timhodson.cominventomatic

After some rather unsettling moments when everything in RAM kept running, but the root filesystem quietly disappeared, I decided to clone the root partition onto a second drive and boot from that. On investigation it looked like the first hard drive – /dev/sda –  had been hitting a max temprature of over 130°C.

The thing that was really puzzling me was how to tell Grub 2 that it should use /dev/sdc6 as the new root partition. I followed several sets of instructions, but no matter what I did, it always chose /dev/sda6 as the root partition.

I then tried using the excellent boot-repair disk to see if that could do what I wanted.  I ticked the option I wanted which was to use /dev/sdc6 a the default boot partition.  I applied the changes and still it booted from the wrong partition.  boot-repair sends a clear and well thought-out report to paste.ubuntu.com, and while looking through this I happened to note that the cloned partition had the same UUID (Universally Unique IDentifier) as the original partition.

light bulb!

I figured that this must be causing the confusion, so a quick google pointed me to this post by Paul Goscicki, which confirmed that it was a likely cause.  So on running tune2fs -U random /dev/sdc6 , and then re-running boot-repair, I now have a system booting from the correct partition.

UUIDs are obviously not always UUIDs!

Since the beginning of this week, I am officially now working for Kasabi.com.  My role here is to find ways to make the adoption of Kasabi.com’s suite of datasets and easy to consume APIs even easier for developers who need a quick way to get at the data that powers their apps.

You probably already know that the core data hosting and ready-provisioned APIs of Kasabi datasets allow you to get a data backed application out of the door much more quickly than you would if you had to incorporate data hosting and access infrastructure within your project.

We want to concentrate on helping you build something fantastic that is compelling for your user groups.  We’re going to be doing this in several ways; by improving our documentation; by holding webinars to demonstrate key features; by running hack days to give you space to play and learn how to effectively work with the Kasabi APIs in order to get the data you want.

This is an exciting time.  We are building something that feels like its moment has come.