We were causing their server more pain than it usually causes itself, so XiBase (the account, which included OTC) was suspended.
I was right there when it happened, so I immediately got onto the owner and the server admin from F5hosting. Allen, the owner, had no idea at the time that I asked him, so I went over to Timothy, their new system admin. He says XiBase was using 30% of server resources, and that it was entirely too much, that the average site used less than 1% (*cough*bullshit... *cough*
BULLSHIT*cough*when pigs fly!), and he recommended we switch to using dedicated servers. At the time, I hadn't known the precise cause.
Well, I asked him for some time to back up the site, and he unsuspended it. While I generated the backups and had a couple people download them, I asked him if we could at least leave a page up explaining what happened, with the boards disabled. He agreed, and allowed the site to remain unsuspended. The load on the server remained under 2.0 after that.
After Sam got home (I had already called and let him know all I knew about the situation so far), he got online and was talking to the host. They let us reactivate the boards and see about the load. Once it (all of it, including OTC) was reactivated, the load started jumping back up. Sam had a hunch, and told me to just deactivate OTC. Instantly, the load dropped down. So that is when we hypothesized that WBB, the software powering it, was the cause of the problem.
So in an MSN chat, MuFu, the OTC mods, Sam, and I were discussing alternative boards. Sam brought up the point of using phpBB, which we are using for The Screensaver. So MuFu offered to do a precisely implemented test on that forum which would simulate a lot of people using it and produce accurate results with detailed layouts of the loads attained, which basically means I stayed in the shell watching the server load like the geek that I am.
The server load went up past 6 or 7, so that wasn't acceptable. (Preferrably, loads under 5 are best)
Then he tried the same thing on the vBulletin install at xibase, and the load shot up to 45 almost instantly (with only 30 accesses!). So we concluded that the server was fucked, and that it had no future in our eyes no matter what board software we used.
Fast forward to approximately late last night.
Sam ordered an account at the new host.
http://www.hostrocket.com/products/shared.htm
We're on the Executive plan. 25 gigs of bandwidth, 450 megs of storage space, 20 MySQL databases, 20 parked domains, unlimited subdomains, unlimited FTP accounts, unlimited email accounts and forwarders.
Excellent service:
When Sam first ordered, by a stroke of bad luck, the server we were assigned to had run out of IP's, so we had a blank email containing no details about how to upload to our server.
Sam got a hold of them on their live chat and got it straightened out, and an email with an IP came at him soon thereafter.
So, then I submitted a support ticket asking for SSH (remote access to the server, basically like logging in directly on the server except it's over dialup so it's about as slow as molasses). They have a priority ticket system that guarantees 10 minute response times, but it costs $3 extra, so I didn't tick the Priority box. But still, the response came in under 15 minutes after I submitted the ticket. Looking good so far.
Excellent servers and connectivity:
Dual Tuallies 1.26 gig
4 gig of ram
RAID 1 protected Quantum Atlas V SCSI drives (10k RPM, Ultra 160 interface)
Multiple OC-12 connection
Methinks this is the right place to be, and el mucho propo's to samb0nium for making that possible.
Anyways, back to the story, I transferred the site over to them, and imported all the MySQL databases, retaining the WBB software for OTC since it wasn't the cause of the problem, then got some sleep.
I woke up this morning and talked to MuFu and asked him to test out the new server with an elaborate program that he uses to make simulated loads, and he went for 30 accesses with me watching the load, and the load never hit higher than 2.60.
I had him try it again, just in case it was just a fluke. Here is a log of me reporting the load numbers and him giving up when the load did not so much as flinch.
[Saturday 04/27 10:53:02] pAuL: right.. go!
[Saturday 04/27 10:53:16] fury: 0.97
[Saturday 04/27 10:53:22] fury: 1.13
[Saturday 04/27 10:53:27] fury: 1.19
[Saturday 04/27 10:53:32] fury: 1.18
[Saturday 04/27 10:53:37] fury: 1.16
[Saturday 04/27 10:53:42] fury: 1.23
[Saturday 04/27 10:53:49] fury: 1.21
[Saturday 04/27 10:53:54] fury: 1.19
[Saturday 04/27 10:53:54] pAuL: nah
[Saturday 04/27 10:53:57] pAuL: it's rock solid
Right at this moment, I have never ever seen the load higher than 2.60, and often it's even below 1.5
2.60 is a long shot from the record 97 that I witnessed a couple weeks ago on the old server, eh?
, just for the hell of it. 
@ Leslie
