bloop.rifle.FailedToStartServerExitCodeException (perhaps due to confusion about which JDK to run?)

[bjornregnell]

Version(s)
Scala CLI 1.0.4
Scala 3.3.0

Describe the bug
When scala-cli run . it craches:

"bloop.rifle.FailedToStartServerExitCodeException: Server failed with exit code 1
  bloop.rifle.internal.Operations$.$anonfun$startServer$8(Operations.scala:263)
  bloop.rifle.BloopRifleLogger.$anonfun$runnable$1(BloopRifleLogger.scala:14)
  java.base@17.0.6/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:539)
  java.base@17.0.6/java.util.concurrent.FutureTask.runAndReset(FutureTask.java:305)
  java.base@17.0.6/java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:305)
  java.base@17.0.6/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1136)
  java.base@17.0.6/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:635)
  java.base@17.0.6/java.lang.Thread.run(Thread.java:833)
  com.oracle.svm.core.thread.PlatformThreads.threadStartRoutine(PlatformThreads.java:775)
  com.oracle.svm.core.windows.WindowsPlatformThreads.osThreadStartRoutine(WindowsPlatformThreads.java:178)"

To Reproduce

It was reproducible on the machine of a student of mine before we fixed it with this workaround:

scala-cli clean .
scala-cli run . --jvm 17

Prior to that it did not work with --jvm system and now I'm unsure if it can be reproduced.

A lot of thing was happening on the machine to try to fix it, including cs uninstall --all, nuking caches etc. But nothing helped until --jvm 17

It seems as if perhaps scala-cli picked up the system jvm but bloop couldn't, but I'm not sure.

Expected behaviour
It should not crash and bloop should use the same JDK as scala-cli (if that was the actual problem).

[bjornregnell]

Perhaps related to oyvindberg/bleep#355

[JD557]

I'm getting a similar issue on CI:

Redacted logs:

cd examples/release
ls | grep .scala | xargs scala-cli compile
[...]
shell: /usr/bin/bash -e {0}
env:
  CI: true
  JAVA_HOME: /opt/hostedtoolcache/Java_Temurin-Hotspot_jdk/8.0.382-5/x64
[...]
Error: Exception in thread "main" bloop.rifle.FailedToStartServerExitCodeException: Server failed with exit code 1
	at bloop.rifle.internal.Operations$.$anonfun$startServer$8(Operations.scala:263)
	at bloop.rifle.BloopRifleLogger.$anonfun$runnable$1(BloopRifleLogger.scala:14)
	at java.base@17.0.6/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:539)
	at java.base@17.0.6/java.util.concurrent.FutureTask.runAndReset(FutureTask.java:305)
	at java.base@17.0.6/java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:305)
	at java.base@17.0.6/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1136)
	at java.base@17.0.6/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:635)
	at java.base@17.0.6/java.lang.Thread.run(Thread.java:833)
	at com.oracle.svm.core.thread.PlatformThreads.threadStartRoutine(PlatformThreads.java:775)
	at com.oracle.svm.core.posix.thread.PosixPlatformThreads.pthreadStartRoutine(PosixPlatformThreads.java:203)
Error: Process completed with exit code 123.

So it seems that bloop is trying to run with Java 17, even though the Java home points to Java 8?

[SethTisue]

fwiw, this happens to me rather often on our Jenkins instance at scala-ci.typesafe.com — I have been sadly lax about reporting it or trying to troubleshoot it, sorry about that

here's an example log where it happened: https://scala-ci.typesafe.com/job/scala-2.13.x-jdk8-integrate-community-build/5801/console

our Jenkins workers routinely use different JDK versions ranging from 8 to 22, so it's quite plausible to me that it's related to running multiple JDKs on the same machine

I do not know how to force it to happen, and I wish I had some insight into conditions under which it tends to happen, but I don't

if there's some sort of troubleshooting I could do or diagnostic options I could enable to get more information about it, I'm happy to try to assist

also interested in workaround information — I guess I could try the --jvm 17 thing

[SethTisue]

Oh, I've been running with --bloop-jvm 17 for a long time now and that hasn't fixed it. Not sure whether to take that out

[tgodzik]

We'll take a look! The team is out until 4th though, so we will dig in then

[tgodzik]

I think the issue might happen if JAVA_HOME is set to a version lower than 17. We removed the support for running Bloop on lower versions, which is why this is breaking in some CIs. Previously, it might break from time to time as it would use less reliable sockets. At least I think that's the issue.

[bjornregnell]

Nice that you have a lead to investigate. Even if 17 is required I still think it should not crash with a strange error msg - at least it should be detected and explained to the user. But best of course if it can work regardless of if it is 17 or 11 or 8.

[Squaess]

I had the same issue. Following the comment above I looked at my JAVA_HOME env variable. It was pointing to java8. I switched to java17 and it started to work.

[yadavan88]

I have the same problem. I am using the JDK 11 and getting this error.

In JDK 17 it works fine, but since we are still on 11 in prod, I need to continue using 11.

[holmsander]

I have the same problem but running java 17. Has anyone found a fix that isn't updating to java 17?

[yadavan88]

@holmsander
For me the issue got resolved after upgrading scala-cli to 1.0.6. Did you try that?