The Afghan Ops Crossroads

[Mike Bobbitt]

It'll be no surprise to anyone here that I've been an absentee admin for Afghan Ops in recent months. I've been slow to respond to reported problems, feature requests have tapered off sharply and heck - even April fools came and went with no clowns this year. Truth be told, I've been tied up with a growing list of other commitments. Unfortunately this means AO has taken a back seat, with no new content to draw people in to play. That, in turn has lead to a lack of MilPoints purchases - nobody wants to play a game that has all but dried up, so there has been no purchases since Sept 2012. In a predictable cycle, I've been trying to focus on projects that have a return on my time causing AO to slip further down my list. I realize at this point it's starting to sound like a plea for cash, but that's really not it so I'll get to the point.

One of the things I have been fighting recently is an insidious performance problem on the server. At seemingly random times, there's a resource spike and the server shuts down, sometimes coming back automatically, sometimes not. Early this morning, I traced that spike to AO.

Sure this is a problem, but it's amplified by the fact that over the last few months some of the other projects I've been working on have included paying customers who have their data and sites hosted on our server. Yesterday's outage hit one project particularly hard and caused ripple effects on several other services on their end. I'm now in the unfortunate situation where AO is impacting other paid projects I have on the go, and I have to make a decision.

Possible solutions include:

• Solve the problem. Yep, that's the obvious answer, but it's not as simple as that unfortunately. The problem is likely inefficient coding, poor database design, server settings that are slightly off spec or (likely) all of the above. Finding those, let alone fixing those is a big task that will take significant time and effort. As noted above, I don't really have a lot of spare cycles to throw at the problem and waiting means putting my other projects in jeopardy until I can get around to it.
• Trim back the game. Spend some time to find out where the inefficiencies are (E.G. equipment proficiency levels) and cut those out of the game. This would be difficult to do without tipping the balance, and may not ultimately solve the problem if I can't find the problem source.
• Get a bigger server. Certainly a consideration, but if the paid projects don't require me to bump up our hosting plan, it seems counter-intuitive to do it for an unpaid project. And there's nothing to say we won't immediately hit the same performance wall again. (See the bit about inefficient coding above. )
• Shut down AO. Not my top choice by a long stretch, but it resolves the performance issue immediately and doesn't put a funding source in jeopardy. I see this as a last resort only. In fact, I will fight this one to the end. AO is still my baby, even though I neglect it.

I'm still grappling with the next move, but I felt I owe an explanation to those who have supported the game and invested so much into it. So what next? I'm going to put some thought into it and hopefully I can find a solution that keeps things chugging along without much fanfare. I invite discussion and feedback, as always.


Cheers and thanks
Mike

 

[Mr.Neville]

  Thanks for the update Mike! As a recent AO addict I would be very sad if the game were to be shut down. I really hope you/the team behind AO can get the issues straightened out!
  Maybe it's time to do some more advertising about the game? I mean there is other games/apps on Facebook that have millions of users (Farmville, Diamond Dash, Scrabble etc.). Is there possibly a way to get several hundred more players, which will result in more people buying Milpoints? Especailly new players, it took me quite a while to accumulate even the small amount I have now.
  I'm not a big fan of micro-transactions inside games, but here's another idea. I see lots of games (mostly on my iPhone) that give you the ability to buy extra perks or abilities for limited time. Perhaps a system like that could be implemented to sell +10% Morale for 24 hours for $0.99 or similar bonuses for other stats.
  I also recognize that you probably don't have the time to implement the above suggestions, I just don't want to see AO be shut down, as it is a big part of my routine in the morning, and before bed 

Thanks for your time, I hope the problems work themselves out!

 

[Mike Bobbitt]

Thanks VIH, I appreciate the input. Some good ideas there for sure. My current plan is to see if I can allocate some time to ID the specific source of the problem and take it from there.

 

[FJAG]

Mike

My experience over the years when with dealing with administrating my office's computers has taught me that sometimes a cash outlay for new hardware is a lot cheaper in the long run than spending hours of analysis and tweaking of an existing system.

If AO is the performance hog and a bigger server isn't in the cards then maybe a second separate and lighter server for AO on its own is the answer.

Obviously I don't know what AO's requirements are or whether or not you would need to buy another OS to run it but, compared to your time, servers are relatively cheap (I've always been partial to the Dell PowerEdge line and they've got some sales on right now).

 

[Mike Bobbitt]

Thanks FJAG, it's a good thought. Although I didn't mention it in the list above, that was sort of what I had in mind. Possibly a second, dedicated database server. The downsides to that are more servers to manage and increased cost, but your point is valid: it may still be faster/easier than chasing a phantom performance problem.

 

[navymich]

Thank you for posting about this Mike.  It makes AO life more bearable knowing the reasoning behind some of the issues.  It was down for me last night too (about 2200 CST).

You have put a lot of work into this over the last few years, and it is great to hear that there are still new addicts around.  Although I am personally pulling pin in a short while, I do hope that you are able to find a way to carry on with it for the others.

 

[larry Strong]

I would come back if there was a new theme - all be it at a less addictive pace. I got tired of doing missions for the sake of doing missions........


Larry

 

[kratz]

I would come back if there was a new theme - all be it at a less addictive pace. I got tired of doing missions for the sake of doing missions........


Larry

Navy.ca "the Facebook game"
Missions include hunting ninjasnpierwannerbes, sousing out Walts, head banging to repetitive recruiting questions and dodging site member's dry wit. Bonus refill points for jumping over the rainbow rules and not reviving old threads. Milpoints will be deducted for MSN speak or not spell checking.  ;D

 

[Mike Bobbitt]

A quick - and positive - update. I've managed to find and fix a large part of the outage issue. While my inefficient coding still causes some server angst here and there, it's not bringing down the server like it used to. At this stage, and for the foreseeable future, there is no need for me to make any nasty cuts to the game.

Thanks for your patience and understanding all!

P.S. Larry, I seriously considered using the AO template for a new game about... well, let's just say it could have been called ZO. May still, if I ever get the time.