Moving the servers from local datacenter to AWS EC2


#1

hello everyone,
I am starting to gather notes to move my servers from a local DC to AWS.
Currently in-order to set the http & https proxies, we use a bootstrap
template. Now that we are moving to the AWS, i assume we will not be
requiring to set this proxies and hence the template can be left aside.

The normal “knife ec2 server create” bootstraps the node/instance. To
install a specific version of chef-client, i think i should be able to pass
the option : --bootstrap-version VERSION
( https://docs.chef.io/plugin_knife_ec2.html )

So apart from specifying a particular chef-client version to be installed,
each time i bootstrap the servers so that they remain consistent across all
nodes. is there any other reason that i should have a bootstrap template.

I also came accross ths blog:

why is it important to have the client.rb, validator.pem on the S3/anywhere
when the same info can be passed from the knife config file itself.


#2

Launch your instances with autoscaling groups or another orchestration
layer and they can self-register with your Chef server using the
config/validator from S3.

On Tue, Apr 28, 2015 at 4:25 PM, niristotle okram nirish.okram@gmail.com
wrote:

hello everyone,
I am starting to gather notes to move my servers from a local DC to AWS.
Currently in-order to set the http & https proxies, we use a bootstrap
template. Now that we are moving to the AWS, i assume we will not be
requiring to set this proxies and hence the template can be left aside.

The normal “knife ec2 server create” bootstraps the node/instance. To
install a specific version of chef-client, i think i should be able to pass
the option : --bootstrap-version VERSION
( https://docs.chef.io/plugin_knife_ec2.html )

So apart from specifying a particular chef-client version to be installed,
each time i bootstrap the servers so that they remain consistent across all
nodes. is there any other reason that i should have a bootstrap template.

I also came accross ths blog:
http://awsadvent.tumblr.com/post/37773106407/bootstrap-cfg-mgmt-aws

why is it important to have the client.rb, validator.pem on the
S3/anywhere when the same info can be passed from the knife config file
itself.


#3

hey Dan, i am just starting on AWS. Could you point me to a doc or
something for “Launch your instances with autoscaling groups or another
orchestration layer”

On Tue, Apr 28, 2015 at 1:41 PM, Daniel Condomitti daniel@condomitti.com
wrote:

Launch your instances with autoscaling groups or another orchestration
layer and they can self-register with your Chef server using the
config/validator from S3.

On Tue, Apr 28, 2015 at 4:25 PM, niristotle okram nirish.okram@gmail.com
wrote:

hello everyone,
I am starting to gather notes to move my servers from a local DC to AWS.
Currently in-order to set the http & https proxies, we use a bootstrap
template. Now that we are moving to the AWS, i assume we will not be
requiring to set this proxies and hence the template can be left aside.

The normal “knife ec2 server create” bootstraps the node/instance. To
install a specific version of chef-client, i think i should be able to pass
the option : --bootstrap-version VERSION
( https://docs.chef.io/plugin_knife_ec2.html )

So apart from specifying a particular chef-client version to be
installed, each time i bootstrap the servers so that they remain consistent
across all nodes. is there any other reason that i should have a bootstrap
template.

I also came accross ths blog:
http://awsadvent.tumblr.com/post/37773106407/bootstrap-cfg-mgmt-aws

why is it important to have the client.rb, validator.pem on the
S3/anywhere when the same info can be passed from the knife config file
itself.


Regards
nirish okram


#4

Hi Nirish -

There’s some good documentation for getting started with AWS
autoscaling here:
http://docs.aws.amazon.com/AutoScaling/latest/DeveloperGuide/GettingStartedTutorial.html

We are using packer (http://packer.io) to build an AMI that’s (almost)
ready to bootstrap on it’s own, then we deploy/bootstrap using
terraform (http://terraform.io) and that has been working out pretty
well for us. We do not currently use autoscaling groups, although I
expect we will need them sooner than later.

Cheers,
-Camden

On Tue, Apr 28, 2015 at 5:01 PM, niristotle okram
nirish.okram@gmail.com wrote:

hey Dan, i am just starting on AWS. Could you point me to a doc or something for “Launch your instances with autoscaling groups or another orchestration layer”

On Tue, Apr 28, 2015 at 1:41 PM, Daniel Condomitti daniel@condomitti.com wrote:

Launch your instances with autoscaling groups or another orchestration layer and they can self-register with your Chef server using the config/validator from S3.

On Tue, Apr 28, 2015 at 4:25 PM, niristotle okram nirish.okram@gmail.com wrote:

hello everyone,
I am starting to gather notes to move my servers from a local DC to AWS. Currently in-order to set the http & https proxies, we use a bootstrap template. Now that we are moving to the AWS, i assume we will not be requiring to set this proxies and hence the template can be left aside.

The normal “knife ec2 server create” bootstraps the node/instance. To install a specific version of chef-client, i think i should be able to pass the option : --bootstrap-version VERSION
( https://docs.chef.io/plugin_knife_ec2.html )

So apart from specifying a particular chef-client version to be installed, each time i bootstrap the servers so that they remain consistent across all nodes. is there any other reason that i should have a bootstrap template.

I also came accross ths blog: http://awsadvent.tumblr.com/post/37773106407/bootstrap-cfg-mgmt-aws

why is it important to have the client.rb, validator.pem on the S3/anywhere when the same info can be passed from the knife config file itself.


#5

Recent article on the topic
http://blog.froese.org/2015/04/12/packer-aws-autoscale-chef/

HTH

-Eric Helgeson
@nulleric https://twitter.com/nulleric
https://usingchef.com

On Tue, Apr 28, 2015 at 4:28 PM, E Camden Fisher fish@fishnix.net wrote:

Hi Nirish -

There’s some good documentation for getting started with AWS
autoscaling here:

http://docs.aws.amazon.com/AutoScaling/latest/DeveloperGuide/GettingStartedTutorial.html

We are using packer (http://packer.io) to build an AMI that’s (almost)
ready to bootstrap on it’s own, then we deploy/bootstrap using
terraform (http://terraform.io) and that has been working out pretty
well for us. We do not currently use autoscaling groups, although I
expect we will need them sooner than later.

Cheers,
-Camden

On Tue, Apr 28, 2015 at 5:01 PM, niristotle okram
nirish.okram@gmail.com wrote:

hey Dan, i am just starting on AWS. Could you point me to a doc or
something for “Launch your instances with autoscaling groups or another
orchestration layer”

On Tue, Apr 28, 2015 at 1:41 PM, Daniel Condomitti <
daniel@condomitti.com> wrote:

Launch your instances with autoscaling groups or another orchestration
layer and they can self-register with your Chef server using the
config/validator from S3.

On Tue, Apr 28, 2015 at 4:25 PM, niristotle okram <
nirish.okram@gmail.com> wrote:

hello everyone,
I am starting to gather notes to move my servers from a local DC to
AWS. Currently in-order to set the http & https proxies, we use a bootstrap
template. Now that we are moving to the AWS, i assume we will not be
requiring to set this proxies and hence the template can be left aside.

The normal “knife ec2 server create” bootstraps the node/instance. To
install a specific version of chef-client, i think i should be able to pass
the option : --bootstrap-version VERSION

( https://docs.chef.io/plugin_knife_ec2.html )

So apart from specifying a particular chef-client version to be
installed, each time i bootstrap the servers so that they remain consistent
across all nodes. is there any other reason that i should have a bootstrap
template.

I also came accross ths blog:
http://awsadvent.tumblr.com/post/37773106407/bootstrap-cfg-mgmt-aws

why is it important to have the client.rb, validator.pem on the
S3/anywhere when the same info can be passed from the knife config file
itself.


#6

thanks folks for the pointers, really appreciated… i will read through them

regards!

On Tue, Apr 28, 2015 at 2:46 PM, Eric Helgeson erichelgeson@gmail.com
wrote:

Recent article on the topic
http://blog.froese.org/2015/04/12/packer-aws-autoscale-chef/

HTH

-Eric Helgeson
@nulleric https://twitter.com/nulleric
https://usingchef.com

On Tue, Apr 28, 2015 at 4:28 PM, E Camden Fisher fish@fishnix.net wrote:

Hi Nirish -

There’s some good documentation for getting started with AWS
autoscaling here:

http://docs.aws.amazon.com/AutoScaling/latest/DeveloperGuide/GettingStartedTutorial.html

We are using packer (http://packer.io) to build an AMI that’s (almost)
ready to bootstrap on it’s own, then we deploy/bootstrap using
terraform (http://terraform.io) and that has been working out pretty
well for us. We do not currently use autoscaling groups, although I
expect we will need them sooner than later.

Cheers,
-Camden

On Tue, Apr 28, 2015 at 5:01 PM, niristotle okram
nirish.okram@gmail.com wrote:

hey Dan, i am just starting on AWS. Could you point me to a doc or
something for “Launch your instances with autoscaling groups or another
orchestration layer”

On Tue, Apr 28, 2015 at 1:41 PM, Daniel Condomitti <
daniel@condomitti.com> wrote:

Launch your instances with autoscaling groups or another orchestration
layer and they can self-register with your Chef server using the
config/validator from S3.

On Tue, Apr 28, 2015 at 4:25 PM, niristotle okram <
nirish.okram@gmail.com> wrote:

hello everyone,
I am starting to gather notes to move my servers from a local DC to
AWS. Currently in-order to set the http & https proxies, we use a bootstrap
template. Now that we are moving to the AWS, i assume we will not be
requiring to set this proxies and hence the template can be left aside.

The normal “knife ec2 server create” bootstraps the node/instance.
To install a specific version of chef-client, i think i should be able to
pass the option : --bootstrap-version VERSION

( https://docs.chef.io/plugin_knife_ec2.html )

So apart from specifying a particular chef-client version to be
installed, each time i bootstrap the servers so that they remain consistent
across all nodes. is there any other reason that i should have a bootstrap
template.

I also came accross ths blog:
http://awsadvent.tumblr.com/post/37773106407/bootstrap-cfg-mgmt-aws

why is it important to have the client.rb, validator.pem on the
S3/anywhere when the same info can be passed from the knife config file
itself.


Regards
nirish okram