No. 0543 min read
huh. What is this string, exactly?
A timestamp, a cron line, an exit code, an invisible Unicode character.
Your Internet works, your email doesn't, and you haven't touched a damned thing on your computer. Your ISP[1]Ref. 1ISP en.wikipedia.org can manufacture that problem all by itself.
At Advertisnet I built redundant user databases for unified RADIUS[2]Ref. 2RADIUS rfc-editor.org, email, FTP[3]Ref. 3FTP rfc-editor.org, and web authentication.
When I said "One user database," I meant one coherent account wherever the customer used it. We could have several physical boxes, I just wanted them to agree about the account.
Redundant user databases help availability: when one component fails, you have somewhere else to get account information. You also have copies that can disagree. I had to worry about that too.
The replication mechanism matters, but first tell me where changes originate and which state wins after an interruption. How do we know a recovered copy is safe to use? I still need those answers even with redundancy in place.
Meanwhile the customer has already checked the password, and they're waiting on us.
