Infrastructure

The topic covers the questions on Cloud infrastructure creation and manipulation.

This topic contains the following sections:

What is a virtualization region?

EPAM Cloud is based on a number of hardware servers. These servers are geographically located in different places. A group of such servers located in one place is called a virtualization region. The infrastructure cost in the regions may vary as the servers maintenance cost is different in different countries.

How can I find out where the datacenter equipment is located physically?

Each virtualiztion region has the "locatoin" parameter indicating the physical location of the hardware serving the virtualization region. You can find the location by hovering your mouse over the region name in the Web UI. The location parameter is displayed in the "Run" wizard, the "Manage Services" wizard, the "Manage Cloud" wizard and the "Schedules" wizard.

In Maestro CLI, you can find the datacenter location in the output of or2dpro and or2dreg commands.

What is a project in terms of EPAM Cloud?

A project in EPAM Cloud is a temporarty logical unit to which the Cloud resources and bills are mapped. An EPAM Cloud project name matches the name of the related UPSA project. The project is activated in Cloud by a Support Request at EPAM Service Portal. The request should be approved by Project Manager/Coordinator.

What is the capacity of CPUs on my VMs? Is it manageable?

EPAM Cloud, as other Cloud services, provides VMs with virtual CPUs (vCPU) that differ from hardware CPUs in their performance. It is impossible to set up the CPU capacity manually, as it depends on the load of the server where the VM is located.

How do I specify RAM and CPU capacities for my VMs?

EPAM Cloud uses a standard approach in providing virtual servers: the VM is provided as a single unit that contains a predefined set of components. These sets are called shapes and include various combinations of CPU and RAM so that you can choose the one which is the most suitable for you. The shape size influences the cost of the VM.

You can find the list of the existing shapes and their resources on Cloud Computing Service (C2S) page (Existing Shape Size Values section). The table also provides the AWS shape mapping in case you want to compare EPAM Cloud and AWS shapes details.

Are there any region-related restrictions?

Yes, there is a number of region-related restrictions in EPAM Cloud, this especially concerns OpenStack-type regions and EPAM-KZ1.

You can find the details on the existing region restrictions in our Maestro CLI User Guide, Annex D - Service Locations.

What is the largest instance shape in EPAM Cloud?

The heaviest type is 3XL (8CPU, 15GB RAM). The existing instance shapes are listed on the Cloud Computing Service (C2S) page. Please note that XL, 2XL, and 3XL shapes can be activated only in EPAM-BY1 and EPAM-HU1 regions.

Are there any limitations on shape/machine image combinations?

Generally, there are no limitations on shape/machine combinations in EPAM Cloud. The only exception is that Windows systems cannot be run on shapes less than MEDIUM.

Is it possible to scale an instance (change its shape)?

Moving to another shape is usually done by creating an image of your instance (the instance will be terminated automatically afterwards, unless it is run in AWS or CSA-type regions) and running a new instance based on a suitable shape, using the newly created custom machine image.

In exceptional cases, you can leave a Support Request at EPAM Service Portal to move to another shape.

Can I resize the instance volume if there is a checkpoint on this instance?

No, this feature is unavailable in EPAM Cloud.

Is there a way to decrease storage volume size?

No, current implementation only allows increasing storage volume size. Decreasing the size can lead to data loss.

How can I check the used disk space in my instance volume?

Maestro CLI allows you to view the used space according to VM environment. To see the disk space actually available and used by your VM you should use the standard methods of disk space monitoring provided by your operating system. However, if you have many instances you should use Zabbix auto-configuration role for disk space control.

More information can be found in Services Guide.

How do I reattach storage volumes between my instances?

You can do it in two steps:
1. Run the 'or2detvol' CLI command to detach the volume from its instance
2. Run the 'or2attvol' CLI command to attach this volume to another instance.
(see Maestro CLI User Guide).

In case it is impossible to reattach volumes between the two instances by means of self service (see command response), please leave a respective Support Request at EPAM Service Portal.

I want to create a checkpoint of my instance. Do I have to stop the instance first?

Yes, you should. Checkpoints can be created on stopped instances only. However, it is recommended to create checkpoints after ACS processes are finished. This can cause ACS failure when the instance is reverted to such a checkpoint.

Please note that reverting to a checkpoint will take less time if the checkpoint was created on a stopped instance.

Can I use checkpoints as permanent system backups? Are there any limitations?

It's not advised to use the checkpoints for system backups. They have a negative effect on instance performance. We strongly recommend to create them only before introducing critical changes to instance configuration and to remove them afterwards.

When deleting a checkpoint, please keep in mind that the process can take significant time, even several hours. During the checkpoint deletion, the instance is unavailable.

Also, please do not try using checkpoints as machine images. They are not designed for this purpose and cannot perform this function properly.

I have a custom image in a certain region, and want to run a VM from this image in another one. How can I do that?

To run such an instance, please submit a service request to HelpDesk, and specify the image, its location and the region in which a VM should be run.

I can see instances in 'missing' state. What does this mean and how can I delete them?

The 'missing' state means that something went wrong during the instance performance and we cannot fix this. You can terminate such instances using the 'or2-terminate-instances' command.

Which time zone is used by orchestrator scheduler?

Scheduler is based on UTC time. So make sure that you use a right time for your zone (for example, Kiev - GMT+3 (Daylight)).

How do I add instances to a VLAN via CLI?

Adding instances to a VLAN needs several steps to be performed. If you have already performed at least one of them, you can skip it and proceed to the next step:

- Create a VLAN for your project. Please, submit a request via EPAM Service Portal to request a VLAN activation for your project.
- Activate the VLAN in EPAM Cloud Orchestrator. Please, submit an activation request via EPAM Service Portal. Specify the name of you project and the ID of the VLAN you would like to be activated (The ID is assigned permanently on VLAN creation). - Once your VLAN has been created and activated, move your instances using the 'or2-move-instance-to-vlan' command (see Maestro CLI User Guide).

How do I move an instance to another VLAN in terms of self-service?

To move an instance to an existing VLAN, please, use the or2-move-instance-to-vlan (or2mivlan) command and specify the instance ID, the project and region the instance is run in, and the target VLAN ID.

The list of the VLANs available in a region for the specified project can be retrieved with the or2-describe-vlans command.

For the details on the mentioned commands please see the Maestro CLI User Guide.

How can I monitor my EPAM Cloud infrastructure performance?

EPAM Cloud Orchestration allows you to monitor the Cloud infrastructure performance and cost alteration at all the structural levels (Cloud, regions, projects, VMs).

The analytics data is represented as a set of graphs and their descriptions on Orchestrator Monitoring and Radar pages.

For more details on monitoring infrastructure in Cloud, please, see the Cloud Analytics page.

How long does it take Orchestrator to perform operations on infrastructure elements?

The time needed to perform an operation on an infrastructure element (change an instance shape, revert to a checkpoint, etc) depends on the Cloud servers load and the parameters of the element.

It is possible that the procedure of reverting to checkpoints that were created over 1 month ago will be performed with issues; this is also true for huge checkpoints. To prevent possible issues, we recommend to review the checkpoints in your infrastructure regularly.

My VM seems to be stuck in transitional state. What should I do?

If your VM is in Transitional state for too long, it is possible that the VM info hasn't been updated yet. In this case, we would recommend you run or2din -i instance_id --force command that results in an immediate VM state info update.

I attached a storage volume to my VM, but the operation system does not see it. What should I do?

In such cases, you should set up the storage volume manually.

  • Linux: 1. Format the volume to the file system of your VM (for example, ext4):

    mkfs.ext4 /dev/sdb
    2. Mount the volume:

    mount /dev/sdb /mnt
    It is recommended to reboot your instance after the volume setup is finished.
    To make your system automatically set up any new additional volume, add the following line to /etc/fstab:

    /dev/sdb /mnt ext4 errors=remount-ro 0 0
  • Windows

To set up a new volume on Windows, go to Computer management -> Disk Management. The system will find a new volume and suggest to initialize it:

Right-click Disk1 and select "New Simple Volume"

Answer the wizard questions the way you like or just click Next button for default values. Then click Finish and wait for the volume to be formatted.

Can I move a personal VM to another project?

You can move your personal VM to a billable project. To do it, please, submit the Move VM to Another Project request. When your VM is moved, all personal project limitations are cancelled for it.

Please note that it is impossible to move a personal VM to another personal project.

How can I get a MacOs VM?

You can get both a virtual MacOS instance and a hardware Mac. All Mac resources are provided in the specially implemented EPAM-MAC region. Virtual MacOS instances are based on the MacOS 10.10 Yosemite operating system.

Instance details:

  • Image name: MacOS_10.10_Yosemite
  • Shape: Any standard EPAM Cloud shape (recommended - MEDIUM)
  • System Disk: 60GB
  • Billing: Same as in EPAM-BY1 region
  • Login: Use or2console command and user/user credentials
  • Limitations: checkpoints are not available
  • Use the standard or2run command to request a virtual MacOS instance.

Hardware Mac servers are also available with the following characteristics:

  • Image name: macmini_os_X
  • Shape: Large (2CPUs, 8Gb RAM) only
  • System disk: 300Gb, not extendable
  • Billing: 30$ per month, calculated hourly
  • Login: via the VNC console
  • Limitations: only or2run, or2kill, or2report, or2settag, or2din commands are available

Use the standard or2run command to request a hardware Mac server.

How do I know the Device name for attaching volumes to AWS instances?

To obtain the --device parameter required for attaching volumes in AWS regions, log in to your VM and run the 'df-h' command. The response to this command will contain the following information:

Find the device name for the root mount point in the 'Filesystem' column (in the example above, the device name is highlighted in bold and underlined). The device name can have one of the following formats: '/dev/xvda1' or '/dev/sda1'. Modify the device name by replacing the last character with any letter following it alphabetically. For example, if the device name is 'dev/xvda1', the modified name may have the '/dev/xvdd' or '/dev/xvde' forms; for 'dev/sda1' - '/dev/sdd' or '/dev/sde'.

Use the modified device name in the commands related to volumes manipulations, for example:

or2addattvol -p project -r AWS-region -s 50 -i instance_id -d /dev/sdd

I didn't find the answer to my question here. Where should I address?

If this section does not contain the answer to your question, please, address our Consulting Team.