Like so many others, I am intrigued by the nature of Linode.com's services and by its obvious popularity both here and on Webhostingtalk. I do, however, have some reservations about signing up, primarily due to the fact that this company is apparently a one man show.

If Chris should wrap his motorcyle around a misplaced tree and end up either permanently or temporarily indisposed, what would happen to his hosting business(es)? While I do realize that many hosting companies out there are entirely operated and controlled by one person only, this is not really the environment that I'd prefer to host the commercial sites for which I'm responsible.

Also, while nearly every post that I've read commends Chris for his competence and /or professionalism, I'd really like to know a bit more about his training and experience. Is he a sixteen year old whiz kid or, possibly, a former systems administrator with years of supervisory experience behind him?

Thanks in advance for any info offered.

Skip

This is from memory, but if I reacall correctly chris worked for a dot com medical company before leaving to startup linode.com full time. If you search the IRC logs you may be able to track down more information on this.

Before linode.com he ran another hosting company based out of the now closed XO datacenter.

Chris should be able to fill in the gaps, and I would not like to guess at his age as I will probably be way off.

Adam

What happens if Chris is 'out of the picture'? (Chris, please don't let that happen! )

Well, it depends... how did he legally structure the company? I don't know, but if it was as a sole proprietorship, then the company will disappear if he is no longer able to assume its duties. With that goes the usual court receivership issues.

If that happens, chance of you getting your data back in a timely manner (or at all) might be a little on the remote or slow side.

A quick search of the SEC's EDGAR system didn't reveal any Linode filings, so it's most likely a private company... and chances are, a sole proprietorship.

For this, and other reasons, it is best to back up your data on a daily basis -- data that you cannot live without. I do that with my Linode. Mostly rsync+ssh for mail backups and the once-every-two-weeks backups of user-generated data (other than email), along with nightly backups of system configuration and log files.

If the original poster absolutely must have certainty about long-term commitments, it would probably be better for the original poster to pay the money to host his/her own machine in a co-located environment.

With that said, my needs are generally not such so critical that I couldn't live without a short-term transition should the unthinkable happen -- and that is partly why I find the pricing here 'just right'. Your mileage may vary, of course.

My impression of Chris? I would suspect he's "been around the block" more than a few times. Beyond that, I wouldn't even hope to correctly guess his age within 10 years.

This has been a concern since the very beginning with Linode.com. Chris is a good sysadmin but it is a one man show. Now that there are 40 or so hosts, and what must be hundreds, possibly over 1,000, accounts, I have to wonder when Linode will be big enough for a little redundancy in the manpower department. Do we have to wait until there are 100 Linode hosts? 5,000 users? When does Linode become big enough to be more than a one man show?

Well, there's a few issues I have yet to see mentioned. I'm not going to defend any one decision, but it's definately a requirement to add these points to the discussion.

The user base of Linode has signed up with Linode at a base asking price, with the knowledge of Linode's terms, conditions, and environment, willing to make the sacrifice of some added risk in exchange for a cheaper service. If hiring on a second expert at 50-60k+/year is going to require an increase in the monthly fees charged to the customers, this is changing the reasons for which most of the people here signed up with Linode. While some would be fine with this change, the majority wouldn't be willing.

Also, while I've come to learn that Chris's experience with linux and his knowledge of servers, networking, and computers in general is amazing, I'm not sure of his experience in business. If Linode is indeed a sole proprietorship, that would mean Chris has to start dealing with employment issues that he hasn't had to deal with before. It would be a move on his part from working in the technical field to working in business management. He may not be ready for that move, or may have other plans in mind. It is his company, and his decision.

There's one other thought I wanted to bring up. Chris may already have everything in place for Linode to continue operation in the event of an emergency.

Getting on with my personal thoughts, I am a customer that would still continue my account with the added fees, but that's not the reason I signed up for Linode, and it's definately not something I'm supporting. If I actually wanted/needed the saftey net of a second expert increasing my chances of being able to get a hold of Linode if something goes wrong, that would mean I have some money at stake and that I should technically not be using a VPS solution in the first place to be honest. I would move over to a dedicated system, and expect to be able to get a hold of a person that could help me on the phone 24/7.

I have a lot of respect for Chris and the way he's run Linode. Where ever he takes Linode in the future, I trust his judgement more than mine right now to be honest.

Is this any better than any *other* hosting company?

My last hosting company was down for several days for no especially good reason. They didn't care because they were "phasing out the brand". This was a large company with lots of people and a corporate strucutre and all that crap. Nobody wrapped their bike around a tree here.

No, the possibility of caker wrapping his bike around a tree doesn't bother me quite as much because linode's not the only game in town right now. Anybody can download the UML patches, set up machines, and write a management console. Or instead of using UML, use Xen or several other solutions. Virtualization is a *big thing* in the Linux server world right now, so there's a lot of mindshare built around it. And, remember, the number of folks on Linode and the other virtual providers is a perfect existence proof that there's a market for it.

Just keep backups and be prepared to migrate stuff over. That's pretty much the only way to play the game.

Like wirehead, I have been screwed by much bigger and more 'reliable looking' hosting companies than Linode. Worrying about caker as a single point of failure is addressing only part of the problem. You need to have a plan which caters for the worst case scenario, i.e. your hosting supplier undergoes total corporate meltdown and disappears overnight, deactivating your service without notice as he goes. This applies to any small or medium sized provider, not just one man bands.

If, like me, you host just personal stuff and 'freebies' on your Linode, then wirehead's recovery plan - rent a virtual server somewhere else, set it up and then restore your backed up data to it - will suffice, but you have to tolerate a downtime of several days.

tronic suggests using your own machine in a colocation environment but this leaves you vulnerable to hardware failure and to problems with your colocation provider. To ensure reliability you need to have two servers in different locations - expensive.

If I was hosting small quantities of mission critical data, I would seriously look at having virtual machines with two or even three UML based suppliers, replicating the data amongst them and using ZoneEdit's load balancing and failover facilities. Triple redundancy is how the big boys get their 99.99999% uptimes - UML means the little guys get to do it too.

_________________
/ Peter

If contingency planning is an issue, I have no doubts Chris would be able to find someone to jump in if he does happen to be incapacitated in some manner..

Thanks for your post and concerns. I'm not planning on wrapping my motorcycle around a tree anytime soon (more likely someone in a cage would cause an accident), but the fact remains that something could happen to me -- It is a risk, just as there are risks that something could happen to the network or the host equipment or a failure higher up the chain with the providers that I use. Reducing risks is part of the game for all of us, as I have providers that I need to make sure will be available as much as is reasonable.

I have a contingency plan in place with two others on board, with documents detailing what to do in the case of an emergency.

I do have plans to go through a hiring phase to meet the goals I've set for Linode.com, although at this time I can't give you anything more than an estimate of when that will happen. Right now, it's looking like sometime in Q2 of 2005. One of my largest justifications for such an infrastructure is to provide dedicated system administrators and customer service personnel. Other justifications are more mundane administration and executive roles.


31 year-old wiz-kid, a former systems administrator, and I do have years of supervisory experience behind me. My last job was at healthstream.com where I joined as a Senior Systems Programmer (employee number 50), a year later we went public, had over 240 employees, and I was promoted to Technical Lead where I managed a team of systems programmers working on HealthStream's core system (called the Learning Management System). The system was a platform for administering online continuing medical education (CMEs) courses to hundreds of thousands of medical professionals in hospitals across the US. I reported to the Senior Vice President for my division. He reported directly to the CEO (Bobby Frist, US Senator Frist's nephew).

Before that, I worked in NYC at a company called BizTV (now defunct) which was a Russian media company working on the then top-secret MTV for Russia. I don't think they ever did anything with it, but it was a neat job.

Prior to that, I taught Digital Media (back in the Multimedia/CD-ROM heydays) at FullSail.com, Center for Real World Education. I had roughly 100 Students a month, in a 4 week long course (full time).

My education includes studies in Mechanical Engineering at West Virginia University, a degree in Audio Engineering and a degree in Digital Media both from Full Sail (they hired me right after I graduated).

I hope that answers your questions, Skip.


You are correct.


This is a good idea regardless.


I don't really see Linode.com raising prices to build up the infrastructure of the company. That's a benefit of increased volume.


With my teaching and team-lead positions, I was directly responsible for the output and mentoring of those individuals. I've hired/fired, done reviews, etc. I'm sure when it's your own business on the line things change, but having employees isn't something I'm afraid of. I'd be a good boss/BEFH (Bastard Employer From Hell).


I agree.

Thanks all for your comments.

-Chris

Thanks to all who took the time and made the effort to reply, particularly Chris who replied with unusual candor. Lots of good stuff.

This is quite a forum!

Skip

I highly suggest hiring technically competent people, who do nice things like pointing Linux Mag reporters your way, and also happen to be members of linode .

Yeah, that's me :-p

_________________
Jay Faulkner
http://oldos.org