How to change the user root to something like ubuntu when executing fabric.
I've been trying this command:

$ aws do demo-server --user=ubuntu get_syslog

INFO: Instance 'demo-server' available
run: tail /var/log/syslog
out: Please login as the ubuntu user rather than root user.

Cheers,
rodney

if we see inside aws.conf ,
server_user = ubuntu

but when i run the : aws ssh demo-webserver
its say please use users rather than root ... im not sure is it a problem/bugs that u guys should look into it ..

btw ... mr.awsome is REALLY AWESOME !! :)

when trying to connect to a jail

ERROR: Failed to connect to ez-instance:i_eena_me (10.0.0.2)
ERROR: username: 'root'
ERROR: port: 22
ERROR: sock: <paramiko.proxy.ProxyCommand object at 0x7f2a80a6bed0>
ERROR: Couldn't validate fingerprint for ssh connection.
ERROR: Private key file is encrypted
ERROR: Is the instance finished starting up?

or configure it:

PLAY [jailhost-i_eena_me] ***************************************************** 

GATHERING FACTS *************************************************************** 
ERROR: Failed to connect to ez-instance:i_eena_me (10.0.0.2)
ERROR: username: 'root'
ERROR: port: 22
ERROR: sock: <paramiko.proxy.ProxyCommand object at 0x7f46fb3ef6d0>
fatal: [jailhost-i_eena_me] => Couldn't validate fingerprint for 'ez-instance:i_eena_me': Private key file is encrypted

TASK: [infopro-digital.apache | APACHE | Include OS-specific variables] ******* 
FATAL: no hosts matched or all hosts have already failed -- aborting

paramiko fails with encrypted keys, because it doesn't forward the ssh agent settings.

This is something btw is something ansible considered: ansible/ansible#4100 but then declined, because their paramiko support is deprecated.

Since updating to the latest mr.awsome 0.6, now when I run the ./bin/aws start command, the following error occurs. Do I need to pass in another argument besides the server config that I want to use?

{{{
$ ./bin/aws start test
INFO: Instance 'test' unavailable
INFO: Creating instance 'test'
INFO: Instance created, waiting until it's available
...
Traceback (most recent call last):
File "./bin/aws", line 16, in
mr.awsome.aws(configpath="/Users/nateaune/code/mrawsome.buildout/etc/deployment")
File "/Users/nateaune/.buildout/eggs/mr.awsome-0.6-py2.6.egg/mr/awsome/init.py", line 652, in aws
return aws(argv)
File "/Users/nateaune/.buildout/eggs/mr.awsome-0.6-py2.6.egg/mr/awsome/init.py", line 646, in call
args.func(sub_argv, args.func.doc)
File "/Users/nateaune/.buildout/eggs/mr.awsome-0.6-py2.6.egg/mr/awsome/init.py", line 444, in cmd_start
self.cmd_status()
TypeError: cmd_status() takes exactly 3 arguments (1 given)
}}}

I can provide a user via the -l flag to ssh, but passing a keyfile using -i is ignored. If I name my keyfile id_rsa, then it works quite nicely due to how the ssh package looks for keys, but working with AWS keypairs this is not so helpful.

It would be nice if it were possible to have command-line arguments to assh passed inward.

we need to find a way to support multiple masters for an instance w/o breaking augmentation of instances (fschulze@c2f6b30#commitcomment-10333109)

https://github.com/ployground/ploy links to http://ploy.readthedocs.com near the top, but clicking this currently results in the Readthedocs "Maze Found" 404 page. Cheers.

im getting error while choosing 64bit images in aws.conf, inside aws.conf didnt specify architecture right? just the AMI ID ...

im not sure either me or others got this issue/problem ...

when calling assh with -c to specify a non-default config file from the command line it seems to get passed through to ssh because the message reads:

assh -c etc/foo.conf bar
Unknown cipher type 'etc/foo.conf'

In the documentation the two lines after entry-points do not appear as indented:

entry-points =
aws=mr.awsome:aws
assh=mr.awsome:aws_ssh

And if you don't indent these lines, then the recipe doesn't work, and no aws or assh scripts are created in the /bin directory.

Also, the two tcp lines are not indented:

connections =
tcp 22 22 0.0.0.0/0
tcp 80 80 0.0.0.0/0

In case anyone is struggling to get mr.awsome running with boto-2.0rc1; that doesn't work. It's possible to start instances but not to terminate or otherwise get info on a running instance.

I don't who's to blame (boto or mr.awsome) :)

To be more precise, after some investigating and source code reading, it seems it is looking into $PWD/etc/ploy.conf as shown below:

1016 [10:36] root@centre:sysutils/py-ploy_ansible# ploy  status www
                                          ^^^^^^^^^^^^^^^^^^^^^^
ERROR: Config '/usr/ports/sysutils/py-ploy_ansible/etc/ploy.conf' doesn't exist.
                                           ^^^^^^^^^^^^^^^^^^^^^^^

I'm not sure that's what we really want or it should be written somewhere (did I miss anything?).

http://web.monkeysphere.info

Docs on pypi mention a "create" command which apparently doesn't exist.
By experimenting I discovered that "start" creates the instance if it doesn't exist.

I'm using the latest 0.8 release, and when I try to run the "do" command to execute a fabric script, the following error occurs:

$ ./bin/aws do plone3-small-production get_syslog
/Users/nateaune/.buildout/eggs/pycrypto-2.0.1-py2.6-macosx-10.6-universal.egg/Crypto
/Hash/SHA.py:6: DeprecationWarning: the sha module is deprecated; use the hashlib module     instead
  import warnings
/Users/nateaune/.buildout/eggs/pycrypto-2.0.1-py2.6-macosx-10.6-universal.egg/Crypto    /Hash/MD5.py:6: DeprecationWarning: the md5 module is deprecated; use hashlib instead
  import warnings

Fatal error: Couldn't find any fabfiles!

Aborting.

Here is what my aws.conf file looks like:

[macro:base-instance]
keypair = gsg-keypair
region = us-east-1
placement = us-east-1a
fabfile = `fabfile.py`_

[instance:plone3-small-production]
<= macro:base-instance
securitygroups = plone3-small-production
instance_type = m1.small 
image = ami-df947ab6

And there is a fabfile.py in the etc/deployment directory. What am I missing? Do I need to specify an explicit path for the fabfile, or will mr.awsome be able to find it if it's in the etc/deployment dir?

One less dependency in the FreeBSD port

<3

Hi,
Thanks for updating mr.awsome, I was secretly hoping for that a long time.

I still have an issue when starting a stopped instance. I'm working on Mac OS X (Lion).

Below is my error. Anybody any idea?

kind regards!

INFO: Instance 'productie-2' available.
INFO: Instance state: stopped
INFO: Instance already started, waiting until it's available
Traceback (most recent call last):
File "./bin/aws", line 18, in
mr.awsome.aws(configpath = "/Users/bowlby/Edubroker/newawsome/etc")
File "/Users/bowlby/Edubroker/newawsome/eggs/mr.awsome-0.10-py2.7.egg/mr/awsome/init.py", line 451, in aws
return aws(argv)
File "/Users/bowlby/Edubroker/newawsome/eggs/mr.awsome-0.10-py2.7.egg/mr/awsome/init.py", line 445, in call
args.func(sub_argv, args.func.doc)
File "/Users/bowlby/Edubroker/newawsome/eggs/mr.awsome-0.10-py2.7.egg/mr/awsome/init.py", line 158, in cmd_start
instance = server.start(overrides)
File "/Users/bowlby/Edubroker/newawsome/eggs/mr.awsome-0.10-py2.7.egg/mr/awsome/ec2.py", line 212, in start
log.error("Something went wrong, instance status: %s", instance.status)
AttributeError: 'Instance' object has no attribute 'status'

There is a small documentation error when you type: ./bin/aws
See the changeset in this fork:
http://github.com/natea/mr.awsome/commit/191ba22140897cc27f8b93ed81bdbe5ea65c922d

If the Amazon account has more than the default 5 Elastic IP addresses, it could be useful to have mr.awsome automatically create a new Elastic IP and associate it with the new EC2 instance, rather than needing to hardcode these IPs beforehand in the .conf file.

When starting an instance,

mr.awsome should first check if it will match the requirements like volume availability / ip prewsent in the pool (and by extension to my previous pull requests) if the load balancer exists in the same zone.

If not, then mr.awsome should prevent the instance to start unless a config option "force_start" (or something like that) is specified in the conf.

The current documentation is missing instructions for how to attach EBS volumes, assign elastic IP, choose an instance type other than m1.small and make a snapshot from an attached volume.

in the aws.conf:
instance_type = m1.large
ip = x.x.x.x
volumes = vol-xxxxx /dev/sdf

to make a snapshot:
./bin/aws snapshot demo-server

The code also suggests that there is a --revision option that would let you install a particular revision of your buildout, or is this just carried over from mr.developer?

Reminder: http://plope.com/Members/chrism/documentation_is_the_differentiator

Some people find ezjail more of an hindrance to manage upgrades compared to "plain jails". There is also the fact that ezjail supports only one base jail which makes mixing versions and/or archs more difficult. It would be nice if ploy could work with plain jails too.

It would be really useful to be able to create new EBS volumes from an existing EBS snapshot.

The aws.conf file could have something like this:

[instance:demo-server]
keypair = default
securitygroups = demo-server
region = eu-west-1
placement = eu-west-1a
image = ami-a62a01d2
snapshot = snap-2d8cd144 /dev/sdf
startup_script = startup-demo-server
fabfile = `fabfile.py`_

When you launch the instance demo-server, it would then create a new EBS volume from the snapshot "snap-2d8cd144" and attach it to the device /dev/sdf.

The info returned from the aws script would look something like this:
INFO: Instance running.
INFO: Instances DNS name ec2-184-73-240-104.compute-1.amazonaws.com
INFO: Instances public DNS name ec2-184-73-240-104.compute-1.amazonaws.com
INFO: Instances volume ID vol-58fc3031 created from snap-2d8cd144

How would I install a linux VM on virtualbox with ploy using bootmedia with custom kickstart or preseed file (debian)?

The fabfile integration seems to be broken for EC2 instances, the fabfile is run with the name of the instance from aws.conf rather than the public_dns_name of the running server. I think this was broken in d671f7f.

if the user does not have ~/.ssh/config any operation that uses plain.Instance will break with a nasty stack trace

When I run the ./bin/aws start command, it mentions that one can now override parameters in the startup script with the -o or --overrides option, but the usage of this is not documented in the README.

Is this the correct syntax, if I want to override a ${svn_url} in my startup-script?

$ ./bin/aws start rhaptos-empty-32-dev -o svn_url=https://software.cnx.rice.edu/svn/devsets/express-edit-reuse/cnx

And then in my startup-script I have this line:

Now Amazon added support to upload your personal SSH public key, rather than needing to use an Amazon generated keypair. http://alestic.com/2010/10/ec2-ssh-keys
Do you have any plans to add this functionality to mr.awsome?

When I try to ssh into a newly created instance with the "assh" command, the following error occurs:

$ ./bin/assh plone3-small-production
/Users/nateaune/.buildout/eggs/pycrypto-2.0.1-py2.6-macosx-10.6-universal.egg/Crypto/Hash/SHA.py:6: DeprecationWarning: the sha module is deprecated; use the hashlib module instead
  import warnings
/Users/nateaune/.buildout/eggs/pycrypto-2.0.1-py2.6-macosx-10.6-universal.egg/Crypto/Hash/MD5.py:6: DeprecationWarning: the md5 module is deprecated; use hashlib instead
  import warnings
INFO: Instance 'plone3-small-production' available
ERROR: Couldn't validate fingerprint for ssh connection.
ERROR: Authentication failed.
ERROR: Is the server finished starting up?

Do I need to specify a user or key somewhere in order for the ssh command to work?

Normally to ssh to the into the machine with this command, which works fine:

ssh -i ~/id_rsa-gsg-keypair [email protected]

Here's an excerpt of my aws.conf:

[aws]

These files must exist and contain your EC2 API keys

access-key-id=/.aws/access-key-id
secret-access-key=/.aws/secret-access-key
region = us-east-1

[securitygroup:openblock-testing]
description = OpenBlock Test Servers
connections =
tcp 22 22 0.0.0.0/0
tcp 80 80 0.0.0.0/0
tcp 8000 8000 0.0.0.0/0

[ macro:openblock-base]
keypair = openblock
securitygroups = openblock-testing
region = us-east-1
placement = us-east-1d
instance_type = t1.micro
server_user = ubuntu

[instance:lucid-64]
<= macro:openblock-base

lucid = ubuntu 10.04

image = ami-63be790a

This works at least enough that I can create an instance from scratch:

$ aws -c aws.conf start lucid-64
INFO: Instance 'lucid-64' unavailable
INFO: Creating instance 'lucid-64'
INFO: Instance created, waiting until it's available
....

But if I try to use assh to connect, no luck:
$ assh lucid-64
INFO: Instance 'lucid-64' unavailable
ERROR: Can't establish ssh connection.
Traceback (most recent call last):
File "/home/pw/builds/openblock/builds/20110519/bin/assh", line 9, in
load_entry_point('mr.awsome==0.9', 'console_scripts', 'assh')()
File "/home/pw/builds/openblock/builds/20110519/lib/python2.6/site-packages/mr/awsome/init.py", line 814, in aws_ssh
return aws(argv)
File "/home/pw/builds/openblock/builds/20110519/lib/python2.6/site-packages/mr/awsome/init.py", line 802, in call
args.func(sub_argv, args.func.doc)
File "/home/pw/builds/openblock/builds/20110519/lib/python2.6/site-packages/mr/awsome/init.py", line 715, in cmd_ssh
user, host, port, client, known_hosts = server.init_ssh_key()
TypeError: 'NoneType' object is not iterable

If I use ssh and explicitly specify my pem file, I can get in, so the server is definitely running:
$ assh lucid-64
INFO: Instance 'lucid-64' unavailable
ERROR: Can't establish ssh connection.
Traceback (most recent call last):
File "/home/pw/builds/openblock/builds/20110519/bin/assh", line 9, in
load_entry_point('mr.awsome==0.9', 'console_scripts', 'assh')()
File "/home/pw/builds/openblock/builds/20110519/lib/python2.6/site-packages/mr/awsome/init.py", line 814, in aws_ssh
return aws(argv)
File "/home/pw/builds/openblock/builds/20110519/lib/python2.6/site-packages/mr/awsome/init.py", line 802, in call
args.func(sub_argv, args.func.doc)
File "/home/pw/builds/openblock/builds/20110519/lib/python2.6/site-packages/mr/awsome/init.py", line 715, in cmd_ssh
user, host, port, client, known_hosts = server.init_ssh_key()
TypeError: 'NoneType' object is not iterable

ploy is now no longer AWS only (split into ploy_ec2)

Rename description accordingly.

Ideally less than 80 chars for fitting into short packaging descriptions

Thanks for a great project!
According to the docs:

fingerprint
(required) The ssh fingerprint of the server. If set to ask then manual interactive verification is enabled. If set to ignore then no verification is performed at all! You can also point this to a public ssh host key file to let the fingerprint be extracted automatically.

After setting the following in ploy.conf:

fingerprint = /path/to/my/key.pub

Then running ploy ssh jailhost results in the following error:

…
ERROR: Couldn't validate fingerprint for ssh connection.
ERROR: Fingerprint doesn't match for [x.x.x.x] (got xx:xx:xx:xx:xx:xx:xx:xx:xx:xx:xx:xx:xx:xx:xx:xx, expected SHA256:xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx)
ERROR: Is the instance finished starting up?

How can one change the type of fingerprinting of the public key from SHA256?

Also according to the docs:

ssh-key-filename
Location of private ssh key to use.

However, according to this comment, ssh-key-filename should be used with a public key?
ployground/bsdploy#24 (comment)
Following this advice results in another error:

ERROR: not a valid EC private key file

I understand that this error might be due to another bug in Paramiko:
paramiko/paramiko#521