Hacker News Reader

So what you do in that case, add a monochrome adapter to the PC and write to the screen memory. Or send [greatly abbreviated] trace out a serial port.

View on HN

In at least one instance of breaking my tools with my tools, the machines in question were in a data-center. Ordinarily one might be able to grab debugging output over an NC-SI link, but that only works if the breakage doesn't also unexpectedly crash your NIC firmware...

If you can’t touch it, you might not be able to debug it. Kudos though for testing^H^H^H^H^H^H^Hdebugging in production.

Case in point. Had a client in the late 80’s with an X.25 connection between their office and a Mainframe across town. It was accumulating hundreds of CRC errors per minute. Couldn’t fix remotely so I flew out there. Had a scope hooked up on both sides.

We loaded the connection in varying patterns for about an hour. ZERO CRC errors. Scratching our heads, we decided to punt. Then suddenly the CRC error count started climbing on my scope. I shout into the speakerphone, excitedly, what did you all just do? They’re like “nothing but unplug the scope from the patch panel”. “Please plug it back in”. CRC error count stopped.

Diagnosis: faulty patch panel.

When your job is to build infrastructure, "in a data-center" does not necessarily mean it's a production server :)

Reply on news.ycombinator.com