GoDaddy, Media Temple, and the Horrible World of Web Hosting
marco.org
marco.org
They're generally more expensive than the competition, but you get what you pay for, you know? I'm sitting here trying to think of a time when Rackspace has ever let me down, and I can't. Being able to have that kind of confidence in your hosting environment is nice.
Marco is correct that shared hosting is a disaster area, so much so that Rackspace doesn't really compete there, so I'm always hesitant when people ask me to recommend a shared host. I generally end up recommending Dreamhost too; it's not great, but it's better than what you'd get for the same money anywhere else.
It's notably worse than a number of other commodity experiences I've had.
Dreamhost aggressively oversells. They're hardly unique in this, but they admit and embrace it like nobody else I've seen.
Because of this, DreamHost accounts have two sets of rules: the ones they sell you on, and the other ones they're counting on you adhering to in order for everything to actually work. If you break the unwritten rules (even if you haven't broken the written ones), they will shut you down (sometimes without notice) and accept your cranky departure if you're unhappy about it.
Or maybe before then they'll have a severe service outage that causes you grief, make a funny blog post about it, and despite your amusement, you'll get the sense that something wasn't really addressed and leave.
If I had to recommend any shared hosting I've been on, it'd be Hurricane Electric. Over the decade I kept a small account there, my experience was the opposite of Dreamhost: they may have given less for the price, but they stood totally behind it (and a little further) and were always up.
Just two examples from my painful membership:
1. I transferred a domain back into Dreamhost. Due to a bug in their system, it was immediately dropped and entered the redemption period, and asked me for a large redemption fee. Support was wholly unable to acknowledge any problem on their part, and only resolved it when I posted the issue on Webhostingtalk.com
2. After requesting a refund onto my credit card after another hideous incident, rather than crediting my card, they debited it. Again, support were completely unable to acknowledge any problem on their part. Instead, when I asked the reason why my card was debited, they said I was charged for "services" - but were unable to itemise the services I had been charged for when challenged. They only refunded the money as "courtesy" rather than an acknowledgement on their part.
Dreadful.
If your storage requirements are small, I strongly recommend medialayer.com. For a budget option, I recommend asmallorange.com
We have to remember that shared hosting is not for heavy or critical websites. There are always some unwritten, but well-intentioned rules attached to unlimited hosting plans. As you would imagine, the server performance on any shared hosting depends on you and your neighbors adhering to them.
I'll agree with you, shutting down a customer without notice is not acceptable. Maybe if a customer is actively trying to scam the hosting provider by violating the terms... but I think even in that case, customer should be given a 24 hour eviction notice so that they have time to transition elsewhere.
If you look carefully, you can find providers of shared hosting who will be honest about what they're selling you in terms of capacity and that you can rely on.
DreamHost is not one of those businesses.
> There are always some unwritten, but well-intentioned rules attached to unlimited hosting plans.
Something something the road to hell something...
As you said though, their goofy attitude toward outages and such isn't good. I like the transparency of their status site, but when shit goes wrong I don't want a cat-meme reference.
I use dreamhost as a sort of project sandbox, and for that it works pretty well. If you need a place to do a one-click install of wordpress or some other project, or just a shell to bypass some firewall, upload some large files for a while, it's great.
I would never, ever put anything where I needed real uptime on Dreamhost, because performance varies wildly. If a project was getting serious, I'd use Linode, Rackspace, or Amazon.
I also recommend to everyone that they separate out their domain name purchases from their hosting provider.
Domain name purchases
DNS hosting
Email hosting
Web hosting
That way any of the layers in the stack can be replace with sufficient planning and if, say, email is down then the web hosting should still work.We've used Rackspace Cloud and Linode for several years and found Rackspace Cloud performance to be significantly worse than Linode for CPU and IO. Like, not even close. Prices are about the same. Support experiences are also similar (quick, competent responses from both sides).
The only reason we keep the Rackspace stuff around is because we want to have a relatively wide separation between our production webservices and our website / monitoring so that if one goes down that we can announce / detect that from a wholly separate network.
We're currently considering a migration from Rackspace to Linode for this very reason so we'll temporarily be running a complete version of our infrastructure on both and replaying traffic on both.
Once we've done this I'll try and post some detailed benchmarks.
All those unscalable things they did for you to get and keep your business? They don't last forever.
So you decide to use a bigger company. And it's not the same, but it's generally more reliable. And some of those companies do a much better job than others. Some are actually pretty good. You like RackSpace, there are a few companies out there that people like as much if not more too. A few even compete in the shared hosting space.
I track most of these big companies and dreamhost is middle of the pack. Better than GoDaddy but below some other brands.
Take a look for yourself: http://reviewsignal.com/webhosting/compare
Maybe next time someone asks you about it, you can let them figure out what they need. No company is perfect, but some are definitely better liked by their customers than others.
They also have a startup program where they'll subsidize $1000 of hosting for six months: http://rackspacestartups.com/
Rackspace cloud sites, terribly unpredictable performance with frequently obscene (20s+) latency, then categoric denial of actual measured issues for months... followed by begrudging admission, a swap of networking hardware, and a respite for a few months... until it all starts again. This pattern is much too frequent in their forums. I suspect they assume the customer's code is always at fault, so won't really look into their network layers unless coerced. A real shame.
I don't agree at all that for many website hosting customers the process is "easy".
A typical web hosting customer is not tech saavy they either have it being handled by their "tech guy" or they can't even remember how their files got onto the server in the first place with their static site and sometimes they don't even know who is hosting their site [1].
[1] Source: We're a registrar and we get the calls and emails of confused customers who have no clue where they are hosted. They don't even know enough to look at the whois and see the dns to give them a hint. Actually you'd be suprised how many times someone will access our whois and think we are their registrar.
We're working on solving this for static sites with BitBalloon (https://www.bitballoon.com).
Both by giving people an instant way to get their site online, but also by making it dead easy to do typical tasks like adding Analytics or getting forms to work.
I think this is partly because of providers like godaddy who have massive up-sell attempts and navigation misdirection to confuse people into purchasing all their value-add services when all you wanted was to purchase a domain or use their DNS services. It's the rare provider who just wants to do one thing well rather than be horizontally integrated and do a bunch of things half-assed. Unfortunately, people prefer one-stop shopping, but the integration doesn't make it easy to differentiate the offerings as services that don't need to be provided by a single provider. Like you can use godaddy as a registrar, but the ability to not use godaddy as the DNS provider too is buried and hidden. They purposely conflate what can be two separate services.
EDIT: I do remember canceling some godaddy service at one time and the process was super easy. Called and said I want to cancel service, wait time on the phone was low, and there was no questions asked.
The less trouble a user can cause, the better -- particularly at that scale. For more advanced users it merely invites frustration, however.
I assume people can also drag a .html etc or a .jpg and it will replace an existing file after verifying (and/or rename)?
Good luck!
So far updates are done by just dragging the whole folder unto the site dashboard again. We'll figure out what files have changes and just upload those.
There's also a version history so people can always roll back changes or see how the site looked like before.
$5 a month is cheap, too cheap in fact.
BitBalloon also makes forms work without having to do any brackens programming (or fitting in an external form service and try to get it to look like part of your site).
There's lots more timesavers included as well, so if your time is with anything to you, those $60 for a year of hosting will pay for themselves.
My father-in-law works for a small business that launched a static site earlier this year. They pay $80/month for "maintenance" on that static site.
I just launched a static site last night. Total cost? $10 / year.
If this was an easy thing for most consumers then they wouldn't be paying $950 / year more than I am.
But you do have to be willing to pay $20 or maybe even $99 for your completely business-critical, complex infrastructure, rather than merely $2.99 pm.
You can't just start an airline overnight. It requires massive capital upfront. A hosting company simply requires a server.
It's a) getting customers but actually even more important b) supporting those customers. c) A good online site to take care of the majority of the issues that come up. d) "c" helps keep down the "rtfm" type calls.
The "iron" and everything else is fairly easy.
So a hosting company is really people if you can solve the people problem you can definitely make money in hosting. Despite what Marco says the average retail hosting customer is not looking to switch their hosting if they get "good enough" service. Most don't even know that there site is down. (They do know when their email is not working of course because that they are constantly checking).
Life is too short to follow the race to the bottom.
This line right here is absolutely sage wisdom. Here are some of the companies I've bought services from, as well as what I remember happening to them:
ClubUptime
Closed in a disastrous closure due to basically being conned.
DirectSpace
Still around, haven't changed much
VolumeDrive
Very sketchy, I don't really know how they're still in business
Fazewire
Local Seattle hosting/colocation company. Originally founded by a guy
when he was 15, he sold the company when he went to college.
URPad.net
Still around, only used them for a short period of time.
OVH
Amazon
Digital Ocean[1] http://reviewsignal.com (reviews scraped from social media / public support requests) [2] http://serverbear.com/ (VPS/Cloud host benchmarks & stats)
I was thinking of my next step when I got an email next day that my account has been upgraded to 512 mb memory plan at no cost.
Definitely made the right move by moving from hostgator to webfaction.
Started with one place I can't remember - started with S. It was cheap and allowed multiple small sites to run off one account. Eventually turned to crap and so we bailed.
WebCentral - gave up on their silly control panel.
Used a place in NSW or Vic that started off OK, then was acquired (BlueCentral, I think?) and turned into a bit of a rude disaster. Abandoned that. Had a site break on Friday of a long weekend, so no support until Tuesday...
Had 60+ clients with CrystalTech who were polite and had great service. They were bought by NewTek and started overloading shared servers and tech support went from efficient and polite to slow and especially dim-witted. Bought again by SBA and things got worse. We've gradually shifted dozens of sites to accounts at Linode and that's been a great move for us.
CrystalTech were very good and only had one issue with a dedicated server in 5 years when a network cable came loose which was fixed after a quick phone call.
WebCentral has a horrific control panel, only one possibly worse is the cutdown Telstra Business version.
Currently using VentraIP/Zuver for a handful of VPS servers and they are always friendly and responsive.
I guess you have to pick your poison, a company that might not be so great but has been around, or someone that is great but fleeting. Although that is starting to change now with Linode, Rackspace, and others.
Plus I know of a smaller company that's intending to move in here. Hope that goes well...
The other consideration is the legals. Moving the server to Germany not only puts your data under EU/German jurisdiction in addition to AU, the new privacy laws to take effect next year will mean disclosing to your clients and they to their clients that personal data is stored outside of Australia.
Funny, my first real host (he.net) is still in business. Maybe it's true you get what you pay for ;-)
One of the big pains in the webhosting world is maintaining legacy systems...we had about 15,000 clients on ancient servers running RHEL4, under a proprietary VPS platform. (And as far as I know, a big chunk of them are still there.) Needless to say, this resulted in a really crappy service for the clients on those servers, and there never seemed to be a big push to get everybody migrated off of them and onto our newer servers running cPanel. We were working towards it, but it was a big endeavor that would leave a lot of clients extremely upset when things invariably went awry. So rather then putting some good development time towards automating the process as much as possible and hiring more support for those accounts that didn't migrate properly, the problem just sat there for years.
Short of ditching them I haven't figured out what to do. I'd hate to be a shared hosting company where thousands of sites break when the server software is upgraded (not to mention Wordpress installs getting hacked every day...)
Also: hacked WordPress sites. The bane of my existence. I must've dealt with several hundred of them in my time.
(My personal site has been on Site5 for over a decade; they have mostly been pretty good)
Is he? It sounds like he's deriding the hosting business (the providers, the prices, the way the customer is passed around, etc.), rather than the idea of shared hosting.
>but that doesn't mean the category is intrinsically bad.
Again, the category isn't bad, all the major players in that category are crap.
But, you could state the same thing about bigbox stores, and any large retail experience. You'll never get the boutique experience with an entity doing their best to squish profit out of volume.
All the 'major players' get a bad rap collectively but only because bad news spreads fast, and reputations get destroyed for random instances of bad news.
So, we are overpaying for our current hardware, but haven't had the stomach for another migration. Contrary to what the article states, small companies with already limited resources don't want to spend time moving a moderately complex infrastructure around, on top of the considerable work already on the table.
But, yeah, GoDaddy engages in questionable practices. Automatically adding stuff to your cart (and/or making it confusingly easy for you to do so), bumping renewals to 5 years by default, and otherwise making their UI "consistently inconsistent" in ways that miraculously always seem to benefit them are part of the equation. To be pushy with upsells is one thing, but they take it a step further.
These are kind of ingrained business practices and part of the same ethos that says selling IT services with sex is OK. It is hard to imagine them acquiring a company without that company getting at least a little of that stink on them.
So, that shadier/confusing stuff is unnecessary. They have a chance to be a really good company. Glad to hear they are working on it.
GoDaddy does a lot to support their customers. Friendly people over the phone. They've walked my dad through some hosting issues he had when he was trying to set a site up. They call me every couple of months to make sure I'm satisfied with everything (and probably try to sell me on that bundled registration). Making them out to be The Devil is too dramatic. And transparent too when he could have linked to the #Philanthropy[1] heading on their Wikipedia page but chose to focus on #Controversies to support a position.
Godaddy could shit rainbows all day long and I still wouldn't want to give them any money.
Fsck him and his company.
His own comment on the affair was "These people have literally nothing and when an elephant is killed it's a big event for them, they are going to be able to eat some protein." Now personally I supect that there are better ways for Bob to go and increase the protein in these people's diets, other than shooting elephants.
I started the transfer to another company an hour later. They have never and will never get another dime from me.
No, this isn't a joke or parody or satire. It's what actually happened. It's a horrible, customer-hostile company.
[1] https://en.wikipedia.org/wiki/Special:Contributions/ParsonsR...
I would argue that Mediatemple kind of killed themselves in many ways, I can't see how GoDaddy will do much worse to be honest. People put them up on such a high pedestal as they got bigger, they just couldn't live up to their glowing reputation because of how big they were growing which is a problem not many companies can say they have, Support stayed timely until the end, but Media Temple lost out to Digital Ocean and Linode big time and just couldn't keep up in the end.
I wish GoDaddy all the best, but for the moment I am very happy with my Linode 1024 virtual server plan which never buckles under anything I've thrown at it thus far. Even hitting the front-page of HN once upon a time didn't cause it to break a sweat.
You were smart to get out. I haven't yet.
We've just built our own data centre (http://blog.bytemark.co.uk/2013/06/04/bytemark-data-centre-w...), have launched our own IaaS product (http://bigv.io) and we get a lot of recommendations though as we're based in the UK we're not as big on most HN readers' radars. Not _just_ a plug, but as someone who wanted to go into video games and ended up running a hosting company, I've found it more interesting than I expected.
We still pick up business (and staff) from great hosting companies whose founders run out of steam. So seeing that happen over and over, our big challenge is to put a company structure in place where (one way or another) we don't have to make that awful "nothing will change - honest!" contractual obligation blog post. I've seen it happen over and over in the UK too - anyone remember DSVR? But also Melbourne last year, RapidSwitch, Redstation ... all beloved names with keen founders that got folded into a bigger company and gradually forgotten.
So I'm 1) trying to put in place products that will last at least another 10 years, and as we develop new ones, also have a transition plan to keep service going without ridiculous legacy maintenance 2) thinking about very long-term private ownership plans, turning it to an employee partnership, or anything else that will give new customers the assurance that we're here to stay even after Pete & me step away from it, as we will do one day. It'll be a few more years before anyone should believes us, but I think 12 years is already above average.
Some of the best hosting companies I've been with were run by "lifers" (which is what you sound like to me, too.) You've got a good team, a system, and you're profitable and growing your customer base. You're enjoying yourself, so there's no point in selling.
There are many others for whom the support gets to be too much, or they haven't charged enough to make a profit--in my case, it was both. If you can't make a profit, you can't put the systems in place you need to run a great business, so you kill yourself getting up at 3AM when a customer texts your "emergency support" line. Eventually, you burn out, sell it and move on.
Unfortunately, with hosting being as price-sensitive as it is (I like the comparison to airlines someone else made in this thread), there will always be companies that start up, have an owner that runs him/herself into the ground, and sell.
You "lifers" are in many ways smarter than the rest of us. I applaud you for not giving up and not giving in to the price war. Here's hoping other hosting company owners will read these comments and realize there is an alternative to full burnout mode.
I guess your life is made easier than many of the hosting companies by not offering shared hosting, and (from the impression I've got) generally working with larger clients. So none of those issues of upgrading PHP and finding clients' sites stop working, or having 1000 hacked and exploited Wordpress-hosting accounts to deal with each morning. If you encounter them, how do you mitigate those two scenarios?
DSVR brings back memories - I used them in 2002-2003? They had such a reputation... and then just disappeared off the radar!
I like the idea of providing a layer that floats on top of a solid network like yours, and solves some of the common problems in between that layer and the final website.
We still have very happy clients from 14 years ago, when we started, but the inventor in me has been inspired by new visions.
Love the depth of your outage notifications incidentally. One big problem for me - sadly stuck on shared hosting - is that the host never tell me a) when things go down and b) when they change things [that often seem to break sites].
Seems like you have a lot of good karma and should get more HN exposure if you ask me.
Side note. Nothing makes someone willing to pay a new hosting company more money for the same basic low end email service than when they have the inevitable problems getting their email from their current host.
The email is a bit less of a problem now then it was years ago.
The reason?
Many people now want the email forwarded to gmail etc. As such they aren't even aware if there is a service interruption because they are getting email addressed to the gmail and the domain and they lose track of which is which. So it could be hours or even a day until they even know something is wrong. As opposed to pop or imap where they get an immediate connection error.
Our company did reseller hosting for about 5 years and went through all of the acquisition stuff Marco mentions. We had to exit SoftLayer because they were horrible, only to be brought right back.
Hosting is a horrible business. To be good at it and have marketplace success you need to deliver over the top support; which is just unsustainable at scale.
They were poking through my code, bitching that I was doing "select *" queries ("those are inefficient" they kept arguing). After a week we were reinstated - turns out there was someone else running spambot on another hosting account on the same server. But... we were "ecommerce" so obviously must have been the root cause of a spiked CPU. I vowed at that point to not use shared hosting again, and haven't since 1998.
It's been nicer with cheap VPSs over the last few years - easier to get clients set up with their 'own' servers. I know that VPSs are shared as well, and I've seen issues where one VPS abused resources to the point where it affected mine, but that's been pretty rare (2x in the past 5 years I can think of), and it's always easier for people to track down, isolate and resolve.
Also, it was one anecdote point, and was 15 years ago, and was just 'bad service' but life is too short to try to be putting projects as risk to save a few bucks. Projects/sites generally have enough troubles - dealing with shared hosting is just one more thing that can go wrong, imo.
I haven't knowingly used virtuozzo for anything so I can't say. Most VPS I've had have been xen-based, but the hardware behind it seems to make more of a difference. I've got two virtualized xen projects right now, and same code on seemingly same virtualized hardware specs still yield pretty different results on some tests.
Then I hit #1 on Hacker News and bad things happened.
Coming from the shared environment, having a VPS is infinitely nicer. We would have an issue every month or two on shared hosting, but now it's an issue every year or so on their VPS knocks on wood
My general experience with DH is that uptime is fantastic... but when it rains, it pours for a day or so. But then right back to clear skies for months.
Am I wrong to have walked away immediately?
I've heard goodish things about Dreamhost, as well.
https://en.wikipedia.org/wiki/Endurance_International_Group
Many people spend time comparing the different services, but in truth they're all the same!
Also, you get much better specs with the free tier of OpenShift, but I guess that will change once enough people switch to it (just like AppFog changed their free tier).
The fact that many of them have the same parent company doesn't mean that their servers, features, support, uptime, etc. are the same. So they're no more the same than Facebook and Instagram are the same social network just because they have the same owner.
They're able to maintain the appearance of having a bunch of competing companies and you can shop around for the best deal, which they will change around periodically, but it's always the same thing underneath.
Source: Worked for two webhosts acquired by EIG.
Pretty much like Apple.
https://www.nearlyfreespeech.net/
Definitely not aimed at large(r) sites, but for my static sites and a few WP installations it works fantastically. The control panel takes some getting used to, but the "pay for what you use" business model more than makes up for the rough edges. Its all À la carte and I love it, I've been a customer for 5 years with no problems.
Really, it's sad, because it's pushing a lot of folks who really shouldn't be running their own servers into the VPS market.
Thats the thing, though; VPSs, generally speaking, have much harder limits. It's harder for that one user to make the server suck for everyone. I mean, it's not as good as a dedicated server, but it's a big step up from the isolation available in shared hosting.
Now that the market price for VPSs has fallen almost to the shared hosting level, I wonder if services that implement a shared-hosting like environment within managed VPSs will take off? Something where the user doesn't have root, where it's managed by the hosting company (presumably automatically) but where there is only one user per virtual.
There are PAAS providers that operate that way, sure, that will let you run languages better than PHP... but there doesn't seem to be an ecosystem of PAAS providers that are all compatible, like there is with php shared hosting.
What interests me about this sort of "PHP as a service" is that unsophisticated users are used to dealing with shared hosting. They understand the limitations. And they want the resource isolation of a VPS solution, even if they are unable or unwilling to put in the sysadmin work required.
That being said, the companies I've had good experiences with, have heard others, and will continue to use/pay are: AWS, Linode, DigitalOcean, and Webfaction (webfaction is amazing for a small cheap shared hosting environment). Other ones that cross my mind are OVH and Hetzner.
If I need to host a more complex or demanding web application it goes onto a dedicated linode (or may share one).
Dedicated servers that are reliable are very very expensive (Hetzner in my direct experience is nowhere near reliable) where with linode across 3-8 linodes at various times I've had no down time in coming up for 5 years.
Fantastic support, they don't oversell their machines.
Sure if I shop around I can get a similar spec (whether it delivers who knows) for half the price but is it really worth saving 20 bucks if I don't sleep at night worrying about my vps provider going down.
I also like DO, I still won't host anything important with them but for a quick dev/test box they are pretty good.
I've never really gotten why the VPS market is quite so price conscious the difference between 5 a month and 20 a month is so meaningless in the grand scheme of things (I suspect I spend a lot more than 15 a month on coffee on the way to work).
About softlayer: Its possible to bargain with them. We have E-2620 servers there, official starting price at $879, and we pay $299/month including more RAM and a small network. So they had been willing to undercut co-location calculation if you ask them. I dont know if this is still possible after IBM. I guess their sales team now knows better how to barter with big customers.
I'm not sure I got it 100% right, but I think there are some valid points.
The support person refused to do so, but instead, asked me to subscribe to a dedicated server. I explained that I didn't need a dedicated server, as it was clear from my statistic that all faults were on their IO throughput side. He just won't listen and still insisted on up-selling me a dedicated server.
What a horrible experience! Anyone encountered the same thing as I do? Is IO throughput a PITA for your hosting experience?
Made this point in another comment but want to stress the biggest barrier to entry is being able to provide customer support and handling the "rtfm" type calls. So it's a people problem. In the sense that you could start doing hosting as one individual but at a certain point you'd have to hire someone to take care of the support calls that a larger customer base (than one person can handle) would require.
I think people will always pay for service, quality, and experience. Whoever can deliver that consistently will make money in hosting.
I would say that usually the hosts that are trying the hardest are the smaller ones that are not "household names" yet. Once they break out and start growing really fast, that's when the people who made it happen tend to check out and let things fall apart.
I would like to know if I'm actually just lucky or if other people have had this experience too.
I was using HG for probably 7 years or so. But I believe they recently got acquired by EIG[1]. I'm not saying there's a connection, but I've experienced several incidents that made me change my mind.
At some point I was moved to new servers. The migration process was very sketchy. HG did not send notification emails about the migration on time, which meant once the migration took place, I had to try to figure out which DNS changes to make (were not provided), whilst the service was effectively down.
Afterwards they've had a rather major issue with their email servers, which got blacklisted as a result of a large amount of spam being sent from compromised Joomla apps[2]. This resulted in many forwarded emails to bounce and (my) customers being both aware of the forwarding (which wasn't always desirable for me), and that their emails bounced. The Hostgator response to this incident was painfully slow and they did not seem to appreciate the gravity of the effect on their customers.
I already had shared hosting services running with webfaction, so I moved most of it away to them. For more heavy-weight hosting, where I need more control, I use Linode and AWS.
[1]http://www.shoutmeloud.com/hostgator-sold-endurance-enduranc... [2]http://blog.hostgator.com/2013/08/23/global-blacklisting-res...
I'm using Hostgator for around 3-4 years now. Never had any problems with them and the support was always great.
but i cant help but feel, the MT story was forced into this post to make the much general point that most hosting companies are horrible
also, he does seem to miss a smallish fact MT or he doesnt raise it clear enough .. MT was not a great host .. it was an expensive mediocre host ... but i think he probably did downplay this a little to make his louder statement that things will get worst for MT ... mainly because the founders left
i have to disagree, MT wasnt, he sort of admit it, the founder was never a believer he admit it ... MT didnt loose much
plus if hosting is such a comodity and MT wasnt good .. the customers should feel they really lost anything
again i believe marco used MT story as just an excuse to make this post ...
I wonder how many of their employees will leave MT
Their datacenters are located in Frankfurt so the roundtrip to the US might come with a latency you'd want to avoid, but you should at least try them anyways. They will gladly offer their support in english, but consider that english isn't their native tongue. You can even send them GPG encrypted mail.
Uberspace is quite young (they started in the beginning of 2011 afaik) but they've only improved during all that time and there was no decline in their service quality after word got around that they are the go-to provider for german customers seeking shared webhosting. They offer anything from ruby/rails to python to nodejs, mongodb, postgresql - and even php in different versions from 5.3 to 5.5 including almost all revelant point releases. They only offer 10GB of storage which is not easily expandable but they place no limit on the amount of Uberspace-accounts you create, so if you are hosting different projects you can simply scale them to different accounts which might end up on different nodes. (which are never too overbooked that it might impact the performance - and if it does, they'll upgrade the hardware to fix that!)
You decide what you want to pay! They want at least 1EUR per month but you can adjust your price to anything you want and you can even change your prices on a monthly basis. They say that they want all people to be able to state their opinion on the internet and that's why they're hoping that people with a bigger budget chose to set their price to their recommendation of 5-10EUR per month. But they will never beg you to pay more if you stay with 1EUR and your service won't suffer either. (I have several accounts and a 2 of them have the default price and I got amazingly fast help via mail despite only paying 1EUR for these accounts)
You don't need to give them any personal data if you simply want to create an account (all you need is <8 letter username and a password or OpenID). They will give you a fully usable Linux-account on one of several dozen CentOS-powered servers in return and you can even run your own services (via djb's daemontools) or ask them to open up a higher (>61000) port for you if you want to host a xmpp-server or something. You can't get unencrypted (as in non-TLS) IMAP/POP3/SMTP and they don't offer FTP because of its bad security. Instead you'll work via a fully capable SSH-connection and have SCP/SFTP-access to transfer files. Their webinterface is simpyl called "dashboard" and offers only the most basic stuff like creating virtual mailaccounts and password and SSH key-management. They're giving you your own IPv6-address and their servers have been dual-stack from the beginning. You can also create your own SSL-certificate for your own domain (which you don't even need to register/transfer at Uberspace, just add the domain to your account and setup your DNS and you're done.
If you're up to the challenge and have lots of experience as a Linux-systemsadministrator, start your own Uberspace-service outside of Germany and you'll get rich within a year if you can offer their commitment and service.
I'll never ever go back to lame hosting providers where I'll have to fiddle with crude and slow webinterfaces. At least for me the roundtrip to Sweden is acceptable and their documentation is understandable when translated to english by Google.
On the topic of Marco's claims regarding hosting providers I can simply state from my experience that Uberspace seems behave diametral to his expectations so far.
(I am not being paid for this and I'm also not affiliated with Uberspace other than being a happy customer.)