How does anyone diagnose bugs on AWS EC2?

[Daniel Farina]

Hello,

I've recently been suffering some problems with sudden crashes that I
believe to be in Linux after upgrading from Ubuntu Precise to Ubuntu
Trusty.

It so happens there are enough computers to get a statistical sense of
the increase in such crashes, and it is about tenfold the rate.

The problem is that no messages I can easily attain are emitted:
crashdump doesn't work on EC2 (which is Xen-based, and fact that
crashdump doesn't work is also documented elsewhere on the Internet as
an upstream issue), and the logs possess nothing except maybe some nul
bytes where I expect unflushed data to be.

While I know LKML is not particularly interested in bug reports such
an a kernel, I'm having trouble finding references on how one goes
about designing a good bug report given the constraints I have (in
particular, no kdump, no logs).  I also was not able to find
definitive references about how to gather data for such cases.

Can anyone dispense advice about the routine way to deal with such
bugs so I can assemble a a decent report given the constraints I have
in post-mortem crash data?

Thanks in advance.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/