This morning I had a data scare. I was backing up my photos from the zoo when I noticed my laptop could no longer see my external hard drive. That happens from time to time, so I did what I always do – unplug the drive and plug it back in. Nada. Tried a different port. Nope. Tried all 4 usb ports. Still nothing. Tried unplugging the power supply and rebooting the computer. Negatory. At this point I was beginning to wonder if my external drive – the one local copy of all my photos – had bit the dust. Turns out I had managed to trip the power strip, so the drive wasn’t getting power despite being plugged in. Phew. The scare was that kick in the pants I needed to start thinking about that home data cloud again.

When I last blogged about our desire for a home data center, Domingo and I were thinking we needed a rather hefty server. We now realize that’s over kill. We thought we needed the horse power to run virtual machines (VMs), however, we can always use the server to store the VMs and run them locally on our laptops. Thus the “server” will basically be a glorified wireless file repository.

The new plan.

The idea is that the home data cloud will house all of our permanent, archival type files: media, photos, etc. Our laptop hard drives will be more like working directories. When I take the files off my camera, for example, I’ll save the unedited RAWs to the shared directory in the home cloud. I may open a few directly on my laptop to adjust the color balance, lightening, etc, and then save the finished versions back to the cloud shared directory. From a user perspective, it’ll be functionally just like using multiple folders on my computers. However, because I’m using the shared drive, Domingo can access any photo or media file he wants as well.

The purple lines show how we expect to need to manually transfer files. Mostly we’ll be pushing files to the home cloud, but occasionally pulling them down. The green lines show what’s managed automatically by crash plan – including backing up our laptop working directories, backing up the archival data to crash plans servers, and creating local mirrors of all our data (effectively raid 1). As mentioned before, crash plan data servers are expensive to recover from and a “worse case scenario” type option, e.g. fire and earthquake. Mirroring gives us chance to recover from the more typical hard drive crash without having to pay that price.

The new game plan means we can cobble together our home data cloud with existing hardware and not spend a single penny.

Today I setup the 9 year old server (unix) and my 4 or 5 year old laptop with the dead battery and flaky AC adapter (windows). I then started the tedious process of transferring the files from each unix external drive, connected to the unix server, to a windows drive connected to the windows server. When it’s all finished, I will reformat the unix external drives and the server. By giving the server a fresh copy of a modern operating system it will be much less vulnerable to malware. We’ll have about 8 TB of disk space, more than enough space for multiple copies of everything. The windows laptop will be officially retired and sent off to an e-waste recycling center.

The process is going to take a while. The server only supports USB 2.0 or 280 Mbit/s. USB 3.1 can support 10 Gbit/s, which makes it roughly 36 times faster. Domingo keeps reminding me that we can get a desktop with USB 3.1 ports for sub $300, but given that it’s not my main computer that’s slowly churning away, I think I’ll just let the process crawl along. Our files are small enough that I doubt we’ll notice a difference in speed when using the file server anyway.

Posted in Internet & Technology | Tags: Frugal Adventures, Home Automation