case they're useful. I'm on engine.io-client 2.1.0 / socket.io-client 2.1.2 and hitting what looks like the same class of issue described here.

   I'm seeing two distinct OutOfMemoryError: pthread_create signatures:

   Crash 1 - EventThread running reconnect, Timer creation fails:

   at java.lang.Thread.nativeCreate
   at java.lang.Thread.start
   at java.util.Timer.<init>(Timer.java:168)
   at java.util.Timer.<init>(Timer.java:140)
   at io.socket.client.Manager.reconnect(Manager.java:491)
   at io.socket.client.Manager$7$1$1.call(Manager.java:513)
   at io.socket.client.Manager$1$3.call(Manager.java:263)
   at io.socket.emitter.Emitter.emit(Emitter.java:117)
   at io.socket.client.Manager$1$4.run(Manager.java:278)
   at io.socket.thread.EventThread$2.run(EventThread.java:80)
   at java.util.concurrent.ThreadPoolExecutor.runWorker
   at java.util.concurrent.ThreadPoolExecutor$Worker.run
   at java.lang.Thread.run

   Crash 2 - Timer fires, EventThread has no live worker:

   at java.lang.Thread.nativeCreate
   at java.lang.Thread.start
   at java.util.concurrent.ThreadPoolExecutor.addWorker
   at java.util.concurrent.ThreadPoolExecutor.execute
   at java.util.concurrent.Executors$DelegatedExecutorService.execute
   at io.socket.thread.EventThread.nextTick(EventThread.java:76)
   at io.socket.thread.EventThread.exec(EventThread.java:57)
   at io.socket.client.Manager$1$5.run(Manager.java:292)
   at java.util.TimerThread.mainLoop
   at java.util.TimerThread.run

   My hypothesis (and I could definitely be missing something) is that the counter/shutdown pattern in EventThread.nextTick causes the executor to shut down and null out whenever it goes idle. The next reconnect attempt creates a new Timer (new OS thread) whose callback calls nextTick, finds service == null, and spins up a fresh newSingleThreadExecutor (another new OS thread). Under memory pressure on mobile this exhausts the process thread limit.

   Happy to provide any additional context if it helps. Thanks!