Server Fault is a question and answer site for system and network administrators. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

On EC2, most of the larger current generation instances have multiple SSD devices. To take advantage of the combined size of these for applications like nosql databases that store large amounts of data, I've been combining them with software raid 0 or 10 using mdadm.

It takes forever when launching new instances though, for instance to run the initial raid10 sync on an i2.4xlarge instance takes several hours to complete.

Is there any way to speed this up? Is it possible to save an AMI after running the initial mdadm sync that will include all the ephemeral disks in the same raid configuration when it lauches?

I haven't used custom AMI's much but from reading the docs it sounds like both instance store backed and ebs backed AMI's only include a snapshot of the root volume.

share|improve this question

closed as off-topic by Frands Hansen, mdpc, MadHatter, Nathan C, Ward May 10 '14 at 17:13

This question appears to be off-topic. The users who voted to close gave this specific reason:

  • "Questions must be relevant to professional system administration. Server Fault is dedicated to professional system and network administrators. End user and enthusiast questions are off-topic (contact your system administrator or hire a professional to help you out). Please see the Help Center for more information." – mdpc, MadHatter, Nathan C, Ward
If this question can be reworded to fit the rules in the help center, please edit the question.

You don't want the ephemeral devices be part of your mdadm-managed array. Using ZFS -- with ephemeral device(s) being caches -- should work much better for you. See… – Mikhail T. Mar 25 '15 at 20:21
@MikhailT. thanks for the link, I'll definitely look into this configuration using ZFS if I run into this issue again – danny Mar 26 '15 at 18:16
Don't think anyone will see this at this point, but can someone elaborate on why this was closed as "not relevant to professional systems administration"? It was an issue I ran into while being paid as a devops engineer/sysadmin, and the database I was setting up is a core part of system that now processes billions of dollars a year in payments. I asked a well-defined question and got a prompt answer that I accepted, why all the haters? – danny Mar 26 '15 at 18:28
Is AWS itself off limits here? – danny Mar 26 '15 at 18:32
up vote 0 down vote accepted

As you've noticed, ephemeral disks are not included in an AMI snapshot; only the root volume is. Because of this it's not possible to "snapshot" the metadata of mdadm because the ephemeral disks will be lost when the instance hosting it is destroyed. You can however create persistent RAIDs with EBS volumes.

share|improve this answer
Thanks, good to confirm I wasn't just missing something. – danny May 12 '14 at 15:15

Not the answer you're looking for? Browse other questions tagged or ask your own question.