4 ms·
One odd omission from the article is why it took an hour to reset everything. Given that they have redundant power, I would imagine that they could remotely sel
by amluto 7y ago
One odd omission from the article is why it took an hour to reset everything. Given that they have redundant power, I would imagine that they could remotely select which power supply to use or at least to remotely reset the breakers.
Alternatively, a good enough automatic transfer switch should transfer the power supply rather than tripping if the primary supply goes out of acceptable parameters.
(This goes to one of basics of high availability: failures, when they happen, should be short.)
- avianlyric 7y agoMost power systems are built with a mix or clever automated systems that can self recover, along with backup dumb systems that can’t. This is done to ensure that if your smart stuff fails, you don’t end up seriously damaging something. Unfortunately the dumb stuff pretty much always has to be reset manually. Usually because the dumb stuff tripping means you’ve encountered a scenario that no one anticipated, you really want someone to double check that nothing got damaged before the safety tripped. It’s a little like a database starting up and discovering that part of its Write-Ahead-Log is corrupt. You really want a human to go in there and see what happened, rather than just ignoring the damage, plowing on and loosing data.
- gsnedders 7y ago> One odd omission from the article is why it took an hour to reset everything. Given that they have redundant power, I would imagine that they could remotely select which power supply to use or at least to remotely reset the breakers. No remote control of them, had to get people to each breaker. (This was mentioned on Twitter closer to the time by NR.)