So I've created two minimal `main` files to run the offending test: 1. [RunScriptUsingBasicLogin](https://github.com/devcsrj/bug-reports/blob/master/jclouds-1047/src/main/java/io/devcsrj/report/jclouds/RunScriptUsingBasicLogin.java#L97) - attempts to cancel using the default identity+credentials. 2. [RunScriptUsingPrivateKey](https://github.com/devcsrj/bug-reports/blob/master/jclouds-1047/src/main/java/io/devcsrj/report/jclouds/RunScriptUsingPrivateKey.java#L97) - attempts to cancel after overriding with user `foo`, then using it plus private key.
Both files reproduce the issue. BUT > ..the weirdest part of this last test, is that, sometimes, it passes. I can now reproduce the *weirdest part*. When debugging any of the two, adding a breakpoint to [`ssh.connect()`]( https://github.com/jclouds/jclouds/blob/master/compute/src/main/java/org/jclouds/compute/callables/SudoAwareInitManager.java#L74), then stepping to `return runAction(action);` will succeed, in contrast to the [original exception](https://gist.github.com/devcsrj/9675b0c3bb04a133fb50/1211b7546f9b17b5d4edaab578092b2b45fb44ac#file-jclouds-profitbricks-live-tests-log-L321). Without doing this, the connection would fail. I wonder if this has something to do with the delay I present before stepping through that, or maybe synchronization issues? Continuing investigation. --- Reply to this email directly or view it on GitHub: https://github.com/jclouds/jclouds-labs/pull/224#issuecomment-165712742
