To add to what I posted above, upgrades done in the recent past have not always been completed by the specified time. Sometimes it has taken several more hours for the systems to come back up.
You'd think Disney would have a mirror site to put up when they do maintenance. Put up the mirror site, do maintenance offline on the original site, then, after it's okay, put back the original site.
You'd think Disney would have a mirror site to put up when they do maintenance. Put up the mirror site, do maintenance offline on the original site, then, after it's okay, put back the original site.
Yes, for hardware and O/S maintainance, they should be running with active/active failover. You take 1 node down, perform maintenance while the other half continues to run.
But for entire system upgrades, you pretty much need an outage window..