Maintenance length

[Taliriah]

I guess it boils down to how conservative vs "yolo" devs are with server updates.

Do they first make a full backup of the servers? Do they first make sure to cleanly uninstall the previous servers and any related data file? Do they also apply os and security patches on the servers? Do they update it piece meal and perform tests after each step before moving on, like maybe they apply any database update first and run a battery of tests to ensure everything is correct before installing new server software? Etc.

If they are really conservative they may also have QA teams perform a battery of tests on the updated servers before opening them to the public and making sure that all game systems still work, that matchmaking still works etc.

[Dragonlancer]

"The people who live in a golden age usually go around complaining how yellow everything looks."

[Shibi]

I guess it boils down to how conservative vs "yolo" devs are with server updates.

Do they first make a full backup of the servers? Do they first make sure to cleanly uninstall the previous servers and any related data file? Do they also apply os and security patches on the servers? Do they update it piece meal and perform tests after each step before moving on, like maybe they apply any database update first and run a battery of tests to ensure everything is correct before installing new server software? Etc.

If they are really conservative they may also have QA teams perform a battery of tests on the updated servers before opening them to the public and making sure that all game systems still work, that matchmaking still works etc.

I've been through some of those YOLO patches before in a previous existence. In one case a certain 3 letter company who ran the data-centre just decided one day, out of the blue, to apply an operating system patch to a live production server, without permission, planning, notification, backups, or anything. Database went down mid-commit. Major incident ensued. Of course, this is the same three letter company who ran a data-centre in Sydney in such a run down state that the roof started leaking water one day. All over the servers.

[VirusOnline]

In one case a certain 3 letter company who ran the data-centre just decided one day, out of the blue, to apply an operating system patch to a live production server, without permission, planning, notification, backups, or anything. Database went down mid-commit. Major incident ensued.

Ah yes, that would be the new junior software engineer.
Not that I would personally ever know. (and if ever, it's the QAs fault)


Also maintenance in this game is few between with the downside of longer maintenance when it becomes necessary.
Personally I'll take that over the once a week 12 hour maintenance some games have.