-
-
Save peplin/470321 to your computer and use it in GitHub Desktop.
# Source accepts the protocol s3:// with the host as the bucket | |
# access_key_id and secret_access_key are just that | |
s3_file "/var/bulk/the_file.tar.gz" do | |
source "s3://your.bucket/the_file.tar.gz" | |
access_key_id your_key | |
secret_access_key your_secret | |
owner "root" | |
group "root" | |
mode 0644 | |
end |
# | |
# Author:: Christopher Peplin (<[email protected]>) | |
# Copyright:: Copyright (c) 2010 Bueda, Inc. | |
# License:: Apache License, Version 2.0 | |
# | |
# Licensed under the Apache License, Version 2.0 (the "License"); | |
# you may not use this file except in compliance with the License. | |
# You may obtain a copy of the License at | |
# | |
# http://www.apache.org/licenses/LICENSE-2.0 | |
# | |
# Unless required by applicable law or agreed to in writing, software | |
# distributed under the License is distributed on an "AS IS" BASIS, | |
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. | |
# See the License for the specific language governing permissions and | |
# limitations under the License. | |
# | |
class Chef | |
class Provider | |
class S3File < Chef::Provider::RemoteFile | |
def action_create | |
Chef::Log.debug("Checking #{@new_resource} for changes") | |
if current_resource_matches_target_checksum? | |
Chef::Log.debug("File #{@new_resource} checksum matches target checksum (#{@new_resource.checksum}), not updating") | |
else | |
Chef::Log.debug("File #{@current_resource} checksum didn't match target checksum (#{@new_resource.checksum}), updating") | |
fetch_from_s3(@new_resource.source) do |raw_file| | |
if matches_current_checksum?(raw_file) | |
Chef::Log.debug "#{@new_resource}: Target and Source checksums are the same, taking no action" | |
else | |
backup_new_resource | |
Chef::Log.debug "copying remote file from origin #{raw_file.path} to destination #{@new_resource.path}" | |
FileUtils.cp raw_file.path, @new_resource.path | |
@new_resource.updated = true | |
end | |
end | |
end | |
enforce_ownership_and_permissions | |
@new_resource.updated | |
end | |
def fetch_from_s3(source) | |
begin | |
protocol, bucket, name = URI.split(source).compact | |
name = name[1..-1] | |
AWS::S3::Base.establish_connection!( | |
:access_key_id => @new_resource.access_key_id, | |
:secret_access_key => @new_resource.secret_access_key | |
) | |
obj = AWS::S3::S3Object.find name, bucket | |
Chef::Log.debug("Downloading #{name} from S3 bucket #{bucket}") | |
file = Tempfile.new("chef-s3-file") | |
file.write obj.value | |
Chef::Log.debug("File #{name} is #{file.size} bytes on disk") | |
begin | |
yield file | |
ensure | |
file.close | |
end | |
rescue URI::InvalidURIError | |
Chef::Log.warn("Expected an S3 URL but found #{source}") | |
nil | |
end | |
end | |
end | |
end | |
end | |
class Chef | |
class Resource | |
class S3File < Chef::Resource::RemoteFile | |
def initialize(name, run_context=nil) | |
super | |
@resource_name = :s3_file | |
end | |
def provider | |
Chef::Provider::S3File | |
end | |
def access_key_id(args=nil) | |
set_or_return( | |
:access_key_id, | |
args, | |
:kind_of => String | |
) | |
end | |
def secret_access_key(args=nil) | |
set_or_return( | |
:secret_access_key, | |
args, | |
:kind_of => String | |
) | |
end | |
end | |
end | |
end |
Thanks for the report - I'm happy to see someone else finds this useful. I'm not able to reproduce the bug you describe - is this an example of what returns a URI::InvalidURIError
for you?
>> URI.parse("s3://bueda/test.zip").path[1..-1]
=> "test.zip"
Hey there,
Try the following:
>> URI.parse("s3://bueda_foobar/test.zip")
URI::InvalidURIError: the scheme s3 does not accept registry part: bueda_foobar (or bad hostname?)
from /System/Library/Frameworks/Ruby.framework/Versions/1.8/usr/lib/ruby/1.8/uri/generic.rb:195:in `initialize'
from /System/Library/Frameworks/Ruby.framework/Versions/1.8/usr/lib/ruby/1.8/uri/common.rb:492:in `new'
from /System/Library/Frameworks/Ruby.framework/Versions/1.8/usr/lib/ruby/1.8/uri/common.rb:492:in `parse'
from (irb):4
Seems it's underscores that are causing issues...
Also, wanted to say that it was indeed very helpful! Thanks so much for coding this up, Christopher!
Thanks - I guess using URI is a bit of a stretch since bucket names don't have to be valid FQDNs. I've updated this to use the URI.split()
method as you do.
How do you actually use this? Sorry, I'm new to Chef and trying to understand this as I build up my stack. Would these files go into a cookbook of their own? Could you give a quick example of how you're using them?
Sure, put s3_file.rb
in the libraries/
folder of any cookbook (create it if it doesn't exist) and it should be automatically imported.
Alternatively, make a standalone s3
cookbook with the file in s3/libraries/
and in other cookbooks, just call include_recipe "s3"
before using it.
hi I added support for all regions and removed the need for the aws-s3 gem
Hi,
I've added to the excellent work of peplin and casualjim (thanks), to add checking for XML error responses from Amazon and to add dual support for S3 and other protocols provided by the out-of-the-box RemoteFile (e.g. http(s)).
https://gist.github.com/1351043
Thanks for the work, folks!
I'm getting an NameError: uninitialized constant Chef::Provider::S3File::AWS.
What I'm doing wrong?
@TiagoBrito @poobury @casualjim @peplin I've created an update to @peplin's original gist using AWS-SDK at https://gist.github.com/DavidAllison/5288249
It includes minor changes that make it work with the newer AWS-SDK gem (the current version as of this writing). Peplin's older gem won't work with the newer AWS-SDK gem since the API is no longer the same.
@poobury and @casualjim, I really like where you guys went with this, but unfortunately I wasn't able to easily make either or your recipes work with AWS's temporary S3 credential granting based on IAM policies (whereas the AWS-SDK gem from Amazon handles this quite nicely). So I ended up modernizing @peplin's original work.
These all have the same problem - to be generally useful, the temporary file needs to be written in binary mode! e.g. for the above:
file = Tempfile.new("chef-s3-file")
file.binmode
file.write obj.value
Getting a "no resource or method named s3_file
for Chef::recipe "default" when we try to use this. ANy ideas?
Have you tried to add include_recipe "s3_file" into your cookbook recipe?
i get "no resource or method named s3_file for Chef::recipe "default" no matter what i try
Hi,
I want how to write (or) the S3 file.
Thanks
Karthi
Can this resource copy a directory of files to a local directory as well as just a single file?
Hello. I am getting the following... any help appreciated:
19: class Chef
20: class Provider
21>> class S3File < Chef::Provider::RemoteFile
22: def action_create
23: Chef::Log.debug("Checking #{@new_resource} for changes")
24:
25: if current_resource_matches_target_checksum?
26: Chef::Log.debug("File #{@new_resource} checksum matches target checksum (#{@new_resource.checksum}), not updating")
27: else
28: Chef::Log.debug("File #{@current_resource} checksum didn't match target checksum (#{@new_resource.checksum}), updating")
29: fetch_from_s3(@new_resource.source) do |raw_file|
30: if matches_current_checksum?(raw_file)
[2016-08-25T09:50:43-04:00] ERROR: Running exception handlers
[2016-08-25T09:50:43-04:00] ERROR: Exception handlers complete
[2016-08-25T09:50:43-04:00] FATAL: Stacktrace dumped to /var/chef/cache/chef-stacktrace.out
Chef Client failed. 0 resources updated
[2016-08-25T09:50:43-04:00] ERROR: uninitialized constant #Class:0x00000003144990::Chef::Provider::RemoteFile
[2016-08-25T09:50:43-04:00] FATAL: Chef::Exceptions::ChildConvergeError: Chef run process exited unsuccessfully (exit code 1)
Hi,
this doesn't work for me when I don't pass access keys, actually the instance already has an IAM role assigned and has enough permission on bucket.
I've ran into an issue with the library, mostly with URI.parse - if the bucket name does not contain a dot in the FQDN portion of the URI, then URI.parse throws a URI::InvalidURIError exception, even if having an FQDN without a dot in S3 is completely valid.
I've used the following bits of code for uri, name, and bucket:
uri = URI.split(source).compact
name = uri[2][1..-1]
bucket = uri[1]