Large Custom Images – Uploading Files > 5GB to Fujitsu K5 Object Storage (Swift)

In a previous blog post I detailed a simple process for uploading custom images to the Fujitsu K5 IaaS platform. One challenge that I overlooked is large custom image sizes which I’ll address here.

Background: OpenStack Swift objects have a maximum size limitation of 5GB. However the Swift containers can have many thousands of these objects. Files that are larger than 5GB need to be split into a subset of smaller files before being uploaded to the container. When all the component files have been uploaded a zero byte file with a ‘manifest’ header is uploaded to the container. This header is composed of the container name and prefix used to build the component file names. When this zero byte file is referenced through the Swift API all the component files are concatenated and the original large file is downloaded.

The following python script will upload custom images smaller than 1 GB directly into K5’s object storage. Larger files are split into 1GB chunks before uploading to the object storage. The default 1GB size can be changed by using the command line parameters.

Once the image is uploaded it is then registered with the the default K5 project. The image must now be shared with other member projects that wish to consume this image – this will be shared in the next post.

Prerequisites : The script relies on a settings file, k5contractsettings.py, which must contain all your contract details and be placed in the same directory  – for example:


#!/usr/bin/python
adminUser = 'username'
adminPassword = 'password'
contract = 'contract_name'
contractid = 'contract_id'
defaultid = 'default_project_id'
project = 'working_project'
region = 'uk-1'

Image Upload Example Script


#!/usr/bin/python
# Author : Graham Land
# Date: 08/10/2016
#
# Purpose: Upload a custom image to K5 Object Storage and then register with K5 Glance
# If the image is greater than 1GB it will be broken down into 1GB chunks
# and then uploaded
# Command line parameters –
# -i image_path
# -c container_name
# -s chunk_size (bytes)
# -n display_name
# -t image_type
# -p project
#
# Prerequisites: k5contractsettings.py file in the same directory with login details
#
# adminUser = 'username'
# adminPassword = 'password'
# contract = 'contract_name'
# contractid = 'contract_id'
# defaultid = 'default_project_id'
# project = 'working_project'
# region = 'uk-1'
#
# blog: https://allthingscloud.eu
# twitter: @allthingsclowd
import sys
import os
import requests
import uuid
import base64
import time
import getopt
import ntpath
# load your K5 contract details from k5contractsettings.py file
from k5contractsettings import *
# get a scoped auth token
def get_scoped_token(uname,upassword,uproject,udomain,uregion):
identityURL = 'https://identity.' + uregion + '.cloud.global.fujitsu.com/v3/auth/tokens'
response = requests.post(identityURL,
headers={'Content-Type': 'application/json','Accept':'application/json'},
json={"auth":
{"identity":
{"methods":["password"],"password":
{"user":
{"domain":
{"name":udomain},
"name":uname,
"password": upassword
}}},
"scope":
{ "project":
{"id":uproject
}}}})
return response.headers['X-Subject-Token']
def get_unscoped_token(uname,upassword,udomain,uregion):
identityURL = 'https://identity.' + uregion + '.cloud.global.fujitsu.com/v3/auth/tokens'
response = requests.post(identityURL,
headers={'Content-Type': 'application/json','Accept':'application/json'},
json={"auth":
{"identity":
{"methods":["password"],"password":
{"user":
{"domain":
{"name":udomain},
"name":uname,
"password": upassword
}}}}})
return response.headers['X-Subject-Token']
# get a central identity portal token
def get_unscoped_idtoken(uname,upassword,udomain):
response = requests.post('https://auth-api.jp-east-1.paas.cloud.global.fujitsu.com/API/paas/auth/token',
headers={'Content-Type': 'application/json'},
json={"auth":
{"identity":
{"password":
{"user":
{"contract_number":udomain,
"name":uname,
"password": upassword
}}}}})
return response.headers['X-Access-Token']
# create a container
def create_new_storage_container(adminUser,adminPassword,project,container_name,contract,region):
# get a regional domain scoped token to make queries to facilitate conversion of object names to ids
scoped_k5token = get_scoped_token(adminUser,adminPassword,project,contract,region)
print scoped_k5token
identityURL = 'https://objectstorage.' + region + '.cloud.global.fujitsu.com/v1/AUTH_' + project + '/' + container_name
print identityURL
response = requests.put(identityURL,
headers={'X-Auth-Token':scoped_k5token,'Content-Type': 'application/json'})
return response
def upload_file_to_container(adminUser,adminPassword,project,container_name,file_name,file_path,contract,region):
# get a regional domain scoped token to make queries to facilitate conversion of object names to ids
scoped_k5token = get_scoped_token(adminUser,adminPassword,project,contract,region)
uploadfile = open(file_path, 'rb')
data = uploadfile.read()
identityURL = 'https://objectstorage.' + region + '.cloud.global.fujitsu.com/v1/AUTH_' + project + '/' + container_name + '/' + file_name
response = requests.put(identityURL,
data=data,
headers={'X-Auth-Token':scoped_k5token,'Content-Type': 'application/octet-stream'})
uploadfile.close
return response
def import_from_container_to_k5(adminUser,adminPassword,project,container_name,file_name,display_name,file_path,os_type,contract,region):
# get a regional domain scoped token to make queries to facilitate conversion of object names to ids
scoped_k5token = get_scoped_token(adminUser,adminPassword,project,contract,region)
k5ContainerURL = '/v1/AUTH_' + project + '/' + container_name + '/' + file_name
image_id = str(uuid.uuid4())
encodedPassword = base64.b64encode(adminPassword)
vmimportURL = 'https://vmimport.' + region + '.cloud.global.fujitsu.com/v1/imageimport'
response = requests.post(vmimportURL,
headers={'X-Auth-Token':scoped_k5token},
json={"name":display_name,
"location":k5ContainerURL,
"id":image_id,
"conversion": True,
"os_type":os_type,
"user_name":adminUser,
"password":encodedPassword,
"domain_name":contract})
return response.json()
def verify_image_import_status(adminUser,adminPassword,project,image_id,contract,region):
# get a regional domain scoped token to make queries to facilitate conversion of object names to ids
scoped_k5token = get_scoped_token(adminUser,adminPassword,project,contract,region)
vmimportURL = 'https://vmimport.' + region + '.cloud.global.fujitsu.com/v1/imageimport/' + image_id + '/status'
response = requests.get(vmimportURL,
headers={'X-Auth-Token':scoped_k5token})
return response.json()
def upload_manifest_to_container(adminUser,adminPassword,project,container_name,file_name,prefix,contract,region):
# get a regional domain scoped token to make queries to facilitate conversion of object names to ids
scoped_k5token = get_scoped_token(adminUser,adminPassword,project,contract,region)
identityURL = 'https://objectstorage.' + region + '.cloud.global.fujitsu.com/v1/AUTH_' + project + '/' + container_name + '/' + file_name
response = requests.put(identityURL,
headers={'X-Auth-Token':scoped_k5token,'X-Object-Manifest': container_name + '/' + prefix})
return response
# list items in a container
def view_items_in_storage_container(adminUser,adminPassword,project,container_name,contract,region):
# get a regional domain scoped token to make queries to facilitate conversion of object names to ids
scoped_k5token = get_scoped_token(adminUser,adminPassword,project,contract,region)
identityURL = 'https://objectstorage.' + region + '.cloud.global.fujitsu.com/v1/AUTH_' + project + '/' + container_name + '?format=json'
response = requests.get(identityURL,
headers={'X-Auth-Token':scoped_k5token,'Content-Type': 'application/json'})
return response
# download item in a container
def download_item_in_storage_container(adminUser,adminPassword,project,container_name,contract,region):
# get a regional domain scoped token to make queries to facilitate conversion of object names to ids
scoped_k5token = get_scoped_token(adminUser,adminPassword,project,contract,region)
identityURL = 'https://objectstorage.' + region + '.cloud.global.fujitsu.com/v1/AUTH_' + project + '/' + container_name + '/manifest'
print identityURL
response = requests.get(identityURL,
headers={'X-Auth-Token':scoped_k5token,'Content-Type': 'application/json'})
return response
def make_out_filename(prefix, idx):
'''Make a filename with a serial number suffix.'''
return prefix + str(idx).zfill(4)
def bsplit(in_filename, bytes_per_file,os_type):
'''Split the input file in_filename into output files of
bytes_per_file bytes each. Last file may have less bytes.'''
in_fil = open(in_filename, "rb")
outfil_idx = 1
out_filename = make_out_filename(os_type, outfil_idx)
out_fil = open(out_filename, "wb")
byte_count = tot_byte_count = file_count = 0
c = in_fil.read(1)
# Loop over the input and split it into multiple files
# of bytes_per_file bytes each (except possibly for the
# last file, which may have less bytes.
while c != '':
byte_count += 1
out_fil.write(c)
# Bump vars; change to next output file.
if byte_count >= bytes_per_file:
tot_byte_count += byte_count
byte_count = 0
file_count += 1
out_fil.close()
result = upload_file_to_container(adminUser,adminPassword,defaultid,container_name,out_filename,out_filename,contract,region)
print "Uploaded Package – " + str(file_count)
os.remove(out_filename)
outfil_idx += 1
out_filename = make_out_filename(os_type, outfil_idx)
out_fil = open(out_filename, "wb")
c = in_fil.read(1)
# Clean up.
in_fil.close()
if not out_fil.closed:
out_fil.close()
result = upload_file_to_container(adminUser,adminPassword,defaultid,container_name,out_filename,out_filename,contract,region)
print "\nUploaded Package – " + str(file_count)
os.remove(out_filename)
if byte_count == 0:
os.remove(out_filename)
# now create manifest file
result = upload_manifest_to_container(adminUser,adminPassword,defaultid,container_name,file_name,os_type,contract,region)
return result
def main():
try:
# ensure minimium commandline paramaters have been supplied
if (len(sys.argv)<6):
print("Usage1: %s -i 'path_to_image' -c 'container_name' -n 'image_display_name' -p '{project1,project2,project3}' -t [ubuntu|centos|rehhat|win2008SE] [-s 'chunk size in bytes'] " % sys.argv[0])
sys.exit(2)
# load the command line parameters
myopts, args = getopt.getopt(sys.argv[1:],"i:c:n:p:t:s:",["imagepath=","container=","name=","projects=","type=","size="])
except getopt.GetoptError:
# if the parameters are incorrect display error message
print("Usage2: %s -i 'path_to_image' -c 'container_name' -n 'image_display_name' -p '{project1,project2,project3}' -t [ubuntu|centos|rehhat|win2008SE] [-s 'chunk size in bytes'] " % sys.argv[0])
sys.exit(2)
# define global variables from the command line parameters
global container_name
global display_name
global bytes_per_file
global os_type
global file_path
global file_name
# set default chunk size for large images that needs to be broken up must be below 5GB for Swift Object Storage
bytes_per_file = 1048576000 #5242880 #1048576000 #262144000 # 250Mb chunks
###############################
# o == option
# a == argument passed to the o
###############################
for o, a in myopts:
if o in ('-i','–imagepath'):
file_path=a
elif o in ('-c','–container'):
container_name=a
elif o in ('-n','–name'):
display_name=a
elif o in ('-p','–projects'):
projects=a
elif o in ('-t','–type'):
os_type=a
elif o in ('-s','–size'):
bytes_per_file=int(a)
else:
print("Usage3: %s -i 'path_to_image' -c 'container_name' -n 'image_display_name' -p '{project1,project2,project3}' -t [ubuntu|centos|rehhat|win2008SE] [-s 'chunk size in bytes'] " % sys.argv[0])
# extract filename from file path suplied at cli
file_name = ntpath.basename(file_path)
# attempt to read the contents of the container to see if it already exists
result = view_items_in_storage_container(adminUser,adminPassword,defaultid,container_name,contract,region)
# check to see if container already exists, if not then create it
if (result.status_code == 404):
# create container
print "\nCreating new container : " + container_name
result = create_new_storage_container(adminUser,adminPassword,defaultid,container_name,contract,region)
print "\nCreated new container : " + container_name
# check size of file to be uploaded is less than 250GB, if not split into smaller chunks for upload
if (os.path.getsize(file_path) > bytes_per_file):
# loop through image file for multi-part upload
print "\n———- Starting multi-part file upload to K5 object storage —— \n"
result = bsplit(file_path, bytes_per_file,os_type)
print "\n———- Finished multi-part file upload to K5 object storage —— \n"
else:
# simple file upload to container
print "\n———- Starting simple file upload to K5 object storage —— \n"
result = upload_file_to_container(adminUser,adminPassword,defaultid,container_name,file_name,file_path,contract,region)
print "\n———- Finished simple file upload to K5 object storage —— \n"
# list container
print "\n———- List container contents K5 object storage start —— \n"
result = view_items_in_storage_container(adminUser,adminPassword,defaultid,container_name,contract,region)
print result
print "\n———- List container contents K5 object storage end —— \n"
# Register image with K5
print "\n———- Registering image with K5 —— \n"
result = import_from_container_to_k5(adminUser,adminPassword,defaultid,container_name,file_name,display_name,file_path,os_type,contract,region)
image_id = result['import_id']
print result
print "\n———- K5 Image import_id : " + image_id + "\n"
# Get import status
print "\n———- Check import status ———- \n\n"
result = verify_image_import_status(adminUser,adminPassword,defaultid,image_id,contract,region)
print result
while ((result['import_status'] != "succeeded") and (result['import_status'] != "failed")):
time.sleep(300)
print "\n———- Check import status ———- \n"
result = verify_image_import_status(adminUser,adminPassword,defaultid,image_id,contract,region)
print result
print "End of Import Process – Import status >>> " + result['import_status']
if __name__ == "__main__":
main()

Example script output:

C:\Users\landg\>python K5ImageUpload.py -i "c:\Users\landg\Downloads\cirrosuploadtest.vmdk" -c uploaddemo12 -n "h
ello k5 milti image"  -t ubuntu -p NotUsed -s 5242880

Creating new container : uploaddemo12
31c0f27e562c4b3089a546c175c144e4
https://objectstorage.uk-1.cloud.global.fujitsu.com/v1/AUTH_eadb882573ac40b1b101
eac93009a313/uploaddemo12

Created new container : uploaddemo12

---------- Starting multi-part file upload  to K5 object storage ------

Uploaded Package - 1
Uploaded Package - 2
Uploaded Package - 3
Uploaded Package - 4

Uploaded Package - 4

---------- Finished multi-part file upload to K5 object storage ------


---------- List container contents K5 object storage start ------

<Response [200]>

---------- List container contents K5 object storage end ------


---------- Registering image with K5 ------

{u'import_id': u'6a0b58c5-bcda-4a64-919e-23f06b8338ad'}

---------- K5 Image import_id : 6a0b58c5-bcda-4a64-919e-23f06b8338ad


---------- Check import status ----------


{u'conversion': True, u'name': u'hello k5 milti image', u'container_format': u'b
are', u'min_ram': u'0', u'ovf_location': u'', u'disk_format': u'raw', u'domain_n
ame': u'YssmW1yI', u'location': u'/v1/AUTH_eadb882573ac40b1b101eac93009a313/uplo
addemo12/cirrosuploadtest.vmdk', u'min_disk': u'0', u'progress': u'0', u'os_type
': u'ubuntu', u'password': u'*', u'user_name': u'landg', u'id': u'70a38639-f819-
4375-b3d2-cfc99c2a148e', u'import_status': u'queued'}

---------- Check import status ----------

{u'conversion': True, u'name': u'hello k5 milti image', u'container_format': u'b
are', u'min_ram': u'0', u'ovf_location': u'', u'disk_format': u'raw', u'domain_n
ame': u'YssmW1yI', u'location': u'/v1/AUTH_eadb882573ac40b1b101eac93009a313/uplo
addemo12/cirrosuploadtest.vmdk', u'min_disk': u'0', u'progress': 0, u'os_type':
u'ubuntu', u'password': u'*', u'user_name': u'landg', u'id': u'70a38639-f819-437
5-b3d2-cfc99c2a148e', u'import_status': u'processing'}

---------- Check import status ----------

{u'container_format': u'bare', u'min_ram': 0, u'updated_at': u'2016-10-08T17:24:
38Z', u'file': u'/v2/images/70a38639-f819-4375-b3d2-cfc99c2a148e/file', u'owner'
: u'eadb882573ac40b1b101eac93009a313', u'id': u'70a38639-f819-4375-b3d2-cfc99c2a
148e', u'size': 41126400, u'conversion': True, u'self': u'/v2/images/70a38639-f8
19-4375-b3d2-cfc99c2a148e', u'disk_format': u'raw', u'domain_name': u'YssmW1yI',
 u'location': u'/v1/AUTH_eadb882573ac40b1b101eac93009a313/uploaddemo12/cirrosupl
oadtest.vmdk', u'progress': 100, u'user_name': u'landg', u'schema': u'/v2/schema
s/image', u'status': u'active', u'import_status': u'succeeded', u'tags': [], u'v
isibility': u'private', u'BaseImageId': u'415b3a0a513aebc27d34c68bd8cdae8c', u'm
in_disk': 0, u'password': u'*', u'name': u'hello k5 milti image', u'created_at':
 u'2016-10-08T17:24:32Z', u'ovf_location': u'', u'fcx.centos': u'true', u'protec
ted': False, u'os_type': u'ubuntu'}
End of Import Process - Import status >>>  succeeded

Happy Stacking!

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s