Atlas On Demand Use

Cloud Execution Environment

1	Introduction
1.1	Description
1.2	Prerequisites
2	Atlas On Demand
2.1	Deployment
2.2	Check Application VM
2.3	Create Backup
2.4	Delete Atlas VM
2.5	Verify Application VM
3	Restore Atlas VM

1 Introduction

Single server configuration of Cloud Execution Environment (CEE) requires low amount of management resources, this makes it possible to disable the Atlas virtual machine (VM) after tenant VMs are deployed.

This document describes how to disable Atlas in a single server CEE installation.

Note:

In this document, memory and storage quantities are represented according to the JESD100B.01 standard:

KB refers to 2¹⁰ bytes
MB refers to 2²⁰ bytes
GB refers to 2³⁰ bytes

1.1 Description

Single server is a special, low cost platform of CEE. It is installed on a single compute host server that runs:

Single vCIC with low resource consumption, including vCPU cores, memory huge pages, local storage

Tenant VMs

Atlas for the time of tenant VM provisioning

This CEE platform does not include any external storage connectivity.

Set up Atlas On Demand by deploying an Atlas VM and VNF. Create a backup of the Atlas VM to save the Atlas VM configuration. If Atlas is no longer required, delete the Atlas VM or image.

Note:: If Heat is used to deploy an application, the state is saved in the Atlas database. If OpenStack command-line client is used to manage (stop, delete) resources, the database in Atlas becomes unsynchronized and Atlas cannot manage these resources.

1.2 Prerequisites

1.2.1 Documents

Before starting the procedure, ensure that the following documents are available:

1.2.2 Tools and Equipment

The following tools and equipment are needed:

Access to the CEE REST APIs
Secure Shell (SSH) logon to CEE

1.2.3 Conditions

Before starting this procedure, ensure that the following conditions are met:

Credentials for logging in to the host are available.
No other maintenance activities are taking place at the same time.

2 Atlas On Demand

This section describes how Atlas On Demand can be achieved.

2.1 Deployment

Deploy an Atlas VM by following the instructions in Atlas SW Installation.

Use an Atlas image file with the following ID syntax:
ecs-atlas-x86_64-<target_atlas_version>-<build_number>.qcow2
and the <name_of_atlas> variable, Atlas_for_backup.

For a minimum configuration, see: Section 2.1.1.

Deploy an application using Atlas.

2.1.1 Atlas Minimum Configuration

Atlas can be launched with a minimum configuration flavor of
Disk: 10 GB
RAM: 1024 MB
VCPU: 2
Ephemeral disk/Attached volume: 10 GB

Note:: Although the minimum size of ephemeral disk is 10 GB, it must be proportional to the size of uploaded images and packages.
Images and packages are stored in /tmp, which is a fixed 30% of the overall disk. So the ephemeral disk must have the size about 4 times bigger that the images and packages altogether.

Update the following parameters in localrc:
# Flavor specifications
Disk=10
RAM=1024
VCPU=2

# Size of data volume
readonly DATA_IMAGE_SIZE=10

DATA_IMAGE_SIZE is defined per use (size of application, or image to upload). Sizes are divided in percentage of total DATA_IMAGE_SIZE attached:

/dev/data-vg/internal-data-lv 15% /dev/data-vg/log-collector 10% /dev/data-vg/var/core 10% /dev/data-vg/var/archives 10% /dev/data-vg/var/lib/mysql 10% /dev/data-vg/var/log 10% /dev/data-vg/tmp 30% /dev/data-vg/var/tmp 5%

2.2 Check Application VM

Check the application from the deployed stack by using nova list, ovft capp-list, and heat stack-list.

atlasadm@atlas:~$ ovft capp-list

id	name	status	type	created_at
`8251da77-332f-41b5-8f60-887f99513419`	`package`	`active`	ovf	2016-02-12T06:26:03.000000

atlasadm@atlas:~$ heat stack-list

id	stack_name	stack_status	creation_time	updated_time
`a5f047d8-75ba-4fc9-b93d-f73220e0a85a`	`stack`	`CREATE_COMPLETE`	2018-03-20T07:03:29Z	2018-03-20T12:05:22Z

atlasadm@atlas:~$ nova list

ID	Name	Status	Task State	Power State	Networks
`6845f0f2-6b4b-42ed-bff5-1883a006347c`	`Atlas`	`ACTIVE`	None	Running	`tenant_3583=10.33.190.38`
`38a57214-c639-4e21-9042-91542052a6e1`	`Controller node1`	`ACTIVE`	None	Running	`BP_1=127.3.0.5`
`ebbf9bc7-eda7-4c9b-8d48-b87c798a431a`	`Payload node1`	`ACTIVE`	None	Running	`BP_2=127.13.0.6`

2.3 Create Backup

Create a backup of the Atlas configuration and store in local storage. See Atlas Backup for more information. To upload the backup to local storage do the following:

Switch user from atlasadm to root with the below command:
atlasadm@atlas:~ $sudo -i
Ensure that the current directory is set to /var/archives:
root@atlas:~ # cd /var/archives
Verify that the created backup directory and files are available:
root@atlas:/var/archives# ls <atlas_backup_directory>
An example of the command is:

root@atlas:/var/archives# ls AtlasBackup1472730328

An example of the printout is:
```
AtlasBackup.1472730328-all-mysql-databases.sql.bz2.enc
AtlasBackup.1472730328-root.master.tar.gz.enc
AtlasBackup.1472730328-etc-puppet-hieradata-passwords.yaml.master.tar.gz.enc
AtlasBackup.1472730328.sha256.enc
AtlasBackup.1472730328-home-atlasadm.master.tar.gz.enc
```
Note:
The backup consists of the above listed files.
Upload the backup for a specified ID:
root@atlas:~# scp -r *<id>* <user@laptop>:<storage_path>

Note:
This step is performed if Swift is not present. If Swift is present, the created backup is automatically uploaded to Swift storage.

2.4 Delete Atlas VM

Deleting the Atlas VM saves CPU, memory and disk size, deleting the Atlas image saves storage space.

Delete the Atlas VM using the following command:

nova delete Atlas_for_backup

Delete the Atlas image using the following command:

root@cic-1:~# rm /root/artifacts/ecs-atlas-x86_64-<target_atlas_version>-<build_number>.qcow2

2.5 Verify Application VM

Verify that the application is running and the resources previously allocated to the Atlas VM are free:

root@cic-1:~$ nova list

ID	Name	Status	Task State	Power State	Networks
`38a57214-c639-4e21-9042-91542052a6e1`	`Controller node1`	`ACTIVE`	None	Running	`BP_1=127.3.0.5`
`ebbf9bc7-eda7-4c9b-8d48-b87c798a431a`	`Payload node1`	`ACTIVE`	None	Running	`BP_2=127.13.0.6`

To verify free space after deleting the Atlas VM, check the nova-compute.log for available resources.

3 Restore Atlas VM

The Atlas VM configuration can be restored using the previously created backup.

There is no need to restore the Atlas VM if Atlas was shut down and not deleted. The image and atlas_install.sh transferred to the vCIC in the original installation can be used if it is still available in the vCIC.

Deploy an Atlas VM in the same way as in Section 2.1.
Restore the configuration of the Atlas VM from the backup stored in local storage by following the instructions in Atlas Restore.
Check the application as described in Section 2.2 and compare the results.