Customers of cloud hoster VPS.net are having problems, after a routine SAN upgrade went wrong and forced a haphazard equipment migration in the company's Chicago data center. Several days ago, a SAN within the company's Chicago facility started to have intermittent iSCSI connectivity, so VPS.net replaced its networking stack. …
I don't personally have any servers with VPS.net, but I do understand simple redundancy: I have one server hosted with a company in the US, and one with a completely separate company in the UK. It may not be perfect, but at least I'm insulated from the "eggs in one basket" syndrome!
Until Cletus pulls up the fiber to your office while digging a hole for a septic tank and cuts off your route to both.
One word: Incompetence
"At the time of writing, VPS.net had not responded to various Register requests for more information about the root cause of its intermittent technical woes."
Root cause? See subject.
You asked about OnApp and storage in another article?
This is the company where OffApp was incubated.
VPS.net had not responded
I know I'm using the title out of context - but it basically sums up all the problems with Cloud and why every time the CEO mentions it, I tell him "no f*cking chance"
Its nothing against IAAS as a principle. Its about the incompetent Organ Grinders who are running it and the 5 layers of monkeys (support lines!) between you and the Grinders. Its impossible to get any reliable information about the problem and the ETA for a resolution!
Thanks, but no thanks :)
Paris.. cause shes gone down on me the same amount of times as my own SAN has, 0 times :(
Ok, if you're one of these customers the situation is surely bad right now. But to all those slating the 'cloud' - are you going to say that your on-premise equipment never had a hiccup?
No tech is going to give you 100% uptime and there will be times when end users will be dis-satisfied by the response of the IT personnel trying to resolve a problem. The advantage that cloud brings is that reliability and response should be a grade better than doing it yourself.
From experience I've seen all manner of in-house tech failures and I've read about plenty of cloud outages. I still feel cloud outages have affected me less than in-house ones. Lets get some perspective.
@JMilies - no, the on-premises equipment isn't necessarily any more reliable than the boxes in the cloud. The difference is in control -- I can see that the reason my SAN is down is because flames are coming out of the fan grille. I could grab a fire extinguisher, put it out, swap the power supply, and be back in business in 5 minutes. If it were a hosted box in the nebulous cloud, all I would know is "We are currently experiencing a thermal issue in a storage node. We have no ETA for a fix." And if it's VPS.net, you're lucky to get even that.
Ditto about the control, but not only that, you're taking away jobs from your community and sending them to who knows where. Add to that network congestion issues, political issues in the countries your servers are hosted in, the security of your data, the fact that if you lose connection to your "cloud" and depending on how many "cloud" services you are using (Office365), you could be looking at a COMPLETE work stoppage from ALL your employees, the list goes on. At least with local in-house storage YOU control the money spent, but once you hand your data over, you are at the mercy of their price increases, and if you don't like, what are you going to do move it all back to your in-house... oh wait, never-mind.
Hmmm, So two things stand out here...
2) Data unavailability when single node/card goes down.
This isn't even "Enterprise class" - there is no redundancy in this set up. It is multiple servers talking over a single LAN path using iSCSI to one or more "cheap" SANs. The SANs here don't seem to have controller node or network redundancy.
In a reasonable in-house Corp SAN setup I would expect to be able to pull a controller, replace the broken card, shove it back in - all without my Hosts even noticing (except maybe a performance drop, but certainly not a complete loss of data path).
So, the Cloud can be great, but it can also be as high (low) tech as a bedroom server and just as reliable.
Long time customer
I am a long time customer of VPS.net - They are great for small non-critical sites and development servers but I wouldn't again host anything crucial with them. The stability of the machines is fine about 95% of the time - but you cant really run a business with that 5% there. It was a couple of years ago when their SAN couldnt talk to its licensing server (a friend who sells Enterprise storage said they were using a cheap "DIY" SAN) and it then proceeded to trash all data on it - a lot of people lost their sites due to VPS.net recent changes in rules over snapshots.
At least once a month I will go into my control panel because my server is down and see it is powered down!