Downtime and Routing Issues

[Xepher]

So I'll start this off by saying "It's not my fault, please don't hit me!" :-) The datacenter hosting xepher.net had a backbone link go down around 2 or 3am CDT this morning, which cut off access to xepher.net for about 1/3 of the internet. That is, most people could still get through, but many couldn't, and it just depended on where you were, and what route your connection came in through. The DC issued an announcement about it, and say they should have it under control now, with only some minor slowdown remaining. I'm not exactly thrilled with their service on this, as they'd been having interrmitent problems of this type for nearly a month now, and each time they claimed to fix it, it came back up. I may be considering a new datacenter if they have many more issues like this, as my patience is wearing thing (and I'm sure that's true for the rest of you too.) Hopefully this one was major enough that they really did fix it and not just tweak something. Anyway, if you (or one of your readers/visitors) is still having trouble connecting, all we can do is wait. Sorry!

[Xepher]

Just got this from FDC (the datacenter)

This email is intended for customers who suffered service interruptions during last 2 days.

One of our core routers has been experiencing hardware malfunctions that has caused service outage or partial outage for some servers on a specific segment of our network.
The memory module on the router failed early morning on 7/16/07. We had a spare one however it turned that the spare had not the capacity to quickly replace it. Our network engineer had to make several adjustments in our routing topology to get affected network segment back online which has pro-longed the downtime. Meanwhile we had ordered extra spares which will be delivered in 12 hours from now.
Than around 1AM CST on 7/17/07 a fiber module has failed on the same router that has impacted same segment of customers and we had to make further modifications, reboot the router and reset the BGP.

This series of events suggest some deeper problem with the router in question and we`ll be working toward replacing it over next few days (most likely this weekend). Should the replacement require any downtime we`ll do everything possible to schedule a maintenance window during night/early morning hours on the weekend so the impact is minimal.

So it looks like it could have a bit of downtime while they actually replace the problem router. I'll give you any more details as/if I get them.

[fesworks]

Jason would be willing to help form a lynch mob and go after some of the people at this datacenter. I'd be willing to go to.

What's your contract with them? Are you stuck with them for a while?

forgive me if this is  too nosey

[Xepher]

No, it's not a problem... There's no contract or anything, it's just I haven't been able to find anything else nearly as cheap as this. I know "you get what you pay for" which is why I'm not TOO upset about the problems this place keeps having. I pay $169/month for "unlimited" bandwidth. That means I can use however much I can, and not get charged more, but in reality it's limited by the size of their outgoing connections, and shared bandwidth with the other servers there. Xepher.net transfers well over a terabyte of data each month (just for web pages, not including email, or uploads or anything else) and I haven't found anywhere else that does colocation with that kind of bandwidth for anywhere near that price. And since I haven't been "gainfully employeed" for nearly two years now, I have a hard time justifying $400-500/month for slightly more reliable service. I mean, yes, it's a pain when the DC goes down for 12 hours. But 12 hours out of 6 months? Is that 12 hours downtime worth the extra $1500 or so that a more reliable datacenter would cost? Still though, if anyone DOES know of a reasonably priced (and reliable) place, let me know! :-)