I would like to respectfully request that in the future, the Dev/Engineering teams only do one major project at a time. For the current downtime (which is perilously close to eclipsing the 2 day mark), we apparently have: Server Merges, and Live patching some rather major changes, and hardware upgrades, and who knows what else they are not telling us. I am unsure if anything else is or could be added to that. Do one of these at a time, work out the bugs before moving on to the next one. And test the changes before going live. That is what the test server is for. It obviously runs on separate hardware, as it is running right now. So take the backups you have of the servers that are to be merged, and load the merge on to that hardware, then work out the kinks. Roll it out to live, and fix the bugs that will inevitably been missed (no amount of testing catches everything, but anything that causes 48 hours of downtime will be caught on test). Same for hardware upgrades...test them on test first, then do the same for live servers. This is already supposed to be the process for the balance/bugfix/update portions of the downtime, but with the extended length this time, it gives the impression (however unwarranted) that it was not tested. Sorry if this has been posted before, I did not see this specific advice, even though it seems to me to be rather basic.
