The ramblings of a developer
As many JavabeanHosting users know, there was a FTP publishing problem that prevented users from updating their site. This occurred shortly after performing a server upgrade on Friday. I wanted to take a few moments to detail what happened, the changes I made, and how things are moving to ensure that JavabeanHosting remains the best (and easiest) hosting solution for RapidWeaver.
Last week, I was excited to announce that JavabeanHosting would be receiving capacity upgrades. This included a nice jump in server capacity to make sure all sites continued to perform at their very best. I like to stay ahead of server load and make sure I'm never caught with an inadequate amount of resources to serve your sites.
The server upgrades on Friday left users unable to publish to their sites on Saturday. They found that RapidWeaver and various FTP applications started reporting errors. As soon as I saw these email reports, I immediately contacted my upstream provider about the issue. I know that my upstream provider has many users, but I became extremely frustrated that it took 4 hours to receive a response. In my mind, this is completely unacceptable. The response I got after 4 hours was:
When I user passive same issue occurred:Command:PASVError:Directory listing aborted by user Afterwards I switched the method to ACTIVE and the directory was listing without a problem.Please change the connection to ACTIVE and try again to connect via FTP. Please let us know it this works for you.
This is completely unacceptable. What worked on Thursday should have continued to work on Saturday after the upgrade. Also, I can not expect my users to understand this nor should I be expected to communicate this to everyone. No one would have liked seeing this in their inbox.
After receiving this email, it took another 4 hours to receive an update with a possible fix. Overall it took 8 hours for a FTP service disruption caused by a server upgrade to get fixed. How is this acceptable? It is, in my mind, absolutely unacceptable to me and my users.
Users may not know, but I have frequent correspondence with my upstream provider. I continually fight for services and technology that make JavabeanHosting the very best. Usually, this takes me through 2 or 3 levels of support to get anything done. I had to fight to get the easier to use Mac Panel installed and could not, for the life of me, get them to successfully install an easier and cleaner control panel installed. I had fought my last fight with them. My users deserve better and this upstream provider was preventing me from delivering on the promises I made to my users.
After a lot of research, I decided to make a switch to a new upstream provider. This new provider allows me to give my users the quality service of service that I expect. I would not offer anything that I wouldn't want to use myself.
The new control panel (yes, the one that I fought unsuccessfully to get installed at the previous provider) is now installed and running. Try it out now.
On the prior upstream provider, uptime was rated at 99.9%. New uptime is 99.99%. I still, unfortunately, need time to perform server upgrades. This will be minimal and with as much notice as possible.
A new DNS system is in place that gives users the benefit of global distribution. Your sites are now faster to respond, DDOS resistant, and fault tolerant. How does it work? It works by distributing the DNS information among 13 different servers all over the world. Now, when a user types in your domain name, the translation between the domain (example, javabeanhosting.com) and the corresponding IP is done by a server that is geographically closest to the user. This is a lot like how a CDN (content delivery network) works.
The new server has faster hard disks. This results is faster server responses to requests. For the techno-savvy, the new drives are in a Raid10 SAN configuration; fast and fully redundant.
Backups will be changing. I still plan to offer 30 days worth of backups but they will be stored in a different location. Backups are currently housed at the old upstream provider. I am going to maintain my account there until I am able to have a full set of backups at my new provider. The new backups will be hosted offsite using Amazon S3. They are securely transferred every night and stored on Amazon with AES256 encryption. This means that your backups are transmitted and stored securely in a reliable Amazon datacenter. Should this change, I will let everyone know.
Currently, users publish their files with FTP. I plan to bring FTP+SSL or SFTP access for a secure publishing option for users. Please continue to use FTP until I can fully test and deploy SFTP and/or FTP+SSL.
Nothing! Not a single coffee bean. Better service shouldn't come at a premium.
I need the flexibility and service to take care of your site. If I can't do that, I need to use a service that will let me. I feel that the migration was in everyone's best interest. I don't take migrating providers lately. It is time consuming and risky. I had about 6 hours of sleep this entire weekend but it was worth it.
Thank you for your continued support. If you have questions about the migration or are curious about more details about the technology, please let me know.
Your Barista, Greg