Instances#

General information#

K2 Cloud provides computing resources as instances. The Instances subsection is used to create virtual machines, change their parameters and advance settings, and display a list and details of all instances created under the project and VPC.

Each instance has several parameters:

ID — the unique instance identifier in K2 Cloud.
Status — the current status of the instance. An instance can have the following statuses:
- stopped
- error
- crushed
- failed
- pending
- running
- terminated
- shutting-down
Name tag is a tag with the Name key and a value set by a user. To learn more about tags, please see Tags section.
Type – the instance type.
Availability zone — the information on the data center in which physical hardware is located.
Subnet — the subnet in which the instance is located.
Created — the instance creation date.
SSH-key — the specific SSH-key assigned to an instance.
Elastic IP — the external IP address of the primary network interface of an instance.
Private IP — the private IP address of the primary network interface of an instance.
Placement group — the placement group that the instance belongs to.

Click the icon on the right to search bar to go to instance parameter display settings.

High availability#

The automatic migration mechanism ensures high instance availability: the instance will be migrated to a backup node if the physical server on which it runs fails.

Billing#

An hourly fee is charged for each instance. This rate consists of:

CPU generation;
vCPU count;
RAM size;
type of OS installed: commercial (proprietary) or free, including free licensed ones;
guaranteed CPU availability: 20%, 50% or 100%. 1

1: The above values of 20% and 50% apply to Shared Core family instances.

Additionally, the price of instances in the GPU Accelerated Computing family also depends on the type and number of graphics cards. Fixed monthly payment is also available. Placing long-term (3, 6, 12 months or more) orders for this family instances is eligible for a discount.

Available quotas#

The following default quotas are allocated to instances:

50 instances in each availability zone in a project;
4 simultaneously running instance export tasks for each user;
3000 GiB total RAM in a project;
3000 vCPU cores in a project.

If necessary, you can increase quotas. To do this, contact the support service via the support portal or by email support@k2.cloud.

Instance types#

K2 Cloud offers users a wide variety of instance types with different combinations of computing resources and memory sizes.

Instance families#

All instance types are grouped and ordered by family. You can select types from the following families: General Purpose for common computational tasks, HPC optimized for resource-intensive applications, GPU Accelerated Computing for ML, AI and LLM tasks, and Shared Core with a guaranteed vCPU availability of less than 100%.

For all instances except for the Shared Core family, 100% resource availability is guaranteed. Several instances of the Shared Core family share a single physical core, with each instance guaranteed a certain percentage of processor time: minimum 50% or 20%, depending on the instance type. It helps reduce costs for services and tasks for which full processor availability is not required. If resources are free, a specific instance can be allocated up to 100% processor time.

The table provides recommendations on instance types for various tasks.

Family	Description	Application
General Purpose	These instances are suitable for wide range of tasks. They offer an optimal ratio of computing power and memory size.	Small databases Microservices Development environments Servers for cluster computing High performance applications Memory-intensive tasks
HPC Optimized	These instances deliver ultra-high performance for compute-intensive workloads. These instances are based on 2nd Generation Intel® Xeon® Scalable Processors with the Cascade Lake architecture.	HPC Enterprise databases, such as 1C, etc.
GPU Accelerated Computing 2	Instances with NVidia GPUs for ML, AI, and LLM tasks.	Generative AI model training 3 Application of trained models (inference) Image analysis, 3D simulation, rendering, and video processing HPC
Shared Core	Instance with a guaranteed processor time availability of 50% and 20%. This is for services and tasks for which full processor availability is not required.	Auxiliary and background computational tasks Automated processes and recurring jobs Test and development benches Runtime environments for CI/CD pipelines Monitoring and data collection services

2: GPU Accelerated Computing instance family is available on request. In an e-mail to your manager or support service support@k2.cloud, specify the desired instance type parameters: model and number of graphic cards, number of vCPUs, and RAM size.
3: If you need out-of-the-box tools to create, train, and use machine and deep learning models, you can use K2 Neurotech-provided services.

CPU generations#

K2 Cloud instances are powered by Intel® and AMD® CPUs. Hardware characteristics for each generation are shown in the table.

Important

In the ru-spb region, only instances powered by gen5 and gen7 CPUs are available.

Generation 4	CPU	Architecture	Characteristics
gen5	Intel® Xeon® Gold 6254	Cascade Lake	Max Turbo Frequency 4.00 GHz
gen5p	Intel® Xeon® Gold 6244	Cascade Lake	Max Turbo Frequency 4.40 GHz
gen7	Intel® Xeon® Gold 6548y+	Emerald Rapids	Max Turbo Frequency 4.10 GHz
gen7p	AMD® EPYC® 9275F	Zen 5	Max Turbo Frequency 4.80 GHz

4: Gen3 (Intel® Xeon® Gold 6140) and gen4 (Intel® Xeon® Platinum 8168) processors are gradually being phased out.

All our servers have the Turbo Boost option enabled. This allows instances to get the most out of the CPU and run at the maximum CPU frequency. Hypervisor does not allow passing this information to an instance. Therefore, if you try to check CPU frequency using OS tools from within the instance, you will see the base CPU frequency.

Graphic cards#

For GPU Accelerated Computing instances, the following graphic cards are available:

Card 5	Architecture	GPU Memory	GPU Bandwidth
NVIDIA H100 Tensor Core GPU	Hopper	80 GB	2 TB/s
NVIDIA L40S GPU	Ada Lovelace	48 GB	864 GB/s
NVIDIA A100 Tensor Core GPU	Ampere	80 GB	1.935 GB/s
NVIDIA T4 Tensor Core GPU	Turing	16 GB	300 GB/s

5: For different graphic cards, please inquire your manager regarding their availability.

Instance types with multiple graphic cards can be added to the GPU Accelerated Computing family upon request. Each particular instance type can only have graphic cards of the same type.

Creating an instance#

You can launch an instance using one of the following methods:

using instance wizard;
using a launch template;
using an existing instance;
using a recovery point;
using an instance image.

Create an instance using the wizard#

To launch the instance wizard:

In the top menu of the web interface, select the region where you want to create an instance.
Expand the side menu by clicking .
Go to Compute Instances. To find the section quickly, you can use the menu search.
Click Create instance. To configure the instance, follow the steps below.

Step 1. Select an image#

To launch an instance, you need to select a pre-created image with the operating system you want to use, for example, Windows, Debian, or Ubuntu.

The following groups of pre-created images are available in the Image selection block:

K2 Cloud images — images with pre-installed system software, which are provided by K2 Cloud.
Partner images — images with pre-installed system software, which are provided by our partners.
Own images — images created in the current project.
Available images — own images and images created in other projects and accessible from the current project.

Note

Own and available images can also include images created during the backup process. To display them, enable Show recovery points. However, we recommend creating an instance from a recovery point directly, in which case you won’t need to specify its parameters again.

To quickly find an image, you can also search by ID, name, or version. Select a pre-created image and click Next to confirm the choice and go to the next step.

Step 2. Select the instance name, placement, and type#

At this step, you need to select where to place an instance in the cloud and which instance type to use. In addition, you can specify an instance name.

In the Placement block, specify the following parameters:

Name tag — (Optional). An arbitrary instance name.
Availability zone — If a region has multiple availability zones, you can select a zone where the instance will be created.
Placement group — When you select a placement group, its member instances will be launched on different computing nodes. This helps minimize the number of failed instances if one of the physical nodes fails. To use this feature, you first need to create a placement group.

Important

Placement groups are not supported for instances of the Shared Core family.
Placement type — If the default value (Shared server) is selected, instances will be started on a shared computing node where instances of other users can be deployed. To launch an instance on a dedicated host, order it first. More information about starting an instance on a dedicated host can be read in the respective section of the documentation.
Number of instances — The number of instances to be created (1 by default).

In the Computing resource block, specify the required configuration and/or instance type on one of the tabs.

Configure yourself

On this tab, you can select the instance family and specify the required configuration more precisely:

General Purpose

If this instance family is selected, specify the required processor generation (generations gen5, gen4, and gen3 are available), the number of vCPUs, and the amount of RAM.
Shared Core

If this instance family is selected, specify the required processor generation (generations gen5 and gen3 are available). For the gen5 generation, you can also specify the number of vCPUs. For the gen3 generation, the number of vCPUs is set automatically.

Also, when the Shared Core family is selected, you can choose the guaranteed CPU availability level: 20% or 50%.
HPC Optimized

If this instance family is selected, specify the required number of vCPUs and the amount of RAM. For this family, only instances based on gen5p generation processors are available.
GPU Accelerated Computing

When this instance family is selected, specify the required number of GPUs and their type.

Specify the instance code

An instance type describes a ready-made configuration. All instance types are grouped by family. Select the appropriate type from the drop-down list.

Note

If you are not sure which instance type is best for your workload, start with the General Purpose instance family. In K2 Cloud, you can change instance type if necessary. For optimal performance, we recommend using the instance types of the current generation gen7 .

To quickly find the required instance type, you can search by instance type, CPU generation, vCPU count, RAM size and GPU type. Select a predefined instance type from the list according to your needs and click Select.

Note

For Windows, we recommend choosing instance types with at least 4,096 MB RAM.

Step 3. Set the network parameters#

In this step, select the VPC and set the parameters of the network interfaces that will be attached to the instance.

Note

For information, the system displays the details of the availability zone selected in step 2.

In the VPC block, select:

Hostname type — Determines the principle by which the instance internal DNS name will be generated.

When the Private IP address option is selected, the instance hostname is generated based on the instance private IP address. Example: ip-192-168-0-1.vpc-12345678.internal.

When the Instance ID option is selected, the instance hostname is generated based on the instance ID. Example: i-12345678.vpc-12345678.internal.
VPC — VPC where the instance will be created.

Note

If the instance is to be attached to a virtual switch, select Non VPC.

Network interface — It allows you to connect an instance to a subnet or switch. By default, a new instance will have a network interface with 0 index. This primary network interface cannot be detached from the instance.

In the Network interface 0 block, set the following parameters:

Hint

To display the block, click Expand all. To hide it, click Collapse all.

Network interface — You can select an existing interface from the list of available interfaces or keep the field intact. A new interface will be created then.

Note

The existing interface will not be deleted together with the instance.
Network — (For a new interface only) Select the subnet or switch, which the network interface will be attached to. If you need to attach the interface to a new subnet, click Create subnet and specify the subnet parameters.
Elastic IP — (For a new interface only) You can set an external IP address only if the VPC where the instance is located has an attached Internet gateway.

To associate an EIP, open the Associate Elastic IP tab if necessary:
- if there are several available Elastic IP addresses, an address will be selected and associated automatically;
- if there is only one available Elastic IP address, information about the address being associated will be displayed;
- if there are no available Elastic IP addresses, click Allocate Elastic IP and specify the address pool (default, user, or DDoS-protected) from which the Elastic IP will be allocated, and (optional) tags.
If associating an EIP is not required, select the No address tab.

Note

If Elastic IP auto-association is enabled for the subnet, the Associate Elastic IP tab is selected by default. If auto-association is disabled, the No address tab is selected by default.

For a new interface, you can also specify a description and internal IP address instead of the automatically assigned values, as well as configure deletion together with the instance. To do this, expand the Additional block and fill in the required fields.

Security groups — (For a new interface only) You can select up to five security groups. If the field is left blank, a default group from the selected VPC will be associated with the instance.

Note

Security groups can only be associated with a network interface attached to a subnet.
Description — The brief description of the interface, for example, which instance it is attached to.
Private IP — The internal IP address of the primary interface at which the instance will be accessible. If the field is left empty, it will be assigned automatically.

Note

An internal IP address can only be associated with a network interface attached to a subnet.
Delete with instance — This option specifies whether to delete the network interface along with the instance (enabled by default).

In this step, you can also create and attach additional network interfaces to the instance. To do this, click Add and set the required parameters.

You can attach max. 16 network interfaces to an instance. For more information on network interfaces, see Network Interfaces.

To go to the next step, click Next.

Step 4. Select volumes#

In this step, set parameters of the volumes that will be created with the instance.

In the Volume 1 (Boot device) block, specify the boot volume configuration:

Parameter	Description
Snapshot	A snapshot from which you are going to create an instance (by default, snapshots that belong to the selected images are shown).
Size	Volume size in GiB (1 GiB = 1,024 MiB = 2³⁰ byte).
Volume type	Sets the desired volume type. For more information on volume features and parameters see page Volumes.
IOPS	The maximum total number of I/O operations per second. For volumes types gp2, st2 and st3 the number of IOPS is calculated automatically, for the volume type io2 you can specify the required number of IOPS.
MiB/s	The maximum total throughput, in MiB/s, at which data can be read from or written to a volume. It is calculated automatically.
Delete with instance	Specifies whether the volume must be automatically deleted after the instance is deleted.
Encryption 7	(Only for `st3` volumes) Specifies whether the volume should be encrypted.

7: Volume encryption is in beta testing and is currently available only on request.

A boot volume cannot be detached from an instance.

To add another volume, click Add and specify the required parameters. You can attach up to 16 volumes to instances with the hvm virtualization type and up to 4 volumes to instances with the hvm-legacy virtualization type.

Once you have added all volumes, click Next.

Step 5. Set additional parameters#

In this step, you can add an SSH key, set a username and password to access the VNC console, select the virtualization type, enable RAM monitoring, configure instance deletion on shutdown, enter user data, and assign tags to the instance and volumes that will be created together with the instance.

Important

The login and password can only be set for a Linux instance on which the EC2 agent is installed (see the installation instructions). In addition, the instance must be attached to one subnet at least.

SSH key — The SSH key is used to authenticate the user when accessing the instance. The public part of the key must be first uploaded to the cloud. You can add the key after the instance has been launched as well.
Username — The username for accessing the VNC console. It can also be used for SSH access. The username must begin with a Latin letter or _, may contain lowercase Latin letters, digits, - and _ and be 1 to 32 characters long. Once set, the username cannot be changed.
Password — The user password for accessing the VNC console (it must meet the password strength requirements).
Virtualization type — You can choose from the two virtualization types. The hvm virtualization supports VirtIO drivers and optimizes the performance of the instance OS (which must support VirtIO drivers). All K2 Cloud images are compatible with this virtualization type.

The hvm-legacy virtualization without VirtIO support ensures correct operation of an instance whose operating system does not support VirtIO drivers.
Enable RAM monitoring — Selecting this option enables RAM monitoring. RAM monitoring is available for Linux instances only. In addition to enabling the monitoring, you need to install the CloudWatch agent on them (see details). The checkbox is cleared by default.
Delete on shutdown — Specifies whether the instance should automatically be terminated after the shutdown. The option is unchecked by default.
User data — This data describes the operations to be performed automatically at the instance launch. For details, see the Instance Metadata API documentation section.

In this step, you can assign tags to the instance and/or volumes. If the Name tag was not specified in step 2, enter its value for the predefined Name key. If you also want to assign the Name tag to the attached volumes, check the Assign to volumes checkbox.

To assign additional tags, click Add tag and enter the tag keys and (optional) tag values. If you also want to assign the tag to the attached volumes, check the Assign to volumes checkbox.

To create an instance, click Create instance.

Instance launch takes some time. Upon completion of the instance launch process, its status will change from Pending to Running. Each created instance is assigned a unique ID in the i-xxxxxx format.

Many instance parameters can be changed without stopping the instance. Click a link with unique ID to go to the instance’s page in order to make changes, attach or detach additional volumes, reboot or shut down an instance, change public address, etc.

Create an instance from a launch template#

Important

Creating an instance from a launch template is not supported in the new web interface yet. Use the previous version of the web interface.

You can also create an instance from a template in the Launch Templates section.

Create an instance from another instance#

You can create a required number of instances with the same parameters as an existing instance has.

In the top menu of the web interface, select the project and region where you want to create instances.
Expand the side menu by clicking .
Go to Compute Instances. To find the section quickly, you can use the menu search.
In the instance table, select an instance with the required parameters to be used as a reference configuration.
In the action panel, select Create an additional instance. If this button is not displayed, open the additional action menu by clicking More.
In the window that opens, specify the number of instances to be created.
- If the instance should be deleted after shutdown, enable the Delete on shutdown option.
- To enable RAM monitoring, check the RAM monitoring checkbox.
  
  Important
  
  RAM monitoring is only available for Linux instances with CloudWatch agent installed (for details, see).
- If additional instance configuration during launch is required, enter the necessary data in the User data field.
- You can also change the Name tag, if it has been set, and specify other tags. If the Name tag has not been set, you can set its value.
  
  Note
  
  If other tags were assigned to the original instance, they will be ignored.
  
  If you need to assign other tags, click Add tag to set the next tag. Specify the tag key and its value. To assign the tag to both the instance and the volumes, check the Assign to volumes checkbox.
After setting the required parameters, click Create.

Deleting an instance#

Attention

Before deleting an instance, make sure you do not lose any important data. If you need to save a volume copy for future use in K2 Cloud, you can create a snapshot of a volume. If you need a volume or instance copy to boot from and run in your local infrastructure, use the export function.

In the top menu of the web interface, select the project and region where you want to delete an instance.
Expand the side menu by clicking .
Go to Compute Instances. To find the section quickly, you can use the menu search.
Select the instance in the resource table and click Delete on the pop-up action bar. You can select several instances for deletion at once.
To confirm the deletion, click Delete.

Note

Volumes and network interfaces for which the Delete with instance checkbox is checked, as well as all instance tags, will be deleted along with the instance.

Connecting to an instance#

Getting the console output#

To view the current state of the instance launch process or to get the password:

In the top menu of the web interface, select the project and region where you want to get the instance console output.
Expand the side menu by clicking .
Go to Compute Instances. To find the section quickly, you can use the menu search.
Find the instance in the resource table and click the resource ID to go to its page.
Click View output on the action panel. If the button is not displayed, click to open the additional action panel.

Accessing the remote console#

Important

Access to the remote console is not supported in the new web interface yet. Use the previous version of the web interface.

Enabling access to the remote console#

You can grant any K2 Cloud user with a remote access to the instance.

In the top menu of the web interface, select the project and region where you want to grant access rights to the instance.
Expand the side menu by clicking .
Go to Compute Instances. To find the section quickly, you can use the menu search.
Find the instance in the resource table and click the resource ID to go to its page.
Click Share on the action panel. If the button is not displayed, click to open the additional action panel.
In the window that opens, set the following parameters:
- email – recipient’s email address;
- duration — period in hours for which access is granted.
After setting the parameters, click Share.

A link to access the remote console will be sent to the specified email.

Important

Access to the remote console can only be granted to the users that have user account in cloud.

Other methods#

You can read about other ways to connect to an instance in Instructions section of documentation.

Actions on instances#

Rebooting an instance#

Rebooting an instance is equivalent to rebooting an OS. In most cases, rebooting an instance takes only a few minutes. To reboot the instance, while gracefully stopping all services:

In the top menu of the web interface, select the project and region where you want to reboot an instance.
Expand the side menu by clicking .
Go to Compute Instances. To find the section quickly, you can use the menu search.
Select the instance in the resource table and click Reboot on the pop-up action bar.
Confirm the reboot of the instance.

You can also reboot the instance on its page.

Activating hibernation#

Important

We do not guarantee that an instance without a proper preconfiguration will enter the hibernation mode after clicking the Hibernate button.

Hibernation is a state in which the memory content of an instance, applications, and processes is saved to the local volume of the instance before it enters the stopped state. It differs from stopping an instance as follows:

To hibernate an instance, its operating system must be correctly configured;
In the hibernation mode, instances save memory and application states that they had when hibernation was initiated;
Hibernation can be used for instance pre-warming if the instance or application takes a long time to boot.

When you hibernate a correctly configured instance, K2 Cloud signals the OS to perform hibernation. To start hibernation, the guest OS uses an installed QEMU Guest Agent. The cloud communicates with the agent and sends it the appropriate command, where thereafter QEMU Guest Agent prompts the OS to enter the hibernation mode. In this mode, the OS saves the memory content of the instance to a volume.

Important

Please note that the volume size must be larger than the RAM size of the instance.

Attention

We do not recommend resizing the volume or changing the type of hibernated instance. Otherwise, the OS will boot normally, and the saved VM memory state will be ignored.

When you start this instance again, it will enter the Pending state, the memory content will be downloaded from the volume, and the states of processes and applications will be restored.

How fast an instant enters/exits the hibernation mode depends on the volume performance. To accelerate the process, you can use higher-performance volume types, such as gp2: Universal (SSD) and io2: Maximum (SSD).

Using hibernation mode reduces your cloud cost during instance suspension periods since the payment is only charged for storing its volumes and Elastic IP.

To hibernate an instance:

In the top menu of the web interface, select the project and region where you want to hibernate an instance.
Expand the side menu by clicking .
Go to Compute Instances. To find the section quickly, you can use the menu search.
Select the instance in the resource table and click Hibernate on the pop-up action bar.
To send a command to hibernate the instance, click Hibernate.

You can also hibernate the instance on its page.

Stopping an instance#

In the top menu of the web interface, select the project and region where you want to stop an instance.
Expand the side menu by clicking .
Go to Compute Instances. To find the section quickly, you can use the menu search.
Select the instance in the resource table and click Stop on the pop-up action bar.
To stop the instance, click Stop.

You can also stop the instance on its page.

Important

If the operating system of the instance does not shutdown within a few minutes, do a forced shutdown (it is similar to turning off the server’s power). To do this, click Stop (force).

Some instance parameters can only be changed when it is stopped.

Creating an image#

Note

Creating an image from an instance that has encrypted volumes is not supported.

An image can be created from both a stopped and a running instance. To create an instance image that can later be used to create new instances with the same parameters:

In the top menu of the web interface, select the project and region where you want to create an image from an instance.
Expand the side menu by clicking .
Go to Compute Instances. To find the section quickly, you can use the menu search.
Select the instance in the resource table and click Create image on the pop-up action bar. If this button is not displayed, open the additional action menu by clicking More.
In the window that opens, enter the image name and description.
If you need to set a tag, click Add tag and specify the tag key and value. To assign more tags, click Add tag again.
After setting the parameters, click Create.

You can find the ready image in Compute Images.

An Image is created in two steps. First, snapshots of all attached volumes are taken simultaneously. This takes a fraction of a second and does not block the running instance.

In the second step, snapshots are written in parallel to object storage. The duration of this stage depends on the maximum amount of data on the volumes and may take a long time. All this time, the image will be in the pending status.

The instance can be restarted, stopped or even deleted without waiting for the second image creation stage to complete. But you can create another image from the instance only after the image creation process is complete and the image status changes to Available. The current image status can, for example, be viewed on its page.

Exporting an instance#

Important

When exporting a running instance, snapshots of all volumes are taken at the same time. All data written to the volume is saved, but data in RAM is not included in the snapshots. To include all data in the image, we recommend flushing the cache and suspending writes to the volumes before exporting.

Note

Before exporting an instance, create a bucket in the Object storage section.

Note

Exporting instances that have encrypted volumes is not supported.

To export an instance in a bucket:

In the top menu of the web interface, select the project and region where you want to export an instance.
Expand the side menu by clicking .
Go to Compute Instances. To find the section quickly, you can use the menu search.
Select the instance in the resource table and click Export on the pop-up action bar. If this button is not displayed, open the additional action menu by clicking More.
In the window that opens, set the following parameters:
- prefix;
- export format;
- description;
- bucket in which the instance will be exported.
Select the appropriate option and enter your email address if you want to be notified when the import process is complete.
After setting all the required parameters, click Export.

You can also export the instance on its page.

Instance export will take some time. Afterwards, the instance image will appear in Object storage. See details on the Import/Export documentation page .

Attach an instance to the Auto Scaling group#

You can attach one or more instances to an existing Auto Scaling group.

Changing instance parameters#

After the instance starts successfully, you can make changes to it. Below are some parameters you can change to suit your needs.

Changing the instance type#

All instance types are grouped and ordered by family. You can choose the desired combination of computing power and memory size and optimize your instance for specific tasks or business needs.

Important

This action is available only for an instance in the Stopped status.

If the current instance type does not suit your workload, you can change it. To do this:

In the top menu of the web interface, select the project and region where you want to change the instance type.
Expand the side menu by clicking .
Go to Compute Instances. To find the section quickly, you can use the menu search.
Find the instance in the resource table and click the resource ID to go to its page.
In the Information tab, click next to the Instance type field.
In the window that opens, select an instance type from the drop-down list.
To change the instance type, click Apply.

Associate/disassociate an Elastic IP address with/from an instance#

Important

An Elastic IP can be associated with the instance only if the primary interface was created in a subnet and if the VPC has an attached internet gateway.

Note

To assign a different Elastic IP to an instance, first release the assigned IP address from the primary interface.

To associate an instance with a public address from the list of allocated IP addresses:

In the top menu of the web interface, select the project and region where you want to associate a public address with an instance.
Expand the side menu by clicking .
Go to Compute Instances. To find the section quickly, you can use the menu search.
Select the instance in the resource table and click Associate Elastic IP on the pop-up action bar.
In the window that opens, you can associate an available Elastic IP address or, if there are no available addresses, allocate and associate an Elastic IP address.
Associate an Elastic IP address
To associate an available Elastic IP address:
- open the Select Elastic IP tab;
- select an address from the drop-down list;
- click Associate.
Allocate and associate
To allocate and associate an Elastic IP address:
- open the Associate Elastic IP tab;
- select the IP address pool from which the address should be allocated;
- if you need to assign tags, click Add tag. Specify the tag key and its value. To assign more than one tag, click Add tag again;
- click Allocate and associate.

You can also associate an Elastic IP on the instance page. The address will be associated with the instance primary interface.

Note

Once the Elastic IP is associated, the network interface becomes available at the specified address within a minute.

To disassociate a public IP address from the primary interface of an instance and make the address available for use:

In the top menu of the web interface, select the project and region where you want to release a public address from an instance.
Expand the side menu by clicking .
Go to Compute Instances. To find the section quickly, you can use the menu search.
In the resource table, select the instance and click Disassociate Elastic IP on the pop-up action bar.
In the window that opens, confirm the action by clicking Disassociate.

You can also disassociate the associated Elastic IP from the instance primary interface on the instance page `.

How to associate/disassociate an Elastic IP with/from an arbitrary instance interface, is described below.

Changing the SSH key#

Important

This operation adds a new key each time a key is changed but old key is not removed. Reboot an instance for changes to take effect.

To assign a new SSH key to an instance:

In the top menu of the web interface, select the project and region where you want to assign an SSH key to an instance.
Expand the side menu by clicking .
Go to Compute Instances. To find the section quickly, you can use the menu search.
Find the instance in the resource table and click the resource ID to go to its page.
On the Information tab, do the following:
- click next to the SSH key field;
- select a new SSH key from the drop-down list;
- click to save the changes.

Set up a password#

Important

To set up a password, install and configure the EC2 agent on the instance according to the instruction.

To set the VNC console password:

In the top menu of the web interface, select the project and region where you want to set up a password for the instance.
Expand the side menu by clicking .
Go to Compute Instances. To find the section quickly, you can use the menu search.
Select the instance in the resource table and click Set password on the pop-up action bar. If this button is not displayed, open the additional action menu by clicking More.
To set a password, in the dialog that opens enter the password in the Password field or click Generate password.
Click on Save to apply the changes.

To modify or delete a previously set VNC console password:

In the resource table, select the instance and click Change password on the pop-up action bar. If this button is not displayed, open the additional action menu by clicking More.
In the window that opens, enter a new password in the Password field or click Generate password to change the password. To delete the password, click Delete password.
Click on Save to apply the changes.

You can also manage the password on the instance page. To do this, go to the resource page and click Set password or Change password.

Note

The button name and the available action (setting or changing a password) depend on whether a password was set earlier.

Changing user data#

To change the instance’s user data after creation:

In the top menu of the web interface, select the project and region where you want to change user data.
Expand the side menu by clicking .
Go to Compute Instances. To find the section quickly, you can use the menu search.
Find the instance in the resource table and click the resource ID to go to its page.
On the Information tab, do the following:
- click in the User data block;
- in the window that opens, edit the previously entered data or paste new data into the input field;
- click on Save to apply the changes.

For details, go to the Instance Metadata API documentation section.

Attaching/detaching network interfaces#

Instances connect to subnets/switches via network interfaces. You can manage interfaces in the subsection Network Interfaces or via API.

To attach a network interface to an instance:

In the top menu of the web interface, select the project and region where you want to attach a network interface to an instance.
Expand the side menu by clicking .
Go to Compute Instances. To find the section quickly, you can use the menu search.
In the resource table, select the instance and click Attach network interface on the pop-up action bar. If this button is not displayed, open the additional action menu by clicking More.
In the window that opens, select a network interface from the drop-down list.
To complete the operation, click Attach.

Alternatively, go to the instance page and click Attach network interface.

An interface can be attached to only one instance at a time.

The following restrictions apply to attaching network interfaces:

multiple interfaces may not be attached to the same instance if they belong to the same subnet or switch;
if the primary network interface of an instance is a virtual switch network interface, then a subnet interface may not be attached to the instance;
up to 16 network interfaces (including a primary interface) may be attached to an instance;
attached interfaces should be in the same availability zone and belong to the same VPC as the instance itself.

To ensure proper operation of the attached interfaces, we recommend performing a minimum configuration of the interfaces or using K2 Cloud images. In these images interfaces are configured automatically. Please see Configure the attached interfaces.

To detach a network interface from an instance:

In the top menu of the web interface, select the project and region where you want to detach a network interface from an instance.
Expand the side menu by clicking .
Go to Compute Instances. To find the section quickly, you can use the menu search.
In the resource table, select the instance and click Detach network interface on the pop-up action bar. If this button is not displayed, open the additional action menu by clicking More.
In the window that opens, select a network interface from the drop-down list.
To complete the action, click Detach.

Alternatively, go to the instance page and click Detach network interface.

Configure hot attachment/detachment of network interfaces#

Interfaces can be attached to/detached from both stopped and running instances. All official K2 Cloud images support hot attachment/detachment of interfaces.

The list of Linux distributions that we have tested to support “hot” attachment/detachment of network interfaces:

CentOS/RHEL/OEL 6/7
Ubuntu ≥ 12.04
Debian ≥ 7
Fedora ≥ 19

If you use CentOS/RHEL/OEL 5.x or Ubuntu 10.x, please download the acpiphp kernel module and add it to the autoload.

This functionality is not guaranteed for instances launched from other images.

If you use other Linux distributions, then make sure the kernel is compiled with HOTPLUG_PCI_ACPI support:

[root@host ~]# grep 'CONFIG_HOTPLUG_PCI_ACPI=' /boot/config-`uname -r`
CONFIG_HOTPLUG_PCI_ACPI=y

If the command output is CONFIG_HOTPLUG_PCI_ACPI=y, it means that “hot” attachment/detachment of network interfaces are supported by this operating system “out of box”.

If the result of the command CONFIG_HOTPLUG_PCI_ACPI = m, then hot attach/detach of network interfaces requires downloading acpiphp or pci_hotplug module, depending on OS.

If the command output is a blank line, then the kernel does not support “hot” attachment/detachment of network interfaces.

Attention

If you detach a network interface from an instance and the instance does not support “hot” attachment/detachment of PCI devices at the OS level, the interface will remain in the Detaching internal state until the instance is stopped.

The cloud will return an error to subsequent attempts of hot attachment/detachment of this network interface.

Change the primary network interface#

Important

You can change the network interface only if the instance is stopped.

If you need to move an instance to another subnet or VPC, or from non-VPC to VPC, change the primary network interface of the instance. To do this, you need to have an unattached network interface in the target subnet or VPC. If there is no such interface, create it first. If necessary, assign the new interface an internal IP address of your choice from the target network.

The network interface to be used instead of the primary one should be:

in the same availability zone as the source one;
created in a subnet;
not attached to any instance.

Note

If you are going to move an instance to another VPC, it should not have any other interfaces in the subnet other than the primary one.

To change a primary network interface:

In the top menu of the web interface, select the project and region where you want to change the primary network interface of an instance.
Expand the side menu by clicking .
Go to Compute Instances. To find the section quickly, you can use the menu search.
In the resource table, select the instance and click Change primary network interface. If this button is not displayed, open the additional action menu by clicking More.
In the window that opens, select a network interface from the drop-down list.
To complete the action, click Change.

You can also change the primary network interface on the instance page.

Important

Primary interface replacement does not include the transfer of its associated Elastic IP. If you want the instance to inherit the old Elastic IP, then first disassociate it from the old primary interface and then manually associate it with the new one.

Associate/disassociate an Elastic IP address to/from the network interface#

Important

Associating and releasing an Elastic IP address for a specific network interface is not supported in the new web interface yet. Use the previous version of the web interface.

Enable/disable RAM monitoring#

RAM monitoring is only available for Linux instances with CloudWatch agent installed (for details, see).

To enable/disable RAM monitoring:

In the top menu of the web interface, select the project and region where you want to enable or disable RAM monitoring for an instance.
Expand the side menu by clicking .
Go to Compute Instances. To find the section quickly, you can use the menu search.
Find the instance in the resource table and click the resource ID to go to its page.
Move the RAM monitoring switch to the desired position in the Information tab.

Changing security groups#

Important

This action is allowed only for the network interface created in a subnet.

Change security groups of the primary network interface#

In the top menu of the web interface, select the project and region where you want to change security groups of an instance network interface.
Expand the side menu by clicking .
Go to Compute Instances. To find the section quickly, you can use the menu search.
Find the instance in the resource table and click Change security groups on the pop-up action bar. If this button is not displayed, open the additional action menu by clicking More.
In the window that opens, select one or more security groups from the list.
- To assign a security group to a network interface, set next to the selected group.
- To cancel the assignment, clear the checkbox.
To save changes, click Change.

You can also change the security groups of the primary network interface on the instance page in the additional actions menu (to open it, click ).

Change security groups of additional network interfaces#

In the top menu of the web interface, select the project and region where you want to change security groups of an instance network interface.
Expand the side menu by clicking .
Go to Compute Instances. To find the section quickly, you can use the menu search.
Find the instance in the resource table and click the resource ID to go to its page.
On the Network and security tab, select the network interface whose security groups you want to change and click Change security groups on the pop-up action bar.
In the window that opens, select one or more security groups from the list.
- To assign a security group to a network interface, set next to the selected group.
- To cancel the assignment, clear the checkbox.
To save changes, click Change.

Note

If source/destination check (source-dest-check) is disabled on the selected interface and at least one of the added security groups is a source in the inbound rules of this or any other group, then you cannot associate such security groups with the network interface.

Change placement group#

Important

This action is available only for an instance in the Stopped status.

To move an instance into another placement group:

In the top menu of the web interface, select the project and region where you want to change the placement group for an instance.
Expand the side menu by clicking .
Go to Compute Instances. To find the section quickly, you can use the menu search.
In the resource table, select the instance and click Change placement on the pop-up action bar. If this button is not displayed, open the additional action menu by clicking More.
In the window that opens, select:
- a placement group from the drop-down list;
- placement type — Shared server or Dedicated host (if available).
  
  If you select placement on a dedicated host, you must also specify the instance affinity, i.e. whether the instance will be placed on a particular host (the Target host option) or on any host (the Any host option).
To complete the action, click Change.

You can also change the placement group on the instance page.

Specifying tags#

To add, modify or delete instance tags:

In the top menu of the web interface, select the project and region where you want to set tags for an instance.
Expand the side menu by clicking .
Go to Compute Instances. To find the section quickly, you can use the menu search.
Find the instance in the resource table and click the resource ID to go to its page.
Open the Tags tab and click Edit to set tags.
If no tags have been set before, you can set them by filling the Key and Value fields. To assign the next tag, click Add tag.

To modify a tag, edit the required fields (Value and/or Key) of the respective tag.

To delete a tag, click the icon next to the tag you no longer need.
To complete editing, click Save changes.

Operations with volumes#

The Volumes tab contains a list of all volumes attached to the instance. Each volume has a set of parameters:

ID — the unique volume ID.
State — the information on the volume status.
Volumes – the sequential number of the volume in this instance.
Size — the maximum amount of data that a volume can hold.
Name tag — a tag with the Name key and a value set by the user.
Volume type.
IOPS — the volume performance (IOPS).

To view volume details in the Volumes subsection, go to the instance page and click the link with its unique ID in the Volumes tab.

Attaching/detaching a volume#

Attach a volume#

Important

A volume can be attached to an instance if they are in the same availability zone.

Only volumes with Available status (which aren’t attached to any instance) can be attached to an instance.

To attach one more volume to an instance:

In the top menu of the web interface, select the project and region where you want to attach a volume to an instance.
Expand the side menu by clicking .
Go to Compute Instances. To find the section quickly, you can use the menu search.
Find the instance in the resource table and click the resource ID to go to its page.
Open the Volumes tab and click Attach.
In the window that opens, you can attach an existing volume or create and attach a new volume.
Attach a volume
- Open the Select volume tab.
- Select a volume from the drop-down list.
- Click Attach.
Create and attach a volume
- Open the New volume tab.
- Set the volume parameters:
  - Select the volume type from the drop-down list.
  - Specify the volume size.
  - (Only for io2 volumes) Set the performance in IOPS.
  - If you need to set a tag, click Add tag and specify the tag key and value. To assign more tags, click Add tag again.
- Click Create and attach.

Detach a volume#

Important

You can detach only non-boot volumes from the instance.

To detach a volume from an instance:

In the top menu of the web interface, select the project and region where you want to detach a volume from an instance.
Expand the side menu by clicking .
Go to Compute Instances. To find the section quickly, you can use the menu search.
Find the instance in the resource table and click the resource ID to go to its page.
Open the Volumes tab, select the volume in the resource table, and click Detach on the pop-up action bar.
To confirm the action, click Detach in the window that opens.

Changing the boot device#

You can select a boot volume (disk1) or a boot menu (menu) as the boot device. If the image from which the instance was created supports CD-ROM, you can also select CD-ROM (cd-rom).

Important

You can only change the boot device when the instance is in the Stopped status.

To change the boot device:

In the top menu of the web interface, select the project and region where you want to change the boot device of an instance.
Expand the side menu by clicking .
Go to Compute Instances. To find the section quickly, you can use the menu search.
Find the instance in the resource table and click the resource ID to go to its page.
On the Information tab, do the following:
- click next to the Boot device field;
- select a boot device from the drop-down list;
- click to save the changes.

Changing the boot volume#

Before changing the boot volume, you should first attach the volume you want to use as a boot one.

Important

You can only change the boot volume when the instance is in the Stopped status.

In the top menu of the web interface, select the project and region where you want to change the boot volume of an instance.
Expand the side menu by clicking .
Go to Compute Instances. To find the section quickly, you can use the menu search.
Find the instance in the resource table and click the resource ID to go to its page.
In the Volumes tab, select a volume that you want to make a boot volume.
Click Set as boot device on the pop-up action menu.
Confirm the boot volume selection by clicking Apply in the window that opens.

Creating a snapshot#

Note

If a volume is encrypted, you cannot create its snapshot.

You can create a snapshot of a volume attached to an instance at any time. To take a volume snapshot and then use it as an image for new volumes or data backups:

Note

For regular data backups, we recommend using Backup service.

In the top menu of the web interface, select the project and region where you want to create a snapshot of an instance volume.
Expand the side menu by clicking .
Go to Compute Instances. To find the section quickly, you can use the menu search.
Find the instance in the resource table and click the resource ID to go to its page.
In the Volumes tab, select a volume you want to create a snapshot of.
Click Create snapshot on the pop-up action bar.
In the window that opens, you can add a volume snapshot description, change the default Name tag, and assign other tags:
- enter a description in the Description field;
- if necessary, change the Name tag value;
- click Add tag to add the next tag and enter its key and value.
After setting the parameters, click Create.

You can also create a volume snapshot on its page.

Snapshot is created in two steps. At the first step, a snapshot of the volume state is created. This is a sub-second operation, which does not block volume reads/writes.

At the second step, the snapshot content is transferred to object storage. It can take a long time depending on the amount of data; meanwhile, the volume snapshot will be in the Creating state. The volume can be attached, detached, or deleted without waiting for this step to complete. You can only create another snapshot of the same volume after the snapshot has been fully transferred to the object storage and the snapshot state has changed to Completed. You can monitor the progress on the volume snapshot page.

Important

For all company projects and users, you can simultaneously create a maximum of 16 volume snapshots.

Important

To ensure volume snapshot consistency of a volume attached to a running instance, stop the instance or flush the cache and suspend writing to the volume.

Create a volume version#

To save the volume content at a certain point in time and then quickly restore it, you can create a volume version, which creates a volume snapshot. However, the snapshot content is not sent to another storage, so the version is created in a fraction of a second.

To create a volume version:

In the top menu of the web interface, select the project and region where you want to create a version of an instance volume.
Expand the side menu by clicking .
Go to Compute Instances. To find the section quickly, you can use the menu search.
Find the instance in the resource table and click the resource ID to go to its page.
In the Volumes tab, select a volume you want to create a version of.
Click Create version on the pop-up action bar.
In the window that opens, you can set the Name tag and assign other tags:
- enter the Name tag value;
- click Add tag to add the next tag and enter its key and value.
After setting the tags, click Create.

You can also create a volume version in the Volumes subsection or on the volume page. Read the volume version documentation to learn more about volume versions, in particular how to restore a volume from an existing version.

Important

As with volume snapshots, to ensure version consistency of a volume attached to a running instance, stop the instance or flush the cache and suspend writing to the volume.

Exporting a volume#

Important

If the volume is attached to a running instance, its snapshot may be in an inconsistent state. To avoid this, flush the cache and suspend writing to the volume before exporting if you do not want to stop the instance.

In the top menu of the web interface, select the project and region where you want to export an instance volume.
Expand the side menu by clicking .
Go to Compute Instances. To find the section quickly, you can use the menu search.
Find the instance in the resource table and click the resource ID to go to its page.
On the Volumes tab, select the volume you want to export.
Click Export on the pop-up action bar.
In the window that opens, enter the following data:
- prefix added to the automatically generated snapshot name;
- format of the snapshot to be created (vmdk, vhd, or qcow);
- description (you can keep the default one or add your own);
- the bucket to which the created volume snapshot will be exported.
If you want to receive a notification when the export is complete, select the Send email when complete checkbox and enter an email address.
To start the export operation, click Export.

Configure volume to be deleted upon instance deletion#

You can modify the Delete with instance option value, which was specified for a particular volume when the instance was created.

In the top menu of the web interface, select the project and region where you want to configure volume deletion together with the instance.
Expand the side menu by clicking .
Go to Compute Instances. To find the section quickly, you can use the menu search.
Find the instance in the resource table and click the resource ID to go to its page.
In the Volumes tab, find the required volume in the resource table and push the Delete with instance toggle to the desired position.

Alarm setup#

You can use alarms to configure the execution of a specified action when the instance metrics go beyond the specified thresholds.

Important

Setup of alarms for instances is not available in the new web interface yet. Please use the previous version of the web interface.

Instance information#

General information about instances can be viewed in the resource table in Compute Instances.

To view a summary table of all instances in the project, select All VPC in the VPC filter. To display instances from a particular VPC, select the desired VPC in the filter.

To go to the page of a particular instance, click its ID in the resource table in the Instances subsection. To facilitate the instance search, select its relevant VPC in the VPC filter or use the table search.

The Information tab provides the detailed information about the instance:

state;
instance name (Name tag);
instance image;
instance type;
SSH-key;
username;
password;
DNS name;
availability zone;
placement group;
placement type;
dedicated host (only for instances running on dedicated hosts);
affinity with the host (only for instances running on dedicated hosts);
virtualization type;
VPC;
network interfaces;
security groups;
network the primary network interface is attached to;
Elastic IP;
Private IP;
boot device;
boot menu status;
user data.

Here you can:

change the instance name (Name tag);
perform all basic operations and actions with the instance;
change its parameters;
use console;
enable/disable RAM monitoring;
set up password.

The Volumes tab displays a table with information about volumes attached to the instance:

volume ID;
volume name;
state;
size;
volume type;
volume IOPS performance;
whether the volume is encrypted;
Delete with instance option value;
device type: boot or additional volume.

The sequence and scope of the displayed data may vary depending on the table settings.

Here you can perform all basic volume operations and actions.

The Network and security tab contains a table with details of the network interfaces attached to the instance and security groups assigned to the network interfaces of the instance:

network interface ID;
network interface description;
subnet ID;
Elastic IP;
Private IP;
assigned security group IDs.

Here you can:

attach or detach network interfaces;
change the primary network interface;
associate/disassociate Elastic IP;
change a security group.

In the Metrics tab, you can view the graphs of the monitored metrics. You can set the period the metric graph covers, statistics, and metric calculation interval. To automatically refresh the graph, enable Auto-refresh; to refresh it manually, click Refresh.

Important

Automatic refresh is not possible when a custom period and/or interval is selected.

The Alarms tab displays details of alarms configured for the resource:

state;
alarm name;
triggering condition.

Alarms can be filtered by a state and/or metrics.

Here you can:

create a new alarm;
modify an alarm;
delete existing alarms with Delete button.

The Tags tab displays all tags assigned to the intance. You can add new, modify existing, and delete no-longer-needed tags.

Instances

In this article:

Instances#

General information#

High availability#

Billing#

Available quotas#

Instance types#

Instance families#

CPU generations#

Graphic cards#

Creating an instance#

Create an instance using the wizard#

Step 1. Select an image#

Step 2. Select the instance name, placement, and type#

Step 3. Set the network parameters#

Step 4. Select volumes#

Step 5. Set additional parameters#

Create an instance from a launch template#

Create an instance from another instance#

Deleting an instance#

Connecting to an instance#

Getting the console output#

Accessing the remote console#

Enabling access to the remote console#

Other methods#

Actions on instances#

Rebooting an instance#

Activating hibernation#

Stopping an instance#

Creating an image#

Exporting an instance#

Attach an instance to the Auto Scaling group#

Changing instance parameters#

Changing the instance type#

Associate/disassociate an Elastic IP address with/from an instance#

Changing the SSH key#

Set up a password#

Changing user data#

Attaching/detaching network interfaces#

Configure hot attachment/detachment of network interfaces#

Change the primary network interface#

Associate/disassociate an Elastic IP address to/from the network interface#

Enable/disable RAM monitoring#

Changing security groups#

Change security groups of the primary network interface#

Change security groups of additional network interfaces#

Change placement group#

Specifying tags#

Operations with volumes#

Attaching/detaching a volume#

Attach a volume#

Detach a volume#

Changing the boot device#

Changing the boot volume#

Creating a snapshot#

Create a volume version#

Exporting a volume#

Configure volume to be deleted upon instance deletion#

Alarm setup#

Instance information#