3 ms·
I’ve seen at least 3 variations of the problem you mention, where power failover caused protracted downtime requiring rush delivery of niche replacement hardwar
by acdha 7y ago
I’ve seen at least 3 variations of the problem you mention, where power failover caused protracted downtime requiring rush delivery of niche replacement hardware. (That last is big: I’ve seen 8-figure enterprise hardware spends down for a week because it requires flying someone in to fix it, whereas AWS/Google has 24x7 staffing along with redundancy).
Anyone thinking this doesn’t happen with private data centers is either very green or selectively excusing problems.