Ask HN: Prgmr is unreliable. Please advise on VPS hosting.
My experience has been OK, except for one thing - they are completely unresponsive. I know their motto is "We don't assume you are stupid." but charging people money for a service and not providing any support is just rubbish. I've emailed them with some questions in the past and they never get back. A while ago they started issuing ticket numbers and I thought they must be improving the service, but they still never get back.
Last couple of days, my server has gone down a few times. It went down again this morning. Now I can access the root domain, but not the subdomains. I have written to prgmr support and I got my ticket number. But I'm not expecting any real help.
I know this is a bit desperate but I'm hoping HN user lsc (who runs prgmr) will take note of this.
Meanwhile, can I ask you to advise me on choosing another VPS host? I know Linode and Slicehost are the most recommended ones here. Given the prices, I might go in for Linode but I'm wondering if any of you have advice on running a beta cheaply until launching a product.
Sorry for the rant. This is my first time running a VPS server. I'd appreciate any help.
EDIT: Thanks a lot for your responses everyone. In addition to Linode, a few other good alternatives have been suggested which I will look into.
I have managed to get my server up again now and also got a response from prgmr. It's a shame that I had to do this publicly, I don't mean to defame prgmr. But I think it's worth pointing out that this is possibly one bad feature of competing on price.
111 comments
[ 4.4 ms ] story [ 180 ms ] threadedit: oh, and possibly consider www.ovh.co.uk, a French company. Never used them myself, and don't know how good their VPS servers are, but have heard many endorsements for their dedi servers (and their prices are very cheap)
I run everything on Rackspace Cloud Servers, from SaaS apps to a very large Ecommerce site, and I'm extremely happy with them.
The linode library is pretty helpful too, especially if you are fairly new to linux or systems administration.
They're big on Django and I guess Ruby (don't use it). No problems getting a Django site up and running. No issues installing packages or other software in my user account.
http://book.xen.prgmr.com/mediawiki/index.php/Complaints
But even with that said, I've found their support to be great. They've helped me with a fair share of configuration issues, even though that's well beyond what I thought they offered.
If you have had good support, that's fair enough. I haven't and complaining on a public forum was my last resort to get any attention or help. The fact that no one from prgmr has yet responded to this thread is telling. It seems that one gets such bad customer service either from really small teams or giants like Google. I am not scared to call businesses out on poor service.
The response time for emails isn't great, but server outages seem to get fixed pretty quickly. Furthermore, I think you need to accept in your case, that it probably wasn't a hardware issue....
Yeah, was gonna say - a VPS provider's job is to make sure you have a running box and a good pipe to it. This is almost surely a configuration issue on the VPS and not the provider's fault.
I certainly understand your frustration, but I think the heart of the matter is that prgmr may be inappropriate for your needs rather than "prgmr is unreliable." I feel like they're quite clear about the fact that they're not for everyone.
If you need high-availability, you can buy 2x prgmr nodes at different locations, for the same cost as most other VPS providers, and you'll get MUCH better reliability that linode or such.. But perhaps you should clarify that it was your fault in your post.. I don't recall ever seeing any long outages..
Also, Prgmr likely runs monitoring software to identify when servers go offline, in which case, they may be able to respond to true outages sometimes the moment they happen.
Their motto is: "No bullshit"
http://en.gandi.net/hebergement/
Also the first month share is free :) which was nice.
I've been a Slicehost user in the past and really liked their service.
The upsides of Rimuhosting are:
* Stellar support. No really, stellar support. Their techs know what they're doing, and they respond quickly * Flexible plans - choose your own distro, VPS size, even semi-dedicated plans in between shared VPS and dedicated servers. * Did I mention the support rocks.
A lot of what I know about system administration, I learned from Rimuhosting techs. Not only do they fix what needs to be fixed, but they send you a great summary of the steps they took, complete with copy/paste from the command line. It's a classic "teach a man to fish" approach to support. They've spoiled me so badly that I get a little upset when I receive terse "problem resolved" responses from other companies.
I can't say enough good things about them. They're worth every penny they charge.
I've also noticed that periodically when I'm researching server setup-related questions, Google points me to their resource pages, which seems like a good sign.
Advantages: - Rock-stable (didn't have any issues so far) - Cheap (e.g., for 3 Euros a month you get a vServer with 384 MB RAM and 8 GB harddisk) - Non-Enterprisy: Support is available via IRC instead of an useless callcenter
Disadvantages: - It seems to be a one-man company. However, it has been in business since 2003 - They are based in Germany. So from the US the latency to their servers is somewhat higher than a US-based provider and their websites are primarily in German
If you check them out have a look at their "vps premium" offers. they are even cheaper then the offers that you find directly on the mainpage.
Their prices are very reasonable, they have a nice, simple online management thing that lets you deal with DNS, re-imaging chunks, etc.
I have used/do use: EC2, Slicehost/Rackspace Cloud, Linode, and prgmr.
I found prgmr to be a pain. There are just some small things that make administration so much easier.
Slicehost/Rackspace has gotten expensive compared to the competition, but they're pretty rock solid.
EC2 has some wonderful features, but it doesn't go down to the cheap levels I want for my personal hosting.
Linode is always a great performer and the only reason I'm not with them is that I'm getting a better price at ChunkHost. While I haven't had problems with ChunkHost, Linode is a larger player if you're looking for business hosting you can sell to your boss. As others will tell you, it's easy to recommend Linode.
Plus, for a beta, the service isn't too bad. But just to clarify, I'm only using it for personal stuff. YMMV
I bet lsc would be interested in your feedback if you sent him an e-mail.
Maybe not, but maybe.
I left linode because I kept getting emails from linode support about disk I/O, and that was for an initial import of a 50GB postgres db.
After that, I kept getting emails from them once in a while, when my app was getting disk I/O busy. I can't run a business with these worries in my mind.
Slicehost is and has been working great for me, and they have great support.
This is not meant as a warning or a representation that you are misusing your resources. We encourage you to modify the thresholds based on your own individual needs.
You bailed out because of that?
As davidw pointed out, those e-mails were generated by completely customizable thresholds. If your default usage results in disk I/O rates higher than the default settings, then you should just raise the settings. The warnings are very handy if your node starts getting starved for RAM and begins to swap stuff in an out a lot.
I have no idea what kind of storage strategy they are using, but if I got support tickets filed because i'm doing a db import, well... they're doing it wrong.
Their articles is another plus point. It's constantly updated. It covers most of the things you need. No doubt they are not the cheapest. But their customer service more than makes up for it.
What was the issue?
I got a reply from support at prgmr but haven't been able to trace down what exactly happened.
In our case, we just got bored of all VPS providers and took a dedicated server. Take the time to compare a VPS to a dedicated server and you may have a few surprises. My current server is 80$/month and a similar one on Slicehost would be 250$/month.
our support, yeah, isn't great. Did you at least get a refund on your last month? we try to be generous with those. (for some things, I know, money doesn't matter... but I think money is the only really sincere way for a corporation to say "I'm sorry")
Slicehost and RimuHosting by far have the best customer-service experience. I really liked the idea of slicehost, but unfortunately in practice trying to scale up a 4GB instance to an 8GB instance to handle load isn't fast-enough to actually be a great production strategy -- your VM contents are effectively copied to another machine and booted in it's place, so if you have a ton of small files, that copy operation will take a long time, especially if the server is getting hammered.
Also, you cannot scale beyond the 15.5GB instance because it is 1 physical server. So if you have potential massive growth down the road, you'll have to look into load-balancing.
As for cost, I disagree with the few that said Slicehost was expensive -- in our experience it was a lot cheaper than EC2 -- go ahead and price out comparable VM's on EC2 and don't forget the bandwidth, that is where they get you.
Unfortunately we didn't stay with Slicehost because of 2 VM host failures in 2 months. About 3-4hours of complete outages both times and 1 time our VM failed for no reason... it was really frustrating to have all these mysterious issues and it seemed there was no "fix" for these things... they were just god's will or something, who knows.
We moved to VPS.net next and it was great, for 12hrs... in the following week we had 2 mysterious complete failures of our VPS (tried, 4, 8 and 12-node configurations IIRC) and each time the 1st tier support would respond with "Our server admins are looking into it and will get back to your shortly", and then another 3 hours of down time before a useless "Ok, your VM has been restarted" response... that's it... no identification as to why the failure occurred so often or what was going on.
Again, another great idea "in theory" that just ended up sucking for us. However, if you need massive scalability, VPS.net will let you scale your individual VMs up to like 64 "nodes" -- which comes out to some insanely large machine.
To their credit, I got tweets and emails from the CEO and head of CS to help me after I blogged about the experience... but it is one of those "thanks but no thanks" situations...I didn't want to keep doing that dance with failing servers.
I eventually ended up BACK on RimuHosting -- 2 years prior we had left to try and find a more easily scalable VPS platform after no downtime on Rimu. We got dazzled by these other AJAX-enabled management sites and so on... Rimu has a very simple/ugly web interface, but an incredibly responsive team of very very smart people all dedicated to server stuff... and in the end we just couldn't come close to replacing that with real world experience on high-load sites (Dugg, slashdotted, etc.)
That being said, there are a lot of options Rimu doesn't advertise on their site well that make it one of the cheapest hosting solutions out there... like going over 2TB of bandwidth a month in their Dallas center only costs $0.10/GB -- Amazon is $0.15/GB
They can also scale you up to an 8-core, 72GB monster dedicated server on the high end for $1k/month -- price that same thing out on EC2 with the same allocated bandwidth and it's like $1700 even with the reserved instances.
Then you throw the really responsive customer support that is willing to do almost anything reasonable for you for free (including configing/installing software, etc.) and even though they have no sexy AJAX on their site, my life is a lot easier hosting with them.
We do all our serving through them, even stuff we have to scale over time with growing VPS accounts -- if you need custom setups that aren't on their site, just ask. They'll likely toss it together for you.
--- Hope that helps, I know you have a lot of feedback to read.
Our app servers (on Slicehost) were getting hammered yesterday, so I issue a resize (1GB -> 2GB). After about 20 minutes it gets to about 97% and then stalls. In a panic I log into the server and kill all the apache2 instances. I look at the Slicehost console again and it says 40%. It actually started over. By the time it actually restarted with the resized slice it was over an hour later. The traffic was already gone by the time the resize finished.