Help debugging a slave going away

[Francis]

I'm doing some work on the ec2 plugin, testing some new changes before
committing them.

Sometimes I'm able to start the ec2 instance and it all works fine,
however more often I get the stack trace below.

When looking at the code this indicates that the slave agent has
unexpectedly gone away.

My question is, is there tracing or some other startup options I can
use with the slave to help me determine why it's failing? I can't see
any helpful options in hudson.remoting.Launcher, and of course
Wireshark does not help since it's SSL.

I'm running the current master for Jenkins (1.451), but the ec2-plugin
is compiled with version 1.399 (per the pom). I'm guessing this is OK
since we want the ec2-plugin to work with any old version after 399.

Any ideas of how to proceed?

-----------------

Connecting to ec2-50-18-232-57.us-west-1.compute.amazonaws.com on port
22.
Connected via SSH.
Authenticating as ec2-user
Executing init script
init script1
Verifying that java exists
java full version "1.6.0_20-b20"
Copying slave.jar
Launching slave agent - with options: -Xmx1000
ERROR: unexpected stream termination
java.io.EOFException: unexpected stream termination
at hudson.remoting.Channel.<init>(Channel.java:375)
at hudson.remoting.Channel.<init>(Channel.java:338)
at hudson.remoting.Channel.<init>(Channel.java:305)
at hudson.slaves.SlaveComputer.setChannel(SlaveComputer.java:307)
at hudson.plugins.ec2.ssh.EC2UnixLauncher.launch(EC2UnixLauncher.java:
131)
at
hudson.plugins.ec2.EC2ComputerLauncher.launch(EC2ComputerLauncher.java:
42)
at hudson.slaves.SlaveComputer$1.call(SlaveComputer.java:194)
at java.util.concurrent.FutureTask$Sync.innerRun(FutureTask.java:303)
at java.util.concurrent.FutureTask.run(FutureTask.java:138)
at java.util.concurrent.ThreadPoolExecutor
$Worker.runTask(ThreadPoolExecutor.java:886)
at java.util.concurrent.ThreadPoolExecutor
$Worker.run(ThreadPoolExecutor.java:908)
at java.lang.Thread.run(Thread.java:662)

[Ullrich Hafner]

Can't you just connect the slave to the debugger and register an
exception breakpoint? Or is this in case of your ec2 setup not possible?

Ulli

[Francis Upton]

Duh, thanks Ulli, of course I could. I need to get more sleep. Attaching to the debugger worked fine. Thanks!