[FFe] New package: ec2-init

Bug #269434 reported by Soren Hansen
6
Affects Status Importance Assigned to Milestone
Ubuntu
Fix Released
Undecided
Unassigned

Bug Description

Hi. I'd like to request permission to upload ec2-init.

Description: Init scripts for EC2 instances
 EC2 instances need special scripts to run during initialisation
 to get hold of ssh keys and to let the user run various scripts.

It can be seen in

    https://launchpad.net/~ubuntu-ec2/+archive

 affects ubuntu
 subscribe motu-release

Related branches

Revision history for this message
Scott Kitterman (kitterman) wrote : Re: [Bug 269434] [NEW] [FFe] New package: ec2-init

Ack from me.

Revision history for this message
Cesare Tirabassi (norsetto) wrote :

Please, excuse my ignorance, but is this of potential benefit to Ubuntu users at large or just to be used for some obscure purpose by the Evil Empire (tm) ?

Please, also allow me to make a couple of observations:

1, Wouldn't be good packaging practise to depend on procps?
2. Shouldn't we set variables in the init script before using them?
3. Security-wise, are we 100% happy to fetch data from an "anonymous" internet source and use it to change (possibly sensitive) user data on his computer? (as a side remark, perhaps "retrieve and install" would be better instead of "get hold")

Revision history for this message
Soren Hansen (soren) wrote : Re: [Bug 269434] Re: [FFe] New package: ec2-init

On Sun, Sep 14, 2008 at 07:06:00PM -0000, Cesare Tirabassi wrote:
> Please, excuse my ignorance, but is this of potential benefit to Ubuntu
> users at large or just to be used for some obscure purpose by the Evil
> Empire (tm) ?

I'm really not sure what you mean or are trying to imply.

This package provides some glue code needed for generally available EC2
instances. Does that answer your question?

> Please, also allow me to make a couple of observations:
>
> 1, Wouldn't be good packaging practise to depend on procps?

Indeed it would. Thanks.

> 2. Shouldn't we set variables in the init script before using them?

Presuming you're referring to the $DAEMON thing, yes, I already fixed
that locally.

> 3. Security-wise, are we 100% happy to fetch data from an "anonymous"
> internet source and use it to change (possibly sensitive) user data on
> his computer?

I didn't design EC2. This is simply an new implementation of what AFAIK
pretty much every other generally available AMI does during startup.
Also, 169.254.169.254 is a link-local address (as per RFC 3927), so it's
not an "internet source".

> (as a side remark, perhaps "retrieve and install" would be better
> instead of "get hold")

Ok.

Revision history for this message
Cesare Tirabassi (norsetto) wrote :

>I'm really not sure what you mean or are trying to imply.

As I said, my question stems from ignorance from my side. I didn't know what ec2 was (had to google that up), and I still don't know how exactly this package fits into the archive. I don't see any other ec2 related packages in the archive (please correct me if I'm wrong) so I was just wondering if this package was meant to cover a specific need highlighted for instance by the server team and, as a side effect, was of possible broader use or not.
Or if there was already an existing package and what the possible relations were; I only find references to rpm packages in the amazon documentation, so, I was wondering if it makes sense to have this package alone in the archive without a corresponding ec2-ami-tools, ec2-api-tools etc. packages.
What I'm trying to understand is if this package as is really pertains to universe (yes, it can be potentially useful to a broad userbase, and no, it doesn't break anything else)?

Thanks for considering my comments (please also note $NAME in the init file).

Revision history for this message
Soren Hansen (soren) wrote :

On Mon, Sep 15, 2008 at 09:48:31AM -0000, Cesare Tirabassi wrote:
>>I'm really not sure what you mean or are trying to imply.
> As I said, my question stems from ignorance from my side. I didn't
> know what ec2 was (had to google that up),

It's Amazon's Elastic Compute Cloud.

   http://en.wikipedia.org/wiki/Amazon_Elastic_Compute_Cloud

> and I still don't know how exactly this package fits into the archive.

..and I'm afraid I still don't understand the question :) It's a package
that provides "Init scripts for EC2 instances", because "EC2 instances
need special scripts to run during initialisation" to fetch login
credentials and possibly do some extra instance specific initialisation.

> I don't see any other ec2 related packages in the archive (please
> correct me if I'm wrong)

python-boto - Python interface to Amazon's Web Services

Debian also has (I think it just missed the DIF):
libnet-amazon-ec2-perl - Perl interface to the Amazon Elastic Compute Cloud (EC2)

I'm not sure what the relevance of this is, though?

> so I was just wondering if this package was meant to cover a specific
> need highlighted for instance by the server team and, as a side
> effect, was of possible broader use or not.

It specifically makes Ubuntu usable as general purpose AMI's for EC2. I
don't know if that counts as broad use. I think so, though.

> Or if there was already an existing package and what the possible
> relations were;

vm-builder (upstream) has code to build and deploy EC2 instances, but
needs ec2-init for it to be really useful, so when ec2-init reaches the
archive, I'll be filing an FFe for that functionality. vm-builder will
be using python-boto to do the actual deployment. In the process of
building EC2 instances, I'll be using the Xen kernels as well.

> I only find references to rpm packages in the amazon documentation,
> so, I was wondering if it makes sense to have this package alone in
> the archive without a corresponding ec2-ami-tools

There's work afoot to either provide a replacement of ec2-ami-tools or
to somehow offer ec2-ami-tools in some form. Even without ec2-ami-tools,
ec2-init still fills a gap.

> ec2-api-tools etc.

python-boto replaces ec2-api-tools.

> What I'm trying to understand is if this package as is really pertains
> to universe (yes, it can be potentially useful to a broad userbase,
> and no, it doesn't break anything else)?

I obviously think it does, otherwise I wouldn't have filed an FFe for it
:)

> Thanks for considering my comments (please also note $NAME in the init
> file).

Good catch, thanks.

--
Soren Hansen |
Virtualisation specialist | Ubuntu Server Team
Canonical Ltd. | http://www.ubuntu.com/

Revision history for this message
Cesare Tirabassi (norsetto) wrote :

Thanks, with the given explanations I'm happy to confirm the FFe.

Revision history for this message
Luca Falavigna (dktrkranz) wrote :

Accepted into archives.

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.