Take the 2-minute tour ×
Server Fault is a question and answer site for professional system and network administrators. It's 100% free, no registration required.

I'm using chef to create RAID10 arrays.

Last friday morning (10/22) I was able to create arrays with 9 EBS volumes on dev/sdz1-9 with no problems

I started having trouble attaching more than 6 EBS volumes on friday night. I put it down to EC2 problems (maybe overload) and stopped work for the weekend.

This morning the problem continues. Volumes 1-6 attach fine (e.g /dev/sdx1-6). Anything after that on the same dev major device stays in the "attaching" state forever.

I can then use /dev/sdv1-6 no problems but again, no more than 6 on that major device.

Anyone else notice behaviour like this? Am I missing something?

share|improve this question
    
Anyone?!?!? I'm wondering if scripting all these EBS create/attach requests is the cause of the problem, I'm now experimenting with sleep() between volumes to see if that helps. AWFUL!! :-( –  Declan Shanaghy Oct 26 '10 at 16:32
    
I'm not sure if this is AMI specific problem. I have even used the older Alestic Hardy AMI's for RAID0 but only with 4 EBS volumes. Hmm..Why not try the newer sets of Ubuntu Lucid or Maverick AMI's found at www.alestic.com –  rquillo Oct 27 '10 at 1:26
add comment

1 Answer

The problem is most likely not AMI specific but more the nature of cloud computing.

In theory (hehe!), you should be able to mount on /dev/sdX until /dev/sdp. And then you can also use /dev/sdX1 through /dev/sdX15.

A demo is here: http://alestic.com/2009/06/ec2-ebs-raid (He's mounting 40 EBS volumes.)

These things happen though -- e.g. yesterday, I started an instance (large) where my usual setup took 44 minutes. What usually takes 7. I shut it down, started a fresh one and was up and running within my usual 7 minutes. I run into a bad instance every so often and the only thing you can do is shutdown and start again.

It could be availability zone related too. E.g., I had way more of these weird issues in us-east-1b, than in any other.

Other than that I can't think of any other restriction to only attach 6 volumes to an instance.

It must be some sort of capacity issues. Have you posted about it on the forums? That seems to be the only place to get conclusive answers from AWS -- unless of course you subscribe to premium support.

share|improve this answer
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

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