Everyone, I wanted to put a post together so everyone can have a good explanation as to why HLS has been having issues as of late and what we are doing about it.
Firstly though, I would like to personally apologize for the headaches. While the actual issue is out of my hands, I made the final decision to put us on Dreamhost (our hosting provider) and this ended up being an awful move in the long run. I can assure you that I am looking to move us ASAP.
Now, here’s what’s going down:
Once MQ resigned from HLS, I needed to move the server to a new host, and more importantly away from GoDaddy due to SOPA BS and general suckage in general. I had moved other sites that I run to Dreamhost after a recommendation and had been having no issues via their shared hosting. Granted, these sites were mostly dead, but I figured there shouldn’t be large issues.
Considering the size of HLS, I needed to move off of shared hosted and upgraded to a VPS (Virtual Private Server). This would allow HLS to have the needed bandwidth and resources to handle our traffic needs, plus allow us room to add resources as we grew.
Then came the issues.
I figured it was initially on my end as high traffic periods tended to bring us down. No big deal, that I knew how to rectify. After 3.0 launched, I immediately went to work on caching plugins for the site. This would allow for our server to store a static copy of our posts instead of having to regenerate the same copy over and over again which was crippling us. On top of that, I added HLS to Cloudflare for additional caching support as well as a CDN for quicker loading times and less server strain.
It seemed to work just fine for posts that happened in the morning, but if we had to update or post new content during the day when we were getting traffic, everything went to hell.
For caching to work properly, those static pages have to be dumped for the new ones. This shouldn’t be a large server strain; however, Dreamhost simply couldn’t handle it. We would receive absurdly large server spikes no matter what plugins our blog had enabled or how I tweaked the caching settings. I even went into our server to do some custom configuration to no avail. I even upgraded our memory to larger than needed resources to counteract this issue, but that failed as well.
tl;dr — When we post new content on the HLS blog, we crash because Dreamhost sucks.
Since the only support I’ve ever gotten from them on this issue is “buy more memory”, I’m done dealing with them. I started doing research on new hosting providers last week, but this weekend was my breaking point as it is hard to enjoy a weekend out at ND when my phone is blowing up with emails and twitter mentions that HLS crashed yet again.
I am hoping to move us during the bye weekend, but I want to make sure that our new host is solid and can properly support our needs.
Again, I can’t apologize enough and appreciate everyone still sticking with us despite this headache. I promise it will be corrected and HLS will run better than ever.