SLA and refund

I'm not an employee. This is just my understanding.

The SLA isn't worth much, if it was down for a full day, you'd be looking to get back 1/30th of $20 (for a 360).

It's not automatic, either, you have to file a support ticket requesting your credit of 67 cents.

And who's fault is a data centre going down, anyway? Where's the demarcation?

So, I personally wouldn't count on the SLA if uptime was valuable to me, I'd be looking at some sort of fail over. Which can get pretty complicated when you try and diversify against a whole data centre going down, the linode website going down and the dns going down.

@backslash:

1. What is exactly your SLA promise and refund policy ?
From the TOS:

@Linode Terms of Service:

Linode.com provides a 99.99% uptime guarantee on all Linode hardware, and on network connectivity. In any given month, if your Linode is down for more than 0.01%, you may request a pro-rated credit for the down-time.
Emphasis added. Feel free to file a ticket if you have any questions.

@backslash:

2. One hour's down time is not short. But we can only see "this issue has been resolved". Can you provide more information and do something to promise you will learn from this problem ?
From the status posting on the issue:

@Linode Status Updates:

8:39am (EDT): From the Dallas NOC: We possibly have a DRS router down. They are working diligently on resolving the issue and we appreciate your patience.

Networking equipment can and does fail, and this is not a problem unique to Linode. Not to sound like I'm transferring blame, but this wasn't our equipment, either, so we couldn't troubleshoot it at all. We do not take downtime lightly, and will be investigating with the folks in Dallas as to why this happened and what can be done to prevent it in the future.

The Internet is a volatile place – these outages are rare, but happen to every provider. People that use the Internet just have to plan for them, and I'm sorry that you ran into trouble. I truly am.

The network in Dallas failed again today. It's the 3rd time this month.

I regret that I didn't move out of Dallas DC after the 2nd fail.

Yes I know "Networking equipment can and does fail" ,

but the stability in Dallas DC is really bad this month.

Do you have any reason to keep me stay in Dallas?

We would like to extend our sincere apologies to any customers affected by today's Dallas networking issues. We understand your frustration with events like this, and would like to convey our assurances that we work diligently with datacenter personnel to identify the causes of such issues and restore full service as quickly as possible. Your patience and understanding is greatly appreciated; please feel free to open a support ticket if you have any further questions. Thank you.

@pparadis:

We would like to extend our sincere apologies to any customers affected by today's Dallas networking issues…

Is it possible to have linode.com failover to another datacenter? I'm not in the Dallas datacenter, I'm in Fremont, but I couldn't get to my Linode Manager this morning and that bothers me.

@devjonfos:

@pparadis:

We would like to extend our sincere apologies to any customers affected by today's Dallas networking issues…

Is it possible to have linode.com failover to another datacenter? I'm not in the Dallas datacenter, I'm in Fremont, but I couldn't get to my Linode Manager this morning and that bothers me.

Agreed. So far as I know cross-data center ip failover isn't an option but they could set up multiple DNS records and DNS failover.

I don't work for Linode, but I was affected by the Dallas outage.

DNS failover wouldn't have done you any good, as your actual Linode itself was not accessible from the outside world due to the networking issues. They haven't updated yet as to what exactly the issue was, however like was mentioned before these problems do happen sometimes. The Planet is a world-class datacenter, and they do try their best to keep these problems from occurring (obviously).

An hour's outage isn't that bad, especially since it has not happened enough that I haven't noticed any of the previous Dallas outages (and I'm on my VPS from sunup to sundown most days). If your data is important enough to require absolute 100% uptime, what kind of failover have you setup yourself for your data? If you are relying on a single server to be available 24/7/365 without a single issue, ever, I'm sorry to say that you're doing it wrong.

"DNS failover wouldn't have done you any good, as your actual Linode itself was not accessible from the outside world due to the networking issues."

I believe the request was for a redundant version of the Linode website and dashboard. I don't know if that is practical, though it is certainly not trivial.

As far as providing redundancy for one's own 'node, that's certainly true. You can put a node in a second DC and use one of several mechanisms for sync and failover. Many things can make a 'node inaccessible and actually most of them are completely outside Linode's control, and your's.

@Yaakov:

…

I believe the request was for a redundant version of the Linode website and dashboard. I don't know if that is practical, though it is certainly not trivial…

Yeah, that's my concern: Failover to another datacenter for the Linode Manager, whose outage affects us all no matter what datacenter our Linodes are in.

@devjonfos:

@Yaakov:

…

I believe the request was for a redundant version of the Linode website and dashboard. I don't know if that is practical, though it is certainly not trivial…

Yeah, that's my concern: Failover to another datacenter for the Linode Manager, whose outage affects us all no matter what datacenter our Linodes are in.

That is a good point. I had not thought about the Linode control panel not being accessible to customers that are not hosting in the Dallas DC :)