I, like most of us, did a stint on dreamhost back in the day. Shared host, 3 dollars a month, who wouldn't want that.
Left when I realized that my email ___domain couldn't consistently send and receive mail. It would sporadically die for a few hours then be up a week, then down for a few hours, etc.
Hoped they would get their act together one day. Seems not.
I've been a dreamhost email and hosting customer for 18 years. My only downtime from them was a 3-day outtage after an earthquake disrupted power to their datacenter near LAX. No downtime since.
I did have a brief outbound email deliverability issue with yahoo in 2010 and again with a small ISP in Wisconsin in 2017. Both resolved within a few days of contacting support.
I'm glad you have had a good experience with them.
For me, in 2006, my email services was the most important thing I had. I ran a family email server that was used by a small group of us. Any outages were hard to accept. We had a lot.
It may have been the host I got put onto, the ip address of the host may have had a shared-account-bad-actor causing an RBL, or any number of reasons.
Just goes to show you that one experience is not all. Thanks!
In my experience, Dreamhost's e-mail has been the part of the service that always worked.
I'm under the impression that the architecture of their e-mail service has been some sort of HA clustered setup for ages, and I think one of their unusually outspoken blog post in the 00s said outright that their customers are less sensitive to web than e-mail outages.
I've had an account with Dreamhost since June 2007. Their web uptime wasn't great back in the day. I seem to remember that they made a bad investment in some sort of very expensive NAS or SAN architecture from NetApp that caused a lot of downtime because the filers just regularly dropped clients.
However, with the dawn of cheap SSD, I think they moved back to local, per-server disks in the early 2010s.
I still have my Dreamhost account. They don't use cPanel, which was appealing to a younger me, for aesthetic reasons, and they also gave away sub account privileges you had to pay extra for as reseller hosting at other places.
My DH account has been a Swiss army knife of sorts for doing small web and hosting related things. This was especially useful back in the days before you could just use free Cloudflare DNS for random things.
These days, I use my own M365 business tenant for e-mail, and all my web customers are on servers in Europe, largely on hosting accounts based on cPanel, which I hated before. Now I like the ease with which I can ask a new provider to migrate all my cPanel accounts. It keeps the web hosting business extremely competitive.
Dreamhost also uses absurdly old hardware, my account is on an 8-core AMD Opteron 4122 CPU. Performance isn't particularly good. They also use Debian, so they don't rely on new features that CloudLinux and LiteSpeed bring to the table for shared hosting, to limit resource congestion/noisy neighbors.
My plan is actually to get off Dreamhost during my upcoming summer vacation.
Email problems are why I ditched DH. I was waiting for some very important visa and job offer emails and though I could send, I wasn't receiving anything and missed some deadlines that ended up losing me a job. Their support said that email was essentially a 'bonus' and there was to be no expectation of it's reliability.
Been a customer since 2004. Shared hosting for most of that time, VPS since 2018 or so. I'm thrilled with their service, and feel like they've always had their act together. I only use them to self-host personal stuff like FreshRSS, pastebins, and fossil, though; nothing professional or high-scale.
It's not super nice of you to automatically assume that cheap(er) devs in cheaper locations like Colombia or Philippines are bad, or at least inferior to more expensive ones.
Of course they aren't, you'll have good and bad devs in expensive and cheap locales alike. There might be language or cultural barriers that make collaboration across country lines, but that's pretty much it.
>Note regarding ___location: This is a remote position within Colombia only. This position is only open to people living in Colombia.
I've never seen anything like this before. What's the reason for hiring this way?
If all their jobs were in Colombia, I'd assume that there's something legally that makes it hard for them to hire internationally, but they have roles lots of roles in the Philippines and a few roles elsewhere, so I'm stumped.
And it may be too harsh to judge a whole country and its talent just because a company whose servers are down hire talent there. What about the AWS engineers in those countries? Or Airbnb, Stripe, Twilio, etc?
To clarify, this means that of the many servers Dreamhost has, only one is down, not that Dreamhost shared hosting is just one server.
And I can confirm that the downtime is not across the board. A podcast that I listen to has its website on Dreamhost shared hosting, and that site is up.
This hit the same day I released a hobby music side-project I'd just spent 200+ hours of time and nearly $10,000 on.
The website has been unreliable, constantly -- for nearly 5 days. Their support has been useless.
Their communication has been useless — I would never have done the release promotion had they clearly communicated that my website would be effectively offline.
They have made no effort whatsoever to make things right with customers. It's as if they don't know what they're doing -- or don't care.
Instead of celebrating the release of geek music, I'm dealing with my listeners sending me emails like:
"Fourth time's the charm"
"got a CloudFlare error once, and some confusion with the cart (0 items, then 2)"
"It eventually worked on the fourth try just now, but the whole site (not just the emailed links) is loading very slowly, with frequent timeouts at the moment... (Both static pages and logged-in account pages.)"
"At checkout it hung for a long time with spinning circles and then gave me a message in red at the top that said something about it having failed."
Dreamhost has an interesting history. Not sure which came first, but they also ran a small design network called Newdream, which was similar to a lot of early Web 1.0 design zines/collectives like Swanky, lockjaw, and others. IIRC, Newdream was responsible for Webring and a few other cool things.
Those collectives and early adopters soon gave way to news/content aggregators like k10k and Newstoday. They were neat in their own way but I personally loved a lot of the anarchic spirit of those early groups. So much of the dominant design aesthetic back then was still very much tied to print principles; it was those groups that were really exploring new layouts, what JavaScript could do, and so on.
The closest analogue I can think of is the demoscene; really early technologists exploring a new space and coming up with really interesting, modern art. It's a shame a lot of that is lost in time.
Just checked my toy project sites that are on Dreamhost shared hosting, and they're up too.
That said, it's not as if I'm particularly happy with Dreamhost. I regularly have problems with (static HTML!) pages just timing out, both from my own browsing experience and from what I see in ahrefs' free monthly website monitoring reports. If these sites were generating revenue I would have moved them off Dreamhost long ago.
"[Investigating] Our Technical Operations team is currently investigating connectivity issues affecting one of our shared web servers (iad1-shared-d12-02). We are closely monitoring the situation and will be posting periodic updates as further information becomes available."
(So the subject 'Dreamhost Shared Hosting Has Been Down for 3 Days' is overstated.
I've been a Dreamhost customer since ~2006 and am not seeing any issues. I've always liked what they have done, with only a few complaints (give me back squirrelmail damn-it, I don't always have access to my systems and atmail is literally the worst webmail client ever). Any time I've had a real issue, they have resolved it quickly, even bending over backwards for me a time or two on billing issues to help me avoid issues with domains expiring. I hope this gets resolved soon, but it does sound like a pretty minor single server. (their shared hostnames are interestingly grouped)
Is the downtime 2 hours or 3 days? They said they pushed out a fix 2 hours after the initial post and root caused it a couple days later.
This title “Dreamhost Shared Hosting Has Been Down for 3 Days” is clickbait and the only customers chiming in here are saying that they’re not affected.
Any good alternatives? I've been using DH to self-host Wordpress (which has been a real bad time), and real simple short-lived public-facing prototypes/etc. It's always been not great, but cheap and not important, and swapping out too much of a pain. But it's time.
(And "Go host a high-volume-spike blog at ___ like a normal person, and then solve the other problem" is entirely valid and welcome)
If you need a place where you can out several websites that don't make money for a sustainable, reasonable small cost, you'll be surprised by the price/quality/performance offered by good, independent cPanel based hosting providers these days.
By independent, I mean not owned by GoDaddy or Newfold Digital (formerly Endurance International Group, EIG). Just as an example, Bluehost, Hostgator etc are owned by Newfold Digital, with big brand recognition and subpar service.
cPanel isn't aesthetically pleasing, but it keeps the market outside the big names extremely competitive. You can ask another cPanel based host to migrate your accounts at any time in a process that's standardized and quick.
With NVMe disks and software built especially for the hosting industry, such as CloudLinux and LiteSpeed, these hosting options can be very performant. On CloudLinux, all hosting accounts run in their own containers with RAM and CPU limits so the system as a whole stays responsive. These environments can easily feel as snappy or better then self-configured virtual machines.
I tried out A2Hosting's Turbo hosting early this year. The performance was very good, but their EU datacenter had connectivity issues.
I now use a British company called Stablepoint, run by people who used to operate TSOHost before selling out to GoDaddy. Stablepoint uses public cloud infrastructure, so the reliability is very good for very reasonable prices around numerous data center regions globally.
I do use Stablepoint's highest-end reseller account, so I can't speak for the performance of their regular servers. A2 is probably very solid network wise in the States, and their non-reseller Turbo plans are affordable and very well regarded. Both companies have generous 2-4 GB RAM limits per cPanel account on their mid-tier accounts.
Easiest is probably a dedicated WordPress hosting provider like Kinsta or WPEngine. More expensive, but they definitely take away $20-30/month worth of hassle and worry. Stick a free CloudFlare plan (they have some WP-specific security and caching options, IIRC—if those are available on the free plan, make sure you turn them on) in front of it and block admin and API paths except from allowed IP addresses, then turn on auto-updating (plugins and core) in WP itself, if you want a reasonably secure and low-effort site that can survive occasional large traffic spikes without sweating. May have to tweak a couple settings to make sure it's serving from CDN nearly 100% of the time, but once that's done, it's done.
If you like wordpress, I had an excellent experience with wpengine. They handle large chunks of security (nontrivial for the junk show that is WP and plugins), backups, came with a staging environment (maybe not all accounts? But ours did), etc. It was as good a time as you can have with wordpress, imo.
Oh, and I got good support who quickly figured out a db issue that a custom plugin was having and got us a solution. Turned around in 1.5 business days. And we weren't on a super expensive plan.
If you're interested in something managed and curious about performance, I just published https://WPHostingBenchmarks.com this week (with 2022 data - 33 companies, 79 hosting plans benchmarked for performance). There was one company on there that offers a command line tool to stand up WP servers on your own Digital Ocean account as well.
I personally would lean towards nearlyfreespeech.net as a first choice. They seem to be highly robust. At least this holds until we get to sites that need a full daemon.
I always wonder, for folks who self-host on Droplets or EC2 - how do you handle provisioning the server to ensure that it's secure and has everything you need? I use Laravel Forge for this. I'm not a server admin and don't want to pretend to be.
For droplets and ec2, you 'own' the inside of the virtual machine and its your responsibility to install software that you need and patch it according to a routine schedule.
That said, both services have firewall rules you put in place that help manage this. IE - you may expose SSH to your local ip address, but only port 80/443 for the rest of the world.
You are right though - its another attack vector. If you don't want to muck with that, and you have a static site, you could put your static site into S3 and then host with cloudfront. With that, you have no risk.
Agreed, Cloudways is great. Just a little expensive.
It essentially provides the power and flexibility of a cloud VM with the convenience and "set-it-and-forget-it" nature of shared hosting.
For those who don't know, Cloudways is essentially a management service that runs on top of a cloud VM of your choosing (choices include DO, Vultr, Linode, AWS, GC). For a premium over the cost of the VM, they take care of all the provisioning and config for you, while still providing a powerful platform (allows SSH access, etc) and some added functionality (Git hooks, staging sites, credential management, and more out-of-the-box). It's really quite slick.
Cloudways is good at providing a control panel with a user experience resembling expensive Wordpress hosts, including developer workflows and such, while also giving a little more control over the server stack. You pay per server and its resources rather than per site.
The default Cloudways Wordpress stack performs very well. Their caching solution is built on Varnish, controlled by a competent plugin. However, I do find that some themes and site builder themes simply don't play well with at least Cloudways' Varnish conf, so one might end up using the included memcached or redis for page caching, with the help of some plugin like W3 Total Cache.
Customer support can be hit-or-miss.
I wouldn't necessarily recommend using Cloudways' lower tier plans, and Cloudways doesn't recommend them for production use either. I pay about USD 100 per month for a server on Digital Ocean that hosts sites I maintain with a web agency. We rely on Cloudways' development tools. For the number of sites we host, it's a very appealing price compared to certain big-name Wordpress hosting companies.
For someone just wanting to host one or more small websites for a sustainable, low cost, I would recommend looking into well regarded cPanel hosts that runs CloudLinux and LiteSpeed on NVMe hardware. These companies also have some very decent, standardized Wordpress management features these days, even though the cPanel GUI us far from pretty. I describe this in more detail here https://news.ycombinator.com/item?id=32255962
I'll second Siteground. My WordPress sites were gradually getting slower on DreamHost - with support saying there weren't any problems. I switched to Siteground a few months ago, and so far, I've been happy with them.
Siteground had a "move my WordPress site" plugin that worked much smoother than I expected.
I had a client which ran a $300M/annum business off a handful of servers with zero redundancy. Their PII/GDPR data was unencrypted on a single spinning disk.
Any advice about risk mitigation fell on deaf ears.
It was the most terrifying 6 months of my career and taught me a harsh lesson in risk assessments of potential clients.
My first job was at a tiny ISP during the dialup->DSL/cable transition, and I got put in charge of setting up the very first RAID array they had, which was a RAID-5 hardware array (SCSI, natch). It was a 4 drive array, we moved most of our fileserving over to it, next project was to set up a hot standby for the server.
Of course, that was when DSL started to become a lot more common, and business started to suffer. Money got a little tighter, and after a month or two we lost a hard drive in the array - no problem, grabbed the spare, swapped the drive. Then I asked for a replacement spare - "sorry no, can't afford a spare drive, let's wait a few weeks"
Two weeks later, another drive fails - and now we have no spare. I go to the CEO and ask for his amex to overnight a replacement drive, and he says, "Ehhh, no", and I said, "well, you know, if we lose another drive, we'll lose the whole array", and he said, "Yeah, but what are the odds we lose another drive in the next few months?"
I left that job two months later, constant nightmares in the meantime. A month and a half after that I heard from a former colleague that indeed, another drive had failed, and turned into a four day downtime for the entire ISP.
Left when I realized that my email ___domain couldn't consistently send and receive mail. It would sporadically die for a few hours then be up a week, then down for a few hours, etc.
Hoped they would get their act together one day. Seems not.