We have an opportunity to move to a new colocation facility in California. There are a group of Facebook employees who share a rack at http://www.he.net and we've been invited to take 2 2U spots in that rack. While we've been with Puregig (now Highwinds) for BP's entire life, there are a number of reasons why this would make a lot of sense for us:
Which brings me to the next point -- it's time to talk about new server(s) again. I put the 's' in parentheses because it seems widely agreed (by the vBulletin developers, by our own sysadmin, by myself (though I am not *at all* a tech-type)) that we need to be on two boxes for a couple reasons. First, splitting the different needs of web and database on to different boxes makes for a significant performance increase. Second, redundancy is good. If we lose a drive or see ram or the motherboard fail on our current machine, then we're down for weeks and out money I just don't have. If we have two identical boxes each capable of, in a pinch, running everything we have -- then losing one for days or even weeks is not the end of the world.
So just like last time, I want to open up a dialog here on the forums and see if we can, as a community, consider our options and the appropriate path. This dialog last time produced a group-designed spec that really improved our load times from my original box, a member (Dryden) who was willing to build it, and of course the surge of funds to buy it. While we may have some purchase/financing options this time that don't require that we have 100% funding in place, these are all still relevant considerations as we stand on the edge of this potential move.
Another thing I'd like to consider is the possibility of someone here having contacts or relationships that can help put us in better boxes. While my history with Dell is well documented here on the forums, I'll use them as an example. If we were able to partner with someone like that, then we could also be putting a "powered by ..." link down at the bottom of every single page and tossing them a few ads here and there in exchange for the discount. The point being that if there are people out there capable of suggesting and/or putting in place a hardware sponsorship/partnership, then we're wide open to those types of suggestions.
Here are the most basic of specs that have been suggested to me.
We've been talking about new hardware for a couple seasons, the discussion has just been accelerated because of the colocation opportunity that's been presented to us. Now, before we commit to that move, we have to figure out what hardware we can and will put in place. I hope over the next couple of days (the group with the rack are looking to place someone in relatively short order, and we're at the top of the list) we can have some good discussion (particularly from our more tech and business-minded types) about the next step in BP's evolution.
This move (hardware aside) is going to make it possible to do more with BP in the future. Keep it better upgraded, involve the tech admins so that the work is better managed and the burden better distributed, introduce better content management so we're breaking things out from the forums that are better suited for static presentation (articles, torrents, images, etc.). More importantly, we're going to be far better equipped to handle heavier traffic during games and on big days like NLOID. There's no reason, with current available tech, for us ever to be seeing page loads over a second.
Good stuff ahead, we just have to sort out what we want and need to do to get there.
BP's growth and direction has always been a community matter, and this is no different. Hope to see a lot of people chime in. We need to figure out our buying power, and then figure out how much bang we can get for that buck.
- We'd have someone local to the machine. We have no one in Arizona anymore (the guy who used to do it for us moved to CA). Right now, if something as simple as one of our 6yo 15k rpm HDDs failed, we'd have to pay over 100/hr to have someone on-site diagnose and reinstall a new drive, and of course we'd be down while we ordered, shipped, installed, configured, and repopulated that drive. It's been a bit of a concern.
- Our monthly expenses would be 1/3 to 1/2 of what they are now. Savings of anywhere from $180-360 a month.
- We'd have more redundancy through access to another server in Cleveland for off-site backups and other needs.
- SysAdmin work and maintenance would be handled by our partner (saving us another $500-700 a year).
- As the result of the sale of Puregig (to Highwinds), the machine (owned by a friend of a friend) we use for off-site backups is leaving the facility and will no longer be available for that purpose.
Which brings me to the next point -- it's time to talk about new server(s) again. I put the 's' in parentheses because it seems widely agreed (by the vBulletin developers, by our own sysadmin, by myself (though I am not *at all* a tech-type)) that we need to be on two boxes for a couple reasons. First, splitting the different needs of web and database on to different boxes makes for a significant performance increase. Second, redundancy is good. If we lose a drive or see ram or the motherboard fail on our current machine, then we're down for weeks and out money I just don't have. If we have two identical boxes each capable of, in a pinch, running everything we have -- then losing one for days or even weeks is not the end of the world.
So just like last time, I want to open up a dialog here on the forums and see if we can, as a community, consider our options and the appropriate path. This dialog last time produced a group-designed spec that really improved our load times from my original box, a member (Dryden) who was willing to build it, and of course the surge of funds to buy it. While we may have some purchase/financing options this time that don't require that we have 100% funding in place, these are all still relevant considerations as we stand on the edge of this potential move.
Another thing I'd like to consider is the possibility of someone here having contacts or relationships that can help put us in better boxes. While my history with Dell is well documented here on the forums, I'll use them as an example. If we were able to partner with someone like that, then we could also be putting a "powered by ..." link down at the bottom of every single page and tossing them a few ads here and there in exchange for the discount. The point being that if there are people out there capable of suggesting and/or putting in place a hardware sponsorship/partnership, then we're wide open to those types of suggestions.
Here are the most basic of specs that have been suggested to me.
- 4-8 cores per box w/HT (vB's devs recommend Xeons; 56xx, etc.)
- >=8gb ram per
- 6 drives in RAID 10 (would love to consider SSDs x4, but cost prohibitive)
- Integrated management
- Two gigabit ports
- 1U-2U each
We've been talking about new hardware for a couple seasons, the discussion has just been accelerated because of the colocation opportunity that's been presented to us. Now, before we commit to that move, we have to figure out what hardware we can and will put in place. I hope over the next couple of days (the group with the rack are looking to place someone in relatively short order, and we're at the top of the list) we can have some good discussion (particularly from our more tech and business-minded types) about the next step in BP's evolution.
This move (hardware aside) is going to make it possible to do more with BP in the future. Keep it better upgraded, involve the tech admins so that the work is better managed and the burden better distributed, introduce better content management so we're breaking things out from the forums that are better suited for static presentation (articles, torrents, images, etc.). More importantly, we're going to be far better equipped to handle heavier traffic during games and on big days like NLOID. There's no reason, with current available tech, for us ever to be seeing page loads over a second.
Good stuff ahead, we just have to sort out what we want and need to do to get there.
BP's growth and direction has always been a community matter, and this is no different. Hope to see a lot of people chime in. We need to figure out our buying power, and then figure out how much bang we can get for that buck.
Last edited: