Comment 4 for bug 2043114

Thanks for getting back to me!

1. The crash seems to happen in a large percentage of the runs of a particular test that does a lot of rapid SSH connections to the machine. Sadly, I don’t have an exact number, but 50% probably wouldn’t be an exaggeration. It was also happening in the setup part of the test harness when it was making a large amount of parallel connections, but we spread those out a bit.
2. The connections all happen from a single test machine. I could try to test if this is repeatable when they’re spread over multiple sources, but that’s not how our test suite works.

As for the debugging information, I will first try to reproduce the crash on a throwaway machine (VM shouldn’t make a difference, I think), as to avoid any potential leaks of host keys or other sensitive information.