When I posted my interview with Terry Myers about cloud hosting, I wasn’t really prepared for the backlash that would come in on that post. Quite a few people responded and a lot of them weren’t too happy. On top of that, there were several instances of downtime in both the AMS and several US clouds over the last week. As you might know, I’m an affiliate for VPS.net, so I get a bit better every time they get a new client through me, but I was seriously debating whether that was still a wise choice. In fact, some people were telling me that I should stop recommending them altogether.
Now, I’ve been with VPS.net for quite some time, from their very beginning in fact, and I’ve always liked the service and the people working there. I’m not one to easily leave and on top of that, having had issues at other hosting companies before, I know that things are bound to go wrong at some point. So, I decided to confront my contacts at VPS.net and “talk to the boss”. That boss is someone I hadn’t spoken to before, Rus Foster. Rus has recently (7 weeks ago) joined VPS.net as their new Managing Director, although he’s been in the VPS business for years, replacing Nick Nelson, who’s gone higher up in the UK2 Group.
As you can see, I asked him some, I think, rather tough questions. I was happy with the answers I got, and must say that the changes he’s talking about below on the AMS cloud have indeed had a very positive effect on my yoast.nl server.
Interview with Rus Foster, MD of VPS.net
Hi Rus, thanks for agreeing to this interview. My goal is to shed some light on what is currently causing all the issues around VPS.net. First of all, it seems as though every day, somewhere in the world, a VPS.net SAN dies. Some locations have more issues than others, but in general, it’s been quite bad lately. What’s causing that? And, more importantly, what will you do about it?
Start with the easy questions eh? Nothing like whats your favourite colour? :)
Seriously though yes there have been SAN issues on some of our newer clouds.
Some of the SANS in multiple locations have been showing instability issues caused by the software running on them. As such we are starting to undertake a very aggressive upgrade schedule to move the SAN software to the latest version. Coupled with this we are also upgrading the storage network within each cloud to increase I/O speeds.
We are however also looking to start using the enterprise class Nexenta storage solution on our new clouds which is engineered fundamentally different to our current SAN solutions. The upshot of this is that the system is much more resilient to failure meaning in the event of a failure there will be no noticeable effect on customers.
This will be going live on our next cloud but more about that later …..
When you say “very aggressive upgrade schedule”, what do you mean, exactly?
We have identified 6 clouds where we feel significant upgrades, or full migrations to new facilities are required. We are doing 1-2 clouds a week and aim to have everything completed within the next 6 weeks, if not sooner. This is a significant investment in our infrastructure to increase resiliency and performance.
To be fair, some people would say that it’s a significant investment to make it work the way it should… Which clouds are we talking about?
There were decisions made in the past that were right at the time. Those choices however no longer reflect customers expectations as cloud hosting is becoming more mainstream. We are making decisions now that prepare us for the future allowing easier service enhancements to be rolled out to all our customers new and old. We are targeting clouds in both the US and Europe. Some of the work, for example on our Amsterdam cloud, has been completed.
Could you say that maybe you’ve grown a bit too fast? (Which would be partly my fault, perhaps, sorry about that)
We learn from mistakes. There is nothing wrong with growing fast. There is something wrong with not managing that growth. I would say that is where as a company we found ourselves tripping over our own feet. My personal focus is split into both pushing the growth of the company, looking at exciting new markets and technologies but also its vitally important that we look after our current customers. This is reflected in us bringing back such old favourites as sales live chat but also focussing on how we can help existing customers. This include things such as the cloud upgrades as well as some exciting upgrades we have coming for all customers.
Apart from the SANs, another issue seems to have been the fact that you weren’t responding during those outages, either on tickets, or on Twitter, with the status blog being deafeningly silent too. What are you going to do about that?
Communications. Communications. Communications. Nothing travels as fast as bad news. We need to be quicker. We need to be more open. We are taking steps to do that. We are now proactively emailing customers in the event of an outage. We are making it the sole responsibility for the manager on duty to communicate with customers. The status blog and twitter are important fields of communication but they can’t give customised responses to each customers. Our support ticket system is in place for that.
I’ve seen you jump in on more and more tickets yourself, are you one of those managers on duty?
Its best to lead by example. I never ask anyone to do something that I wouldn’t be willing to do myself. When things get busy in an outage its really a matter of needing all hands to the pumps. Thats why you will see me replying to all sorts of support tickets from simple reboots to full implementations of load balancing.
I get the feeling, talking to some of the people who bought a VPS, that they actually shouldn’t have a VPS. They have two types of downtime, the one caused by themselves as they don’t really know their way around a web server, and the ones caused by SANs or other hardware failures at VPS.net. I think that for people who have a lot of those first issues, the new Cloud Hosting, or a somewhat faster WestHost package, both a bit more managed, would be better for them.
Do you agree? If people think that’s true for them, is there any way for you to help in that?
We all love power. As a petrol head I dream of having a TVR Sagaris on the drive but do I need the power? Of course not. Some people do however need that power. It’s all about what matches your individual needs. If people just have a few websites that need to “just work” then a VPS isn’t always the best choice and there are other things in our current, and upcoming, product range that could suit them better. If however people do require large amounts of customisation VPS are still a good way to go. If they do require that little bit of hand holding they can always get an On-Demand ticket for a one time issue or pop management on top for unlimited help.
If a customer wants to move to another product we can move a reasonable number of sites free of charge.
So you mentioned a new cloud… Tell me more.
Yes our Japan cloud is the first cloud to be running our new Nexenta storage system and OnApp 2.2. It will also be live next week. Coupled with this, as we know people might be a bit wary of the new solution, we are doing first month, free of charge, for customers new and old. There is a limit of 10 free nodes per customer.
What can customers do if they think they deserve more attention?
They can contact me directly at firstname.lastname@example.org.