I have switched to HE.com a few months ago after I've heard a few good things about them. Well, my server is totally unreachable for over 2 hours now, it had a 20%-50% packet loss for about 8 hours before that, and this is the second time it happened in the past three months. Their e-mails support is terrible, it usually takes them anywhere from 8 to 72 hours to reply, even if it's an urgent reboot request. I've also seen packet loss of 5%-10% from time to time during peak hours. And their SLA guarantee is laughable -- you get refunded for the actual network downtime, i.e. if the server is inaccessible for a full 24 hours, you only get 1/30th, or about 3.3% of your monthly fee back. I thought I would never need to use SLA guarantee since this company supposed to be one of the most reliable network providers with multiple redundant connections, otherwise I would've never signed up. Unfortunately I'm locked in a 12 month contract, so I'll have to suffer this company for quite some time more. What a mistake.
What's even more frustrating is that their main website http://he.com is still up and apparently has no network problems whatsoever. I.e. either they took care of it first, while their customers' servers are still offline, or they provided it with a much more redundant connection.
As bulletproof as their SLA is, considering past experience hosting cabinets upon cabinets of servers with them, we've never even considered their SLA. Period. Why? The datacenter has been near flawless for 6 years.
I really can't say enough about the place.
As much as this is turning into a horrible day for me, personally, I know that things will get ironed out and service will be restored to the stability and quickness we're used to.
I really can't have any other thoughts at this time.
We have dozens of people that want to slaughter us now, true.
However, considering how the HE datacenter has treated us, I have no fear that this was a hugely isolated event.
A single router failure can take their whole facility down??? Granted, I don't know much about routers and how to connect thousands of servers to internet, but I'm sure there must be a way to make it redundant enough not to depend on a single router.
YOU WANT TO KNOW WHY? -- 1 CISCO GSR ROUTER - there is no backup - no HSRP - no multihoming - so if they are DoSed..and thier OC blades blow up *ahem* its not going to work.
HE's network sucks ass. I've been with them forever, but now I've setup links at other networks so this crap wont happen again. Its BS, everytime I call...blah blah blah and blah blah. I'll send an CW/Exodus/Level3 CCIE out there and they'll fix the issue within a few minutes.
They were down since 11:00am, all the way until 7:00pm, what kind of reliability is that. WHat kind of refund are you going to get?
4 hours outta your day..hmm..and depending on how many megs you use.
At while you look at larger datyacenters they are willing to credit you the whole month
MSI - down for 12hours due to GlobalCenter's BS, free month credited for a T3.
You have a point, to some extent.
However the issue was far more advanced than your particular version.
Additionally, I know they have a handful of future redundancy plans in the works - it really isn't my place to discuss the details publicly. But they are definitely working on many of the issues you have brought out below, if those facilities aren't in place already.
They aren't nearly as primitive as you think.
You've had machines with them "forever" after all.
At any rate in a nutshell I'm just saying you've made the situation extremely simple as compared to the reality of what occurred today @ HE. It's not my place to go into further details.