Noodlings

I just switched PotionStore to use Phusion Passenger. Also known as mod_rails, Passenger is an Apache module that allows you to run Rails with Apache. Unlike other Apache plugins like mod_php, your application is still run in separate processes. Previously, I had been using Apache as a proxy to a mongrel cluster. On the surface, this doesn’t sound much different but Passenger does give you a couple things:

• It maintains the pool of Ruby processes for you. It can adjust the pool dynamically as needed in case you want to reclaim memory when it is not busy, for example. You don’t have to worry about setting up and maintaining a separate set of servers like you do with mongrel. It gets restarted with Apache and you can also trigger it to restart just the Ruby stuff. One less thing to administer and monitor.
• Lower memory footprint if you use Enterprise Ruby (also made by Phusion). It will share resources between the Ruby processes.

Luckily, Andy Kim already played guinea pig and tried it out to make sure it worked. Many thanks to him for that (and for the whole PotionStore thing to begin with, of course).

While the setup was fairly simple, I ran into a couple odd issues. For one, the Enterprise Ruby installer seemed to screw up the permissions of some of its files. All of its .so files and a directory or Ruby file here and there were set to be only readable by the owner. Make sure to check this before deploying. Note also that it installs as a totally separate Ruby installation so run its version of gem to make sure your Ruby packages match what you had on “regular” Ruby. For those of you are running PotionStore, make sure to do a rake rails:update otherwise it’ll bomb and log a message telling you to do so.

Unfortunately, I didn’t record the memory usage beforehand so I don’t know the exact gain. Based on my recollection, it does seem like I have maybe 20M or so more than I did before (for two Ruby processes). One odd thing I’ve noticed in my graphs is that my interrupts and context switches plummeted immediately. Not sure why that is but it seems like a good thing to me.

While this doesn’t remove Rails’ lack of thread-safety problem (resulting in a separate process per request), it does at least make the deployment much, much easier and with the memory savings, a bit more scalable as you take less of a memory hit with each extra Ruby process. Especially for those of you that have not deployed yet, this will save you a bit of a headache in configuration (no proxy and mongrel setup). It’s only been up for a couple days so it may be too early to tell but so far it’s been running fine.

It’s been over a month since I’ve been with Slicehost so I figure it’s enough time to make an assessment. Especially now that the MacUpdate promo is over, I actually have a sense of how well things hold up under load.

For those who don’t know, Slicehost is a hosting provider. What sets them apart from shared hosting providers is that they provide you with what they call a “slice” (other similar providers may call it a virtual private server or VPS). What this is is a virtualized server of your own. From your perspective, it’s like getting a dedicated server. You choose what OS you want (which right now consists of different Linux distributions) and you get root access so you can do whatever you want.

It differs from shared hosting in that your slice is like it’s own machine. It gets a guaranteed amount of memory and CPU so even if your neighbor is a hog, it won’t affect your slice. Because of the way that the Xen virtualization works, it is impossible to oversell on capacity.

Compared to getting a dedicated machine, it’s much cheaper and you aren’t tied to specific hardware. Blown power supply? Not your problem. I don’t know exactly what they do in this case, but I imagine they can move your slice to different hardware as needed.

Set up

You can read my original report on getting set up. You are expected to set up and administer the slice yourself. If you have the inclination and need a high level of control, then this is probably for you.

It only took me a weekend from getting my slice to having everything migrated over. Of course, being the tweaky type that I am, I spent some days playing with it and optimizing it. One of the benefits and dangers of having full control.

Upgrading

When I launched the new store, there was a problem. Ruby on Rails is a memory hog and as a result, I needed to upgrade to a bigger slice. Fortunately, Slicehost automates all that. Just log into the management console and request the larger slice. It takes a little while for the slice to get prepped but during that time your slice is still up. The downtime for the reboot was short (less than a minute) and that was it. The fact that it’s automated is a big deal to me as it means I can do it on the fly without doing a drawn out back and forth with a support person.

Traffic

As you may or may not know, Hazel was included in the recent MacUpdate bundle. Before the launch, I upgraded my slice again to 1024M in anticipation of the load. Turns out, this was unnecessary. The 1024 slice never broke a sweat. The load went up briefly to 0.4 once. Apache connections stayed below the upper limits of what a smaller slice would have been able to handle. Traffic was about 150K requests a day at its peak. I don’t really have a frame of reference for that except that it’s a good bit more than I usually get. In short, the 512 slice would have been able to handle it fine. The slice has performed better than with my previous providers and I haven’t noticed any slowdowns or downtime (except for when I restart things for maintenance). With the promotion over, I’ve downgraded my slice and things are still running smoothly.

After all this, I’ve only contacted support twice. Once in the beginning just to say hi and once today for an issue that ended up being an Ubuntu thing. In both cases, I received responses within the hour. Granted, I’m doing a lot of the things that support at other hosting services would do for you. It’s a trade-off between effort and control and I’m at that point where I need more of the latter. For the things I really care about, keeping the machine and network reliable, there has been nothing to report, and that is how it should be.

It’s too bad most providers price on bandwidth and storage space. I would have happily paid more per month if I could get higher availability and reliability (with the ability to run RoR - sorry Pair). Of course, with everyone claiming 99.999999% availability, it’s hard to differentiate oneself on this front so providers seem to just pile on the bandwidth/space like extra gravy hiding the bad meat.

I feel like VPSes are the future of hosting. The amount of computing power that you can cram into a 1U rack space is far more than most of us need or want to pay for. But virtualize it and divvy it up and you have a great scalable model for doing dedicated hosting. It’s probably greener too but I’ll let the hippies make a determination on that.

If you are reading this right now, then that should mean that the host migration was successful. No, I didn’t accidentally re-post my last article. I’ve moved Noodlesoft to yet another hosting provider. Noodlesoft is now on SliceHost.

Unfortunately, things didn’t work out with WebFaction so I decided to go shopping again.

While many people were recommending SliceHost, I resisted as I wasn’t relishing the idea of doing sysadmin work. I’ve done my fair share in my lifetime, including compiling kernels, setting up firewall rules and mucking with sendmail.cf files. The whole purpose of using shared hosting was that I could pay someone else to do it. Nonetheless, I found myself in need of a higher level of reliability and control. I had to bite the bullet.

I signed up and within minutes, was logged in to my slice. And oddly enough, I found myself actually enjoying setting up my server. Sure, it took some time to get everything up and running but not as long as I expected. It’s put together the way I want and if something goes wrong, it’s hard to beat the response time of going in and fixing it myself. It’s not for everybody but if you know what you are doing then I say give it a try. I would recommend paying extra for the backup feature where you can save snapshots (or have snapshots automatically taken daily or weekly). Restoring slices from these snapshots is quite easy in case you mess up or something else goes wrong. They also have articles on how to install most of the software you’ll need to get up and running. My only beef right now is that I’d like more snapshots slots available (you only get 3).

Here is the obligatory/shameless SliceHost referral link in case you are interested in signing up.

Oh, and I am keeping my DreamHost account for things such as hosting my mail. Also, since DH’s plans offer ridiculous bandwidth, it’ll be good to have as a backup when there’s some sort of burst.

And because it’s what I’ve been dealing with for the past couple weeks, you’ll probably see some sysadmin articles from me in the near future. I’ve even added a “System Administration” category. I’ll post a report on SliceHost and my experiences with VPS hosting in a month or two.

If you are reading this right now, then that should mean that the host migration was successful. You probably didn’t notice anything but Noodlesoft has switched hosting providers. It’s been a long time coming, but I’ve moved the site from Dreamhost to Webfaction (note: this is an affiliate link).

The transition had its share of bumps and glitches as the two providers do things a bit differently. Except for the forums being offline during the migration, hopefully you shouldn’t have noticed a thing. If you do find anything not quite right, please let me know.

It’s too early to tell whether this was a good move but I’ll post a report/comparison after I’ve developed a track record with the new provider. In the meantime, I’m keeping my fingers crossed that nothing blows up.