|
a multi-day outage, basically a day before a major release...
this does not look like infrastructure to me (or we would have a more limited/selective outage).
this seems more like application-wide functional defects... like certain stuff either didn't get tested, or for some reason certain stuff is behaving completely differently in production than it did in test (load/timing-driven defects or whatever, which equates to failure to do a different sort of testing).
and now it looks like we are having patches fail and require yet more patches, and then those patches fail... and each new domino requires yet more downtime to analyze/fix/"test"/deploy. pretty standard result when your QA is ineffective and you have to (choose to?) rely on paying customers in your production environment to system-test for you.
this is odd to begin with, given the size of the WoW revenue stream (they should be able to manage production-scale QA at this point, with the tools to similate as large a load of whatever nature they want), and i have to believe at this point (4 years after rollout) that if continuous availability really was Blizzard's top priority, this would be mitigated by now (in a variety of ways: by better methodology, better architecture, over-engineered infrastructure, etc; maybe they think they are already doing all of these things?).
at least it is rare these days... probably just a matter of a few shortcuts taken at a critical (and bad) time.
|