Comments on: Fedora 12 Bootable Root EBS on EC2

By: Frannack

Frannack — Thu, 02 Sep 2010 11:57:35 +0000

@Tim

yum -y install cpipe

By: Installing Cent OS 5.5 on EC2 with the Cent OS 5.5 Kernel

Installing Cent OS 5.5 on EC2 with the Cent OS 5.5 Kernel — Tue, 10 Aug 2010 10:51:53 +0000

[...] follows is similar, both steps and concepts, to the "from scratch" section of my post on Fedora 12 on EC2 using a root EBS. I've also bundled all the instance building commands up into one script (centos5.5.sh). If [...]

By: Olivier Delrieu

Olivier Delrieu — Tue, 10 Aug 2010 07:03:55 +0000

@Bryan.
Same problem here.

I've been running a cluster of home made F11 instances for a year. Instances are S3 based, and were created using a script very similar to Carson's.

The same script, apply to F13, produces bootable F13 instances (console output shows couple of problems with udev, requiring adding MAKEDEVs in rc.sysinit).

However when launching a new instance created from an ec2-bundle-vol of the instance, the new instance self terminates (Client.InstanceInitiatedShutdown: Instance initiated shutdown) , the console output is empty, and IPs are not attributed. Note this worked well (and still works well) with the F11 instances.

As my F11 and F13 scripts are identical, and because they still work with F11, I believe this is due to changes made in F12 & F13. But, as the console output is empty I cannot figure out when the problem is. I've tried using a F11 rc.sysinit in F13, it made no difference.

Any idea?

By: carson

carson — Sat, 07 Aug 2010 11:41:13 +0000

@Bryan It should all work as described here although a few things have changed recently. I'll be creating a new post soon on how to use some of the new tools.

By: Bryan Field-Elliot

Bryan Field-Elliot — Fri, 30 Jul 2010 22:39:03 +0000

I followed these instructions and am very pleased to be running FC12 now on EC2! However, I am having trouble trying to launch a new instance created from an ec2-bundle-vol of the instance I ran these instructions agains. ec2-bundle-vol succeeds, as does ec2-upload-bundle and ec2-register. But when I launch the instance, it's stuck in "Pending" for several minutes, and then goes immediately to "Terminated". I have no idea why, there is no console output and nothing I can retrieve from EC2 seems to offer any assistance. Has anyone else seen this? Do I need to use non-default options for ec2-bundle-vol or ec2-register? I'm just trying to launch it on ephemeral storage, not ECB.

By: AWS User

AWS User — Wed, 14 Jul 2010 21:00:52 +0000

Hi. I was very excited about your script and tried it out on my AWS account. However, I get the following message. The only change I made was that i changed the device from /dev/sdh to /dev/sdc. I also tried changing everything to 13 for Fedora 13 and that did not work either – I saw the same errors. Thanks very much for any help with this!

…lots of rpmlib errors…
rpmlib(PayloadIsXz) is needed by glibc-common-2.11.2-1.i686
rpmlib(FileDigests) is needed by grep-2.6.3-1.fc12.i686
rpmlib(PayloadIsXz) is needed by grep-2.6.3-1.fc12.i686
Complete!
(1, [u'Please report this error in http://yum.baseurl.org/report')
365 package files removed
createfedora12bootebs.sh: line 50: /mnt/etc/ssh/sshd_config: No such file or directory
createfedora12bootebs.sh: line 51: /mnt/etc/ssh/sshd_config: No such file or directory
cp: cannot create regular file `/mnt/etc/sysconfig/network': No such file or directory
cp: cannot create regular file `/mnt/etc/sysconfig/network-scripts/ifcfg-eth0': No such file or directory
cp: target `/mnt/usr/local/sbin/' is not a directory
cp: cannot create directory `/mnt/lib/modules/': No such file or directory
mv: cannot stat `/mnt/lib/tls': No such file or directory
createfedora12bootebs.sh: line 84: /mnt/etc/ld.so.conf.d/kernelcap-2.6.21.7-2.fc8.conf: No such file or directory
chroot: cannot run command `chkconfig': No such file or directory
chroot: cannot run command `chkconfig': No such file or directory

By: Lingji Chen

Lingji Chen — Sat, 22 May 2010 14:34:47 +0000

Thanks a lot, Carson!

I followed the instructions and got an instance running. The first thing I did was of course to issue the command "uname -a" and expect to see the new kernel version, but it reports "2.6.21.7-2.fc8xen". I kept thinking that I did something wrong; two comments above asked basically the same question (e.g., "what kernel version is that AKI/ARI supposed to be?") but the answers did not explain.

In any case, when I do something like "yum install octave," the FC12 packages are fetched, so IT IS FC12 after all, even though it is not reported so by uname. Anybody cares to enlighten us on this?

By: Ernest Mueller

Ernest Mueller — Mon, 05 Apr 2010 14:17:51 +0000

Hey, I'm having a problem with my instances built this way.

Every time I reboot, /dev/null goes to permissions 600.

udev is set up to do the right thing…
/lib/udev/rules.d/50-udev-default.rules:KERNEL=="null|zero|full|random|urandom", MODE="0666"

But I saw some posts about udev + Ubuntu kernels on ec2 not running?

I rm -f /dev/null;mknod -m 0666 /dev/null c 1 3, and then it's OK, but once I reboot it's bad again. Any ideas?

By: carson

carson — Thu, 01 Apr 2010 00:51:34 +0000

I'm not sure if you could get the modules working under Fedora or not, you could give it a try and see if it boots. Someone mentioned that there is work going on to make the Fedora 13 kernel and ramdisk available on EC2 so that may be worth waiting for.

By: Jeffrey W. Baker

Jeffrey W. Baker — Mon, 29 Mar 2010 18:05:41 +0000

Carson, thanks for the great tips. I made an fc12 x86_64 AMI based on your page but the problem with using the Ubuntu kernel is that you don't have the Ubuntu kernel modules on disk. There's no way to load md, for instance, or xfs. Any way around that?