This animated GIF is always perfect…
Question is whether or not the 10 -> 14 hour maintenance will be extended further. Of course, it was just a minor issue that caused the 4 hour extended downtime. I guess I will find out when I wake up 8 hours from now.
Seriously, we used to do test rollouts well before any kind of actual transition in order to catch any possible snafus because we couldn’t have anything more than 1 hour of downtime whenever doing any sort infrastructure maintenance; just regular system patches needed to be tested against mission critical applications, network upgrades meant testing a prototype design first, and then doing a lot of onsite preparation first including installing the new equipment, identifying and testing open fiber pairs, connecting and testing connectivity of this temporary network, and then coordinating the actual cut over – there were more granular details that I left out; the gist is that there’s a lot of testing that was entailed and we normally did these kind of major transitions within a 30 minute timeframe. And some of these transitions were non-trivial (like actual IP routing transitions from RIP to OSPF, static to dynamic DNS, moving from one network topology like FDDI to a collapse switched layer 3 network, etc) when the network changes needed to be done piecemeal in sections.
The above includes all the usual backing up and verification (standard operating procedure). Blizzard’s issues when it comes to rolling out major patches, doesn’t surprise me though. GIGO when it comes to their craptastically outdated/outmoded architecture.
