Maintenance Watch

[Aidan8et]

True, but it looks better if they just make us all wait a little longer so they don't need to roll backs or repeated hot fixes. They're taking the "Do it right the first time" approach. Well... I suppose this would "Do it right first time because the OTHER first time was just practice" lol

[NuadaAkali]

True, but it looks better if they just make us all wait a little longer so they don't need to roll backs or repeated hot fixes. They're taking the "Do it right the first time" approach. Well... I suppose this would "Do it right first time because the OTHER first time was just practice" lol

This right here. It looks a lot better on a company if they make one long maintenance period rather than having two or three more maintenance periods after the first one because they rushed themselves the first time. It's better to do something once and do it right rather than wasting everyone's time with random "Emergency" updates.

[Nabian]

Ahhh...I completely forgot about the servers. Being naive I would expect them to have two teams working on it simultaneously.

Here is also some food for thought. I've done very very limited coding (I got a C in java many years ago and decided to switch majors --about 10 years ago? idk anyways besides the point). Why wouldn't they have all the coding pre-prepared and simply "copy and paste" it over and reboot everything? Obviously it's not that simple but I guess it would make more sense to me that way.

Perhaps a live test server where they could have all the "down time" and patching nonsense and then for the real game just copy stuff over.

I guess it comes down to man power and them not having enough of it?

[Lurkios]

Ahhh...I completely forgot about the servers. Being naive I would expect them to have two teams working on it simultaneously.

Here is also some food for thought. I've done very very limited coding (I got a C in java many years ago and decided to switch majors --about 10 years ago? idk anyways besides the point). Why wouldn't they have all the coding pre-prepared and simply "copy and paste" it over and reboot everything? Obviously it's not that simple but I guess it would make more sense to me that way.

Perhaps a live test server where they could have all the "down time" and patching nonsense and then for the real game just copy stuff over.

I guess it comes down to man power and them not having enough of it?

It's not really a copy-paste operation. The changes involve running scripts, probably several dozen if not more against the rather large databases that make up the entirety of the worlds we play. Not only does all of the new data have to be introduced, but everything that was changed has to be updated and all the new bits integrated with the old. Honestly my expectation would be that the SE team is sitting there twiddling their thumbs, getting coffee, perusing the forums and occasionally checking to see if the little green spinning "operation in progress" animation has stopped spinning so they can move onto the next thing.

Once you introduce databases into the mix updates become a whole different game. For example, regarding your "update a test server and copy" idea, the test server would have a copy of the database frozen as it was when the copy was created - generally a live database cannot be updated, certainly not with sweeping changes such as are being introduced with 2.1. So while that database is updating, the still online production server with players crafting, gathering, fighting, downing titan, running coil, etc. etc. would be receiving thousands of updates per second - every item moved from an inventory slot to a retainer, every gathering node accessed, every exp earned recorded... And then the now updated test database is ready, the patch is applied... but all the data is several hours old. All those accomplishments do not exist in that copy of the database. Could you move them over? Probably. You'd likely have to significantly alter the data to fit into the new schema, and then... yep, update the old database with all the new data... During which time both servers would have to be offline, or you'd run into the exact same problem again. So instead of just updating the existing database they'd be making a copy, updating the copy, stopping the original, copying the data from the original (again), updating the copied data, applying the copied data, then applying the updated updated data to the production system.

We'd be waiting for a week.

[Aidan8et]

Ahhh...I completely forgot about the servers. Being naive I would expect them to have two teams working on it simultaneously.

Here is also some food for thought. I've done very very limited coding (I got a C in java many years ago and decided to switch majors --about 10 years ago? idk anyways besides the point). Why wouldn't they have all the coding pre-prepared and simply "copy and paste" it over and reboot everything? Obviously it's not that simple but I guess it would make more sense to me that way.

Perhaps a live test server where they could have all the "down time" and patching nonsense and then for the real game just copy stuff over.

I guess it comes down to man power and them not having enough of it?

There's so many things that can still go wrong even with a seemingly simple copy-paste job. Likely/hopefully they're going through each server individually to verify them. Also, the different servers are going to be in different geographical locations. Think about how hard it is to even get a large raid to work together & multiply it by the player base of all of FFXIV. Also, they've got to keep in mind the legacy vs non-legacy servers & what all changes might be needed between the 2.

[NuadaAkali]

Ahhh...I completely forgot about the servers. Being naive I would expect them to have two teams working on it simultaneously.

Here is also some food for thought. I've done very very limited coding (I got a C in java many years ago and decided to switch majors --about 10 years ago? idk anyways besides the point). Why wouldn't they have all the coding pre-prepared and simply "copy and paste" it over and reboot everything? Obviously it's not that simple but I guess it would make more sense to me that way.

Perhaps a live test server where they could have all the "down time" and patching nonsense and then for the real game just copy stuff over.

I guess it comes down to man power and them not having enough of it?

The way gaming companies work is they have different "teams" and resources for various projects, but they're all under the name of the company. SE is no exception. A team popped up with ideas for FFXIV ARR and that's all she wrote. The resources are company-provided, but only after a game has a certain amount of success does the company offer more of its resources to a certain team. It's all gauged by numbers and profit.

How does this answer any of your questions? Simple. The team simply doesn't have the resource available (and yes, probably not the manpower) to open up a PTS while they work on patches. You have to remember that servers do not come cheap and are often times easily broken if not managed properly. Another thing to take into account is that the servers would most likely have to -double- in number to do a request of this multitude.

[Raist]

Take into account that this also restarts your download. I've also restarted the process several times and it's been a no-go on speeding up. Plus, I just ran my connection speed on Ookla, and i'm getting a very speedy 0.53 Mbps (most likely contributed to the fact that i'm downloading it at the time I ran it)

eh... as always, mileage may vary. Just kicked it off for the laptop again, and it started at 1.43MB/sec... more than double the speed it was about 40 minutes ago. Estimated time then was right at 1.5 hrs, now it's saying about 45 minutes. And that's with my Roku streaming a channel ATM too, so a portion of my pipe is consumed streaming a movie. In the end it's taking about the same time frame as if I had just kept it going I guess, but I was able to tool around on my laptop for about a half-hour while it downloaded to the PC at near full bandwidth and if not for using my Roku right now, the lappy could probably be doing the same and it would have saved me time.

It almost always does that for me though... even with FFXI. Get a crappy speed, cancel, wait a few minutes and restart and it usually gets a better stream going and cuts my time considerably. Guess it's just luck of the draw as to whether you hit a busy cluster or not... so far, it's worked to my benefit most the time.

[NuadaAkali]

eh... as always, mileage may vary. Just kicked it off for the laptop again, and it started at 1.43MB/sec... more than double the speed it was about 40 minutes ago. Estimated time then was right at 1.5 hrs, now it's saying about 45 minutes. And that's with my Roku streaming a channel ATM too, so a portion of my pipe is consumed streaming a movie. In the end it's taking about the same time frame as if I had just kept it going I guess, but I was able to tool around on my laptop for about a half-hour while it downloaded to the PC at near full bandwidth and if not for using my Roku right now, the lappy could probably be doing the same and it would have saved me time.

It almost always does that for me though... even with FFXI. Get a crappy speed, cancel, wait a few minutes and restart and it usually gets a better stream going and cuts my time considerably. Guess it's just luck of the draw as to whether you hit a busy cluster or not... so far, it's worked to my benefit most the time.

Well, i'm in a good spot as any really. My work shift consists of me working from 4pm-1am, so I can basically let things sit and download and they'll be done when I get home. And I usually only sleep about 4-5 hours anyway, it's just how my body works. Gaming all morning, erry morning!

[Ganfal]

The way gaming companies work is they have different "teams" and resources for various projects, but they're all under the name of the company. SE is no exception. A team popped up with ideas for FFXIV ARR and that's all she wrote. The resources are company-provided, but only after a game has a certain amount of success does the company offer more of its resources to a certain team. It's all gauged by numbers and profit.

How does this answer any of your questions? Simple. The team simply doesn't have the resource available (and yes, probably not the manpower) to open up a PTS while they work on patches. You have to remember that servers do not come cheap and are often times easily broken if not managed properly. Another thing to take into account is that the servers would most likely have to -double- in number to do a request of this multitude.

Another thing no one else said here. It's pretty common that a content has been fully tested on the Test Server but doesn't work as expected on the Production Server. Some other times, there are content they applied on the last Test Server patch, but it wasn't applied to the final Production Server patch. Or, even worse, a hotfix that was already applied to the Production server but wasn't on the final patch, since it was applied after the Final patch was generated (for instance, can you imagine if the board bug came back and you started losing your precious Gil?). Even if everything went all fine, there are chances that, with all the data from the Production Servers, something go south. That's why the QA teams have to complete a checklist on every server. It takes a lot of time.

[Nabian]

Thank you all for the posts. Has been very informative.