Slicehost Review – Excellent VPS Hosting Experience

We are currently in the process of building out a farm of upgradeable virtual servers that will comprise Break it Down blog and the rest of The Buzz Media’s blog network. RimuHosting has been a top-tier host for 3 years now for our member sites, but the beauty of RimuHosting and it’s well-deserved reputation as having some of the best customer service in the hosting game is really centered around dedicated servers or what I’ll call “committed” server setups. What I mean by that is you don’t have the instant control over the size or number of VPSs you are running like you would Amazon EC2, Slicehost or other virtual hosts that provide management interfaces *specifically* for resizing instances, launching more or shutting down some. For a budding company that really cannot anticipate it’s traffic more than a few months out at a time, going with a host that provided virtual servers that were resizable AND persistent was huge for us.

With that said, if you have more specific server needs and know you might want a few VPSs and a dedicated machine or something, I’d recommend RimuHosting any day of the week.

Amazon EC2? No thank you…

As for virtual/resizable hosting experiences, Amazon EC2 was the first thing we looked at. AWS certainly has the “virtual” thing down, but after pricing out an average month with a Small instance a few GBs of traffic and either an EBS store for our persistent database data OR some cut of the AWS MySQL Server our monthly expenditures just for that single small machine were climbing up to a few hundred dollars a month — we are already spending a few hundred a month at RimuHosting and getting a dedicated quad-core beast-machine for it. AWS hardly felt like a steal.

So we did some projection over time to see if AWS became cheaper as the sites grew, possibly making it worth it for us to stick with it — again, all the prices did was climb to the moon. While the whole “cloud computing” thing is attractive in that everything is nebulous, these service providers are printing gold (Google App Engine included in this statement as well) by charging you for every damn thing having to do with all the CPU, memory, disk space, bandwidth in, bandwidth out, API service calls, times you go to the bathroom during the day and how often you call your mother.

So we kept hunting…

1 Week with Slicehost

We got a suggestion from Marc Chung and Preston Lee over at Open Rain (now part of Red Pear) to try out Slicehost — they had been utilizing the easy-to-manage virtual hosting service for a few months and so far liked it.

I signed up for an account and a 1GB slice and was off to the races. It’s a standard virtual-linux-server setup, when you create your slice you can choose from different “sized” slices, there are 7 of them and their names pertain to how much memory they have, e.g. “256 slice” == 256mb of RAM, “512 slice” == 512mb of RAM and so on. There are also preset limits on disk space and bandwidth for each account — these slices are all sold as-is, no buffet style resizing of individual specs here.

Implied in the sizes are actually the priority that your slice gets on the host physical server under-load. The Slicehost racks are all filled with what I guess are Quad or 8-core 16GB memory servers. If you have their biggest slice (15.5 GB, they need 500mb for management software) then you get full control over the machine, you don’t have to share with anyone. If you have an 8GB slice, you are splitting the resources on that machine 2-ways with 1 other slice. If you have a 1GB slice, you get 1/16th of the resources under load and so on.

NOTE: These are resource allocations under load — if the machine has idle CPU and disk time, any individual slice is free to spike and claim it — so these time splits are really “worst case scenarios”. Best case, you are on a server with a bunch of lazy slices and get all the hardware to yourself

It’s pretty straight forward and one of the reasons why we liked it right out of the gate.

A side effect of this simplified approach of 7-choices is that everything in the front-end management system is based around images of your slices. You can quickly backup your slice, clone it to create a new web server (e.g. if you want to quickly setup duplicates for load-balancing) or even resize bigger or smaller without loosing state — this was one of the biggest reasons we went with Slicehost over EC2, we didn’t have to manage additional stateful storage with Slicehost as we did with EC2/EBS and it simplified our lives.

In addition to individual slice management, managing of additional slices (aka “servers”) was straight forward with the web-based management interface:

You can also add new slices (servers) any time you like and Slicehost supports every Linux distribution out there that you might want to make a server from:

Again, one of the nice features here showing through — that when you go to create a new slice, you can even choose the backup image of an existing slice that you would like to use to seed the new instance. Similar to the “clone” operation for an individual slice, but gives you a chance to resize-and-clone in a way.

The overall flexibility has allowed us to get a fairly robust layout of servers configured quickly and rolled out. Our migration isn’t done yet, but so far it’s been smooth.

Slicehost Support – Accessibility is Key

In the week we’ve been on Slicehost, we had a few hickups with 1 of our hosts not rebooting correctly and not coming back online. We noticed the issue about 10am on Saturday and before sending a support ticket, I hopped into the no-login/instant-on Slicehost Chat Support channel and asked if someone could help. Much to my happiness one of the support techs was sitting in the channel and took a look, said the server was acting squirrelly and said they would look into it. During the next hour I got 2 or 3 emails from Slicehost’s support system notifying me that there was an issue with the server, it was identified, what the solution was and that they were monitoring the server to make sure it didn’t act up again.

I didn’t have to do a thing… nice.

Later that evening, starting at 5pm and ending around 7pm I got more support emails from the system starting with SH identifying a faulty piece of hardware causing the slices on the physical server to act up and that they were doing an emergency replacement to avoid a complete loss. I continued to get update emails automatically from them until the situation was totally resolved — again, I didn’t have to do anything.

After the move was complete I noticed my slice was booting into a state with no network access and using the AJAX console from the manager app, I couldn’t figure out why. Again, I hopped into chat, a Slicehost tech was there that immediately looked into some routing hardware on their end and had to flush a cache for the slice to behave itself — there was some remaining cleanup post-move that they were wrapping up.

All in all, I have had absolutely 0 stress and sub-10-minute support turn around time from Slicehost thus far. They make themselves very accessible (web-chat or email) and everyone I’ve dealt with knows what is going on… I’m not getting routed through call centers or dealing with folks that ask me if my modem is plugged in or what the lights on my AOL look like.

Slicehost Customer Appreciation

I’ve outlined our experience with Slicehost customer support so far — I’m very happy with the accessibility and responsiveness of them so far. Hardware failures happen, I’m not loosing sleep over it, I’m just happy smart people were there to take care of it.

Much to my surprise I have another email from the Support team this morning (keep in mind all of this is taking place over the weekend) saying that I’ve been completely refunded for this month for my slice that was giving me hassle because they felt the downtime was not fair to me as a customer… that’s some much-appreciated customer management right there as far as I’m concerned.

I’ve been dealing with a week of failed DSL server from Qwest with no end in sight and never once has a customer rep offered to reimburse me for anything. I really don’t expect a response this positive and pro-consumer from any companies, so when it happens, I really feel like it deserves some serious praise.

Conclusion

Slicehost is good — real good. If you are looking for cloud-computing and the bare-metal-ness and cost of EC2 gives you diarrhea and you absolutely need the ability to resize machines easily on the fly to respond to demand, I’d strongly recommend giving Slicehost a go.

If you are looking for dedicated hosting and need a strong team behind it, give RimuHosting a call. Either way, both are excellent companies run by smart people that know what is going on — it really just boils down to what you need.