MySQL doesn't replicate/load balance worth a damn. If the system we're based on Oracle, you could easily do such a thing but doing so on the current environment would actually make reliability even worse.

I don't see it as a Decentralization, maybe because I don't depend on it (the public LS). There are many people into Minilogin, and they stay with the project. The developers are working for the 'public' version of the emulator, and I don't see where it would matter, just so long as there's an LS available. It just doesn't seem that important to me, what is important is the frustration people get when all this is down and now one can play or do anything else.
Anyway, it's just a solution to the Public LS - something needs to be done - if LS and this dns went black today, then we would have a big problem, eqemu would probably be dead for a while, as we are not prepared for it, and would have to take time out to re-organize.

I'm think they might have a cable connection - cable tends to get laggy during peak hours like in the evenings and on weekends (cable users have to share bandwidth with the TV viewers)

I would like to clarify, we don't want to decentralize the loginserver (look at the WoW emulators), but I think making it a distributed platform would help for redundancy (like a P2P network?), although you still need to find a way to share the account information between everyone running the loginservers.

The problem is majority of the lifeblood many servers trickles in from players looking for something they heard about on a website, which is almost always gone or not what they really wanted afterall. The login server makes it easy to go and try something else without any talent since the majority of our players do seem to struggle.

That considered, if we enhanced the web site to include a list of servers supporting mini-login list where you easily controlled what is up and down (a new hook or telnet) and join from there, this would provide the same affect and could be mirrored without security risk. This same application could also be built into a minipatcher that does nothing but simply get a list of servers and connect to the selected just as minilogin performs.

I'm a bit behind on my current project to jump in head first at the moment but I wouldn't mind spear heading such a task in the near future.

It seems pretty clear that the issues isn't with bandwidth. Otherwise it would eventually resolve on it's own when usage died down, which is not happening. My understanding is that they are having to actually reboot the entire server to resolve this. They are hosted on http://www.serverpronto.com/ and I have heard they have the lowest package available from there. Even the lowest package there should be well more than enough to handle what we need.

I am sure many of the old timers remember the last hosting before it was moved to this one. I think it was about 2 years ago. From everything I remember about the old hosting, we never had this issue on it. This issue started almost immediately after being moved to the new hosting. It then got increasingly worse and worse every month and more frequent until it got to the point that it is today where it can fail multiple times almost every week. That tells me that it is very likely that the problem is related directly to the host hardware or software/configuration. But, that means it should be resolvable one way or another. My guess is that the problem is with the configuration for max connections and/or idle session timeouts either with MySQL or something with the OS itself. I am definitely no guru in either of those, but I am sure I could help find a resolution to it if I knew more of the specifics.

I joined IRC last night (which I never do), and made mention of the issues and asked if Rogean or Doodman would be able to discuss it with me. Both of them are logged into the IRC and Secrets says that they do stop by there fairly often. They weren't on when I left the message, but I am trying to get something started and would like to speak with them personally to see what can be done if anything. Doodman won't know me, but I think Rogean will recognize who I am. But, either of them can at least check the forums and see I am a mod :P

Either way, I am glad there is a bit of an active discussion going on about this now. For a long time, I was just ignoring it, because it was only happening on occasion and would remain up nearly a month at a time between major outages. Some outages are acceptable, and since everything here is free to us, we are lucky to have it all in the first place. But, the Login Server has proven that it won't be getting any better, only worse. So, it is time to take some sort of action. I am sure we can all come up with a good solution to this though.

IMO, keeping the Login Server centralized is one of the most important things we should be keeping in mind. As Yeormom said, it is nearly impossible to keep track of new or old or retired independent servers. Some servers pop up and disappear so quick that it would be hard for a list to ever be anywhere near accurate, at least not for long. The centralized login server is nice and easy and is guaranteed to have an accurate list of available servers! WoW emulators may be popular, but their community is completely split up all over the place. By not having a central Login Server, it means that each server has to advertise like crazy to get new members. I considered trying to start up a WoW server, but the community I could find and the fact that there are so many different versions and no clear central site or login was just too much for me to want to deal with. I didn't like it at all. To make eqemulator more like that would be a mistake.

Minilogin does have it's place, but I don't think it is the solution we need. Minilogin should be for what it is already used for; a close group of people/friends. It shouldn't be used for public logins. Mostly due to the same reasons I mentioned above about keeping it centralized.

I know that nobody here knows me really, but I really feel that what has been said recently is a important point. I say this based on being fairly new to the EqEmu community.


I think that if this project is going to continue on in the direction it has been, its time to trust some new keepers.

I posted this on Trev's forums, and I continue to feel strongly about it.

I think given all that Trev has done to help get the SoF Emu started with all the opcode work I would like to hope that a bit of trust has been earned there as well. Lets hope once the site comes back online in earnest that something can be worked out. Fingers crossed :)


So lets give some support to the folks who are trying, and lets see if we can't re-unite the community on a great LS, and hopefully soon a new client :)


Thanks.

I sincerely hope you all get somewhere with this, If you go back through my posts, you'll see all what I tried. All what's mentioned here, I've also tried. All I got back was, frustration and got ignored.
I didn't post the suggestion for a new LS just to be a pain in the ass, or break up the project, I did it because after all the effort I put in to this, I decided it's better to look for another direction.
Since there is no other public LS, and this LS is probably on a collision coarse to disaster, I've already made sure I have something working if it does happen. MiniLogin will keep the Emulator running.
If you don't get started on a new LS in whatever way you want (Open source, Closed source, Semi, etc) - you may find you have no LS one day.
This topic is very old - in fact, I remember when we would all get banned for even suggesting another LS.
We're beating on a dead horse.