I want to know how upgrading a server is done? Is it as simple as replacing hardware in its place or does it require some tinkering/tweaking.
Can replacing parts lead to worse performance? I mean I've changed hardware in computers several times and it has always improved the computer. Is it the same with servers?
I want to know how upgrading a server is done? Is it as simple as replacing hardware in its place or does it require some tinkering/tweaking.
Can replacing parts lead to worse performance? I mean I've changed hardware in computers several times and it has always improved the computer. Is it the same with servers?
I cannot say how ZOS does it. The way that I am familiar with will build a second server environment, test it with live data, and then on The Day, backup the "live" environment and restore it to the new environment. Finally, test again, then do the final configuration so that the second server environment is the "live" server environment. Typically, this is done every couple/few years because it is not good to let server hardware get too old, or out of support from the manufacturer.
The reason to build a second server environment, rather than trying to swap out hardware, is to allow for testing. The new hardware is rarely identical to the old hardware, and that can introduce problems. By building a new system, it allows testing and tuning of configurations and can isolate hardware that is not performing properly or does not work well with others.
Again, no idea how ZOS does it.
I want to know how upgrading a server is done? Is it as simple as replacing hardware in its place or does it require some tinkering/tweaking.
Can replacing parts lead to worse performance? I mean I've changed hardware in computers several times and it has always improved the computer. Is it the same with servers?
I cannot say how ZOS does it. The way that I am familiar with will build a second server environment, test it with live data, and then on The Day, backup the "live" environment and restore it to the new environment. Finally, test again, then do the final configuration so that the second server environment is the "live" server environment. Typically, this is done every couple/few years because it is not good to let server hardware get too old, or out of support from the manufacturer.
The reason to build a second server environment, rather than trying to swap out hardware, is to allow for testing. The new hardware is rarely identical to the old hardware, and that can introduce problems. By building a new system, it allows testing and tuning of configurations and can isolate hardware that is not performing properly or does not work well with others.
Again, no idea how ZOS does it.
Thank you for the information. So it's not that simple as installing hardware and letting it run? Then when new better stuff comes out just replacing it?
How come this game is running almost smooth as butter during event/anniversary? How can someone fiddle with the existing installed hardware to "boost" it?
If you're on PC, ZOS did a server maintenance right before the event started. I've noticed a clear pattern of the server running great right after a restart, then slowly degrading over the next few weeks until the next restart.
I just logged off from a 3 hours Cyrodiil session and lag is starting to creep back in, and is worse right now than during the event, even if the campaign isn't totally pop-locked.
Could be, note that this should be very visible in Cyrodil as these instances never shuts down outside of maintenance.If you're on PC, ZOS did a server maintenance right before the event started. I've noticed a clear pattern of the server running great right after a restart, then slowly degrading over the next few weeks until the next restart.
I just logged off from a 3 hours Cyrodiil session and lag is starting to creep back in, and is worse right now than during the event, even if the campaign isn't totally pop-locked.
Sounds like a server code memory leak... ?
If you're on PC, ZOS did a server maintenance right before the event started. I've noticed a clear pattern of the server running great right after a restart, then slowly degrading over the next few weeks until the next restart.
I just logged off from a 3 hours Cyrodiil session and lag is starting to creep back in, and is worse right now than during the event, even if the campaign isn't totally pop-locked.
Sounds like a server code memory leak... ?
Haven't ZOS said they'll do a code rework/refresh/optimisation once all the server hardware has been upgraded?
Tenthirty2 wrote: »I do NOT envy their coding teams at all
It's definitely possible for scripts/programs of a smaller scale to run without memory leaks or errors, *however* when the code gets sufficiently complex and large, memory leaks and other types of bugs start to creep in.If you're on PC, ZOS did a server maintenance right before the event started. I've noticed a clear pattern of the server running great right after a restart, then slowly degrading over the next few weeks until the next restart.
I just logged off from a 3 hours Cyrodiil session and lag is starting to creep back in, and is worse right now than during the event, even if the campaign isn't totally pop-locked.
Sounds like a server code memory leak... ?
Very possible. My limited understanding of programming says that every action we take in the game requires a script to run, and it's basically impossible for every script to run and end 100% perfectly, leaving small bits of code left running in the background doing nothing useful. These bits accumulate over time until they build up to a point that they have a noticeable drag on server resources, which is corrected by restarting the program. I guess this is the infamous "memory leak".
Ever notice how if you don't reboot your cell phone occasionally, it starts to develop glitches? Same principle.
Tbf this is coming from a non-programmer, so feel free to doubt.