System and method for provisioning hyper-converged infrastructure
The automated provisioning of Ceph-based HCI on bare metal servers using cross-paired network interface cards and orchestration components addresses the labor-intensive nature of manual configuration, achieving high performance and availability in HCI setup.
Patent Information
- Application Number
- PCT/KR2023/006325
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2023-05-10
- Publication Date
- 2024-11-14
AI Technical Summary
Provisioning hyper-converged infrastructure (HCI) to bare metal servers using Ceph is labor-intensive and time-consuming, requiring manual configuration of operating systems, virtualization software, and cloud management software, with no automation for resource setup and high availability.
A system and method for automatically provisioning Ceph-based HCI on bare metal servers using high-speed network switches and Ceph-based operating systems, which includes components for orchestration of storage, cloud management, and monitoring, ensuring high performance and availability by cross-pairing network interface cards and automating the setup of virtual machines and resources.
Minimizes human and time waste by automating the provisioning process, ensuring high performance and availability of HCI on bare metal servers, and allows for seamless expansion or replacement of components without reworking the entire setup.
Smart Images

Figure KR2023006325_14112024_PF_FP_ABST
Abstract
Description
Systems and methods for provisioning hyper-converged infrastructure
[0001] The present invention relates to hyper-converged infrastructure (HCI), and more particularly, to a system and method for provisioning Ceph-based hyper-converged infrastructure (HCI) on a bare metal server.
[0002] Hyper-Converged Infrastructure (HCI) is a system that combines computing, storage, and networking elements. HCI clusters standard servers to provide virtual server functionality, external storage, and networking capabilities. The most crucial system for provisioning HCI is a distributed storage system that combines the disks of each server to function as a single storage unit.
[0003] Ceph is a distributed storage system that clusters multiple bare metal servers, allowing their disks to be combined and used as a single storage unit. It offers storage capabilities suitable for provisioning HCI. However, Ceph only provides storage using bare metal servers and does not provide the compute and networking capabilities required by HCI.
[0004] Therefore, when provisioning HCI using Ceph, the operating system, virtualization software, Ceph software, and cloud management software were previously manually provisioned on bare metal servers. This provisioning method is extremely labor-intensive and time-consuming, as the entire process must be reworked manually if some bare metal servers need to be replaced or expanded.
[0005] This calls for a more robust provisioning method that is independent of any individual bare metal system and can address issues that can arise from manual provisioning, for provisioning HCI on bare metal servers via Ceph.
[0006] Accordingly, the present invention is an invention created to solve the above-described problems, and the main purpose of the present invention is to provide a system and method that can conveniently and automatically provision an operating system, virtualization software, Ceph software, and cloud management software when provisioning HCI to a bare metal server using Ceph, an open source distributed storage software.
[0007] Furthermore, the present invention proposes an architecture and process of HCI that is automatically configured through orchestration when configuring HCI on a bare metal server using Ceph, and an additional purpose is to enable the automation of the creation and configuration of various resources required for HCI construction through an orchestration device created using the architecture and process.
[0008] In addition, another object of the present invention is to provide a system and method that can ensure high performance and high availability when provisioning HCI on a bare metal server using Ceph.
[0009] A system for provisioning HCI on a bare metal server according to an embodiment of the present invention to achieve the aforementioned purpose comprises at least two high-speed network switches and three or more bare metal servers,
[0010] Each of the high-speed network switches has a NIC connection port for connecting to different network interface cards (NICs) existing in each of the bare metal servers and an external network connection port for connecting to an external network, and one of the high-speed network switches further has an IPMI connection port for connecting to an IPMI port of each of the bare metal servers, so that the NICs provided in each of the high-speed network switches and the bare metal servers are cross-paired, and each of the bare metal servers has:
[0011] It features a Ceph-based operating system installed and running, including components for provisioning HCI.
[0012] Furthermore, in a system including the above-described configuration, each of the bare metal servers,
[0013] IPMI traffic is processed through the IPMI port connected to the above IMPI connection port and the cable,
[0014] Management traffic is processed through the management traffic port on the NIC side, which is connected by a cable to the management traffic port among the ports for the above NIC connection.
[0015] Among the ports for connecting the above NIC, the virtual machine traffic is processed through the port for virtual machine traffic on the NIC side that is connected by a cable.
[0016] Storage service traffic is processed through the storage service traffic port on the NIC side, which is connected by a cable to the storage service traffic port among the ports for the above NIC connection,
[0017] It is characterized in that data replication traffic is processed through a port for data replication traffic on the NIC side that is connected by a cable to a port for data replication traffic among the ports for the above NIC connection.
[0018] In addition, the Ceph-based operating system is characterized by including, as components for provisioning HCI, a controller virtual machine image, a component for preparing cluster configuration, a component for orchestrating a storage controller virtual machine, a component for orchestrating a cloud management virtual machine, and a component for orchestrating a monitoring center.
[0019] In addition, the Ceph-based operating system is characterized in that it is downloaded and installed on each bare metal server from a user computer that can be connected through an external network connection port of the high-speed network switch.
[0020] Furthermore, it is characterized by further including a recording medium on which Ceph-based operating system program data including components for provisioning the HCI is recorded.
[0021] Meanwhile, a method for provisioning an HCI according to another embodiment of the present invention is a method for provisioning an HCI to each of bare metal servers having a plurality of high-speed network switches and a plurality of network interface cards (NICs) cross-paired with the plurality of high-speed network switches,
[0022] Steps of connecting a user computer to a bare metal server via an IPMI network and mounting a Ceph-based operating system image existing on the user computer to the connected bare metal server;
[0023] A step of booting a connected bare metal server to install a Ceph-based operating system, selecting and setting a management network NIC and IP address to install the operating system;
[0024] A step of executing the cluster configuration preparation part included in the operating system installed on each of the above bare metal servers to prepare preliminary information for provisioning the HCI;
[0025] It is characterized by including a step of sequentially executing the virtual machine orchestration units so as to provision a storage controller virtual machine, a cloud management virtual machine, and a monitoring center on a bare metal server by linking the controller virtual machine image included in the above operating system and the plurality of virtual machine orchestration units.
[0026] In the method including the above-described configurations, in the step of executing the above-described pre-information preparation,
[0027] It is characterized in that a key file and a host file for encrypted shell execution are created by the execution of the above cluster configuration preparation part, and time server information is set to synchronize the time of the bare metal server and controller virtual machine that constitute the HCI.
[0028] In the step of executing the storage controller virtual machine orchestration unit, which is one of the above virtual machine orchestration units,
[0029] It is characterized by setting the size of compute resources among the devices to be configured in the virtual machine, selecting the configuration method and configuration target device of the disk to be used for Ceph storage configuration among the virtual machine devices, selecting the NIC for management traffic, the NIC for the server to be used as the storage traffic NIC, and the NIC for replication among the virtual machine devices, and then registering the host information, host name, IP address of the management NIC, IP address of the storage server NIC, and IP address of the storage replication NIC, and registering the encrypted shell execution key information to be used for connection to the storage controller virtual machine and command line execution, and then provisioning the storage controller virtual machine and orchestrating the registered settings information.
[0030] The above storage controller virtual machine orchestration unit is characterized in that it generates an initialization script using the registered setting information and transmits it to the virtual machine so that the IP address, encrypted shell execution key, and host information of the virtual machine are set.
[0031] Furthermore, in a method for provisioning the above-described HCI on a bare metal server, a Ceph orchestration unit configuring the provisioned storage controller virtual machine registers the storage controller virtual machine as a storage cluster node, the Ceph orchestration unit executes a Ceph storage server within the provisioned storage controller virtual machine, the Ceph storage server identifies an inventory of a storage disk of the provisioned storage controller virtual machine, selects a device for an object storage daemon by selecting from the identified disk inventory, and executes an object storage daemon in conjunction with the selected object storage daemon device and a Ceph disk controller to create a Ceph storage pool.
[0032] In the step of executing the cloud management virtual machine orchestration unit, which is one of the above virtual machine orchestration units,
[0033] It is characterized by setting up failover cluster information for providing high availability of a cloud management virtual machine and compute resources of a cloud management virtual machine, selecting a network configuration and a network NIC to be allocated for the cloud management virtual machine, setting up host information, a host name, a management NIC IP address, and a service NIC IP address, and registering key information for executing the encrypted shell, and then provisioning and orchestrating the cloud management virtual machine.
[0034] In addition, provisioning of the cloud management virtual machine uses the registered compute resources and NIC resource information, and orchestration of the cloud management virtual machine is characterized by executing a cloud management database service included in the controller virtual machine image and starting a cloud management service in conjunction with a cloud management platform.
[0035] In the step of running the monitoring center orchestration unit, which is one of the above virtual machine orchestration units,
[0036] It features provisioning and orchestrating the monitoring center by entering the IP addresses of the target hosts and virtual machines to be monitored and the SMTP information that the monitoring center will use when sending alarms, using the input information.
[0037] In this case, the monitoring center is characterized by running as a service within a cloud management virtual machine provisioned by the cloud management virtual machine orchestration unit.
[0038] According to the technical problem solving means described above, the present invention has the advantage of ensuring high performance and high availability when provisioning HCI to a bare metal server using Ceph by adopting a structure in which a plurality of network interface cards provided in each of high-speed network switches and a bare metal server are cross-paired with each other.
[0039] Furthermore, the present invention connects a user computer to a single bare metal server via an IPMI network to install a Ceph-based operating system, and sequentially provisions a storage controller virtual machine, a cloud management virtual machine, and a monitoring center on the bare metal server by linking a controller virtual machine image and a plurality of virtual machine orchestration units included in the operating system, thereby minimizing waste of human and time compared to the conventional method of manually reworking all processes when a part of the bare metal server needs to be replaced or a part needs to be additionally expanded.
[0040] FIG. 1 is an example diagram of the connection status of bare metal servers (230) and high-speed network switch (210, 220) devices that constitute a system for provisioning hyper-converged infrastructure (HCI) according to an embodiment of the present invention.
[0041] Figure 2 is an example of a network cable connection status between a high-speed network switch (210, 220) device and a bare metal server (230) in Figure 1.
[0042] FIG. 3 is an example diagram of the connection state between a user computer (310) and a bare metal server (230) required to install an operating system on a bare metal server (230) according to an embodiment of the present invention.
[0043] FIG. 4 is a drawing illustrating a procedure for installing an operating system on a bare metal server (230) from a user computer (310) illustrated in FIG. 3.
[0044] Figure 5 is an example of components included in the operating system installed on the bare metal server (230) by Figure 4.
[0045] FIG. 6 is a diagram for explaining a cluster configuration pre-preparation method for provisioning Ceph-based HCI according to an embodiment of the present invention.
[0046] FIG. 7 is an example diagram of the configuration of a storage controller virtual machine provisioned by the storage controller virtual machine orchestration unit (503) illustrated in FIG. 5.
[0047] FIG. 8 is a diagram illustrating an embodiment of provisioning a storage controller virtual machine through the storage controller virtual machine orchestration unit (503) illustrated in FIG. 7.
[0048] FIG. 9 is an exemplary diagram illustrating a procedure for provisioning a storage cluster to provision HCI according to an embodiment of the present invention.
[0049] FIG. 10 is an exemplary diagram illustrating a procedure for provisioning a cloud management virtual machine to provision HCI according to an embodiment of the present invention.
[0050] FIG. 11 is an exemplary diagram illustrating a procedure for provisioning a monitoring center to provision HCI according to an embodiment of the present invention.
[0051] The detailed description of the present invention, which follows, refers to the accompanying drawings, which illustrate specific embodiments in which the present invention may be implemented, to clearly illustrate the purposes, technical solutions, and advantages of the present invention. These embodiments are described in sufficient detail to enable those skilled in the art to practice the present invention.
[0052] Furthermore, throughout the detailed description and claims of the present invention, the word "comprise" and variations thereof are not intended to exclude other technical features, additives, components, or steps. Other objects, advantages, and features of the present invention will become apparent to those skilled in the art, in part, from this description, and in part, from practice of the invention. The examples and drawings below are provided by way of illustration and are not intended to limit the invention. Moreover, the present invention encompasses all possible combinations of the embodiments set forth herein. It should be understood that the various embodiments of the present invention, while different from one another, are not necessarily mutually exclusive. It should also be understood that the positions or arrangements of individual components within each disclosed embodiment may be varied without departing from the spirit and scope of the present invention. Therefore, the following detailed description is not intended to be limiting, and the scope of the present invention, if properly described, is defined only by the appended claims, along with the full scope of equivalents to which such claims are entitled.
[0053] Meanwhile, unless otherwise indicated or clearly contradicted by context, items referred to in the singular encompass plural references unless the context otherwise requires. Furthermore, when describing the present invention, detailed descriptions of related known structures or functions will be omitted if they are deemed to obscure the gist of the present invention.
[0054] First, FIG. 1 illustrates the connection status of bare metal servers (230) and high-speed network switch (210, 220) devices that constitute a system for provisioning hyper-converged infrastructure (HCI) according to an embodiment of the present invention, and FIG. 2 illustrates the network cable connection status between the high-speed network switch (210, 220) device and the bare metal server (230) in FIG. 1.
[0055] Referring to FIG. 1, a Ceph-based HCI (100) according to an embodiment of the present invention is composed of two high-speed network switches (210, 220) and three or more bare metal servers (230).
[0056] The physical devices for provisioning HCI must be connected via network cables between the switch and the server, and between the external network and the HCI.
[0057] Accordingly, as illustrated in FIG. 2, the high-speed network switches (210, 220) each include a NIC connection port (211, 221) for connecting to a network interface card (NIC, 232, 233) of a bare metal server (230), an IPMI connection port (212) for connecting to an IPMI port (231) of the bare metal server (230), and an external network connection port (213, 222) for connecting to an external network.
[0058] The bare metal server (230) includes two network interface cards (NICs) (232, 233) that guarantee a transmission speed of more than 10 Gb per second, and an IPMI port (231).
[0059] In order to ensure that the network is available normally even if one high-speed network switch fails or one network interface card in the bare metal server (230) fails, the high-speed network switches (210, 220) and NICs (232, 233) are cross-paired as shown in Fig. 2.
[0060] Cable connections must be designed to handle the network traffic generated by the HCI. This includes IPMI traffic, management traffic, virtual machine traffic, storage service traffic, and data replication traffic.
[0061] Accordingly, IPMI port 231 provided on the bare metal server (230) and IPMI connection port 212 of the active switch (210) among the high-speed network switches are connected by a cable so that IPMI traffic is processed.
[0062] Meanwhile, among the NIC connection ports (211, 221) of each high-speed network switch (210, 220), the ports for management traffic (211A, 221A) and the ports for management traffic (232A, 233A) of the bare metal server (230) are connected by cables so that management traffic is processed, and among the NIC connection ports (211, 221) of the high-speed network switch (210, 220), the ports for virtual machine traffic (211B, 221B) and the ports for virtual machine traffic (232B, 233B) of the bare metal server (230) are connected by cables so that virtual machine traffic is processed.
[0063] Meanwhile, among the NIC connection ports (211, 221) of the high-speed network switch (210, 220), the storage service traffic port (211C, 221C) and the storage traffic port (232C, 233C) of the bare metal server (230) are connected by cables so that the storage service traffic is processed among the storage traffic, and among the NIC connection ports (211, 221) of the high-speed network switch (210, 220), the data replication traffic port (211D, 221D) and the data replication traffic port (232D, 233D) of the bare metal server (230) are connected by cables so that the data replication traffic is processed.
[0064] In an embodiment of the present invention, an operating system is installed centrally on a bare metal server (230). This is more effective in maintaining consistency than a method of installing an operating system on a bare metal server (230) using operating system installation media (which can be considered operating system recording media) such as a DVD, and has the advantage of allowing the same operating system to be installed on all servers as long as a user (administrator) computer is present.
[0065] FIG. 3 illustrates the connection state between a user computer (310) and a bare metal server (230) required to install an operating system on a bare metal server (230) according to an embodiment of the present invention.
[0066] Referring to FIG. 3, it is assumed that an operating system (installation) image (311) file including software for HCI exists (on a separate storage medium or on an internal storage medium) on the user computer (310). The user computer (310) is connected to a network connection port (312) of the user computer (310) and an external network connection port (213) of a high-speed network switch (210) for IPMI connection with a bare metal server (230) via a cable.
[0067] Since the bare metal server (230) is connected to the IPMI connection port (212) of the high-speed network switch (210) via the IPMI port (231), connection is consequently possible between the user computer (310) and the bare metal server (230). When the bare metal server (230) and the user computer (310) are connected via the network in this way, an operating system can be installed on the bare metal server (230).
[0068] FIG. 4 is a drawing illustrating a procedure for installing an operating system on a bare metal server (230) from the user computer (310) illustrated in FIG. 3.
[0069] Referring to FIG. 4, first, a user connects to the IPMI management tool of a bare metal server (230) via a user computer (310) and an IPMI network (step 401). Next, the user connects the operating system image (311) of the user computer (310) to a virtual USB device of the bare metal server (230) via the IPMI management tool (step 402) and mounts it on the bare metal server (230).
[0070] Next, the user boots the bare metal server (230) using the operating system image as the boot image (step 403). When the bare metal server (230) is booted, a program guiding the steps for installing the operating system is executed, and a display is provided asking the user to select a disk on which to install the operating system (step 404).
[0071] Accordingly, the user selects a disk on which to install the operating system among the disks on the bare metal server (230) and proceeds with the operating system installation. Since the configuration of a management network is required to provision the bare metal server (230) as an HCI, the user selects a management network NIC during the operating system installation step and sets an Internet Protocol address (step 405).
[0072] Meanwhile, the operating system installation program automatically installs the Linux kernel, HCI software, and virtual appliance image on the selected disk of the bare metal server (230) (step 406). Upon completion of the installation, the operating system installation program reboots the bare metal server (230) (step 407). Once the operating system installation on the bare metal server (230) is complete, all components for provisioning the HCI are installed.
[0073] Figure 5 illustrates components included in the operating system installed on the bare metal server (230) by Figure 4.
[0074] The bare metal server (230) on which the operating system is installed includes a controller virtual machine image (501), as illustrated in FIG. 5, and further includes a cluster configuration preparation unit (502), a storage controller virtual machine orchestration unit (503), a cloud management virtual machine orchestration unit (504), and a monitoring center orchestration unit (505).
[0075] Each orchestration unit (503, 504, 505) interacts with the controller virtual machine image (501) to provision the necessary virtual machine on the bare metal server (230).
[0076] The virtual machines provisioned by the orchestration unit (503, 504, 505) are allocated the CPU (506A), memory (506B), and storage (506C) of the bare metal server (230) in a virtualized manner or through a pass-through manner. Once all the bare metal servers (230) are ready, preliminary information for provisioning the HCI is prepared.
[0077] FIG. 6 is a diagram illustrating a cluster configuration pre-preparation method for provisioning Ceph-based HCI according to an embodiment of the present invention.
[0078] Referring to FIG. 6, the user first executes (step 601) the cluster configuration preparation unit (502) included in the operating system of the bare metal server (230). The cluster configuration preparation unit (502) provides a function for inputting or selecting configuration information for configuring a cluster using software.
[0079] The first task that the cluster configuration preparation unit (502) must perform is to create a key file for executing an encrypted shell (step 602). The key file is used jointly by the bare metal server (230) and the controller virtual machine (501) that constitute the HCI, and authentication is performed between the server and the virtual machine through the key file without exchanging separate IDs and passwords, and the encrypted shell is transmitted and executed between the server and the virtual machine.
[0080] Once the key file is created, the cluster configuration preparation unit (502) creates a hosts file (step 603). The hosts file is a file that manages pairs of host names and Internet Protocol addresses of Linux servers, and has pairs of names and Internet Protocol addresses for the bare metal server (230) and controller virtual machine (501) that constitute the HCI. Using the previously created key file and hosts file, fast command line transmission between each server becomes possible.
[0081] Next, the user sets time server information to synchronize the system times of the bare metal server (230) and controller virtual machine (501) that constitute HCI (step 604).
[0082] Finally, the cluster configuration preparation task is executed using the input and selected information (step 605).
[0083] When the cluster configuration preparation is complete, the controller virtual machine image (501) installed on the bare metal server (230) and each of the orchestration units (503, 504, 505) are linked to provision the storage controller virtual machine, the cloud management virtual machine, and the monitoring center on the bare metal servers (230), and the cluster configuration preparation unit (502) is terminated.
[0084] FIG. 7 illustrates the configuration of a storage controller virtual machine provisioned by the storage controller virtual machine orchestration unit (503) illustrated in FIG. 5.
[0085] The storage controller virtual machine image contains all components of Ceph distributed storage, the cloud management platform, and the monitoring platform. The storage controller virtual machine image links with the orchestration unit installed on the bare metal server (230) to provision and configure the necessary services and perform the assigned role.
[0086] To provide these functions, the storage controller virtual machine image is composed of a Ceph orchestration unit (701), a Ceph storage server (702), a Ceph disk controller (703), a Ceph management unit (704), a cloud management database (705), a cloud management platform (706), a monitoring unit (707), and a notification unit (708).
[0087] Meanwhile, the storage controller virtual machine orchestration unit (503) illustrated in FIG. 5 provisions a storage controller virtual machine in conjunction with the controller virtual machine image (501).
[0088] FIG. 8 is a diagram illustrating an example of provisioning a controller virtual machine image as a storage controller virtual machine through a storage controller virtual machine orchestration unit (503) installed on a bare metal server (230) according to an embodiment of the present invention.
[0089] Referring to Fig. 8, first, the user executes the storage controller virtual machine orchestration unit (503) (step 800). Next, the user sets the size of the compute resources among the devices to be configured in the virtual machine (step 802). The devices to be configured in the virtual machine are the sizes of the CPU and memory. The CPU and memory are allocated by virtualizing the CPU (506A) and memory (507B) of the bare metal server (230). The CPU and memory of the storage controller virtual machine are appropriately selected according to the number of disks to be used to configure the Ceph storage. If there are 6 or fewer disks, select 8 cores and 16 GB of resources. If there are more than 6 disks, select 16 cores and 32 GB of resources.
[0090] Next, the user selects the configuration method and the target device for configuring the disk to be used for configuring Ceph storage among the virtual machines (step 803). The disk configuration method is selected between the RAID pass-through method and the LUN pass-through method. The RAID pass-through method assigns the entire disk connected to the RAID device to the storage controller virtual machine, and the LUN pass-through method assigns the disk LUN unit to the storage controller virtual machine. If the LUN pass-through method is selected among the above methods, the disk of the bare metal server (230) to be assigned to the storage controller virtual machine is selected.
[0091] Next, a NIC for management traffic is selected from among the virtual machines, and a NIC for server use and a NIC for replication are selected as the NIC for storage traffic (step 804). The NIC for management is allocated by virtualizing the bridge of the NIC for management traffic (231) of the bare metal server (230). The NIC for storage traffic is allocated by bonding the NIC for server use (232C, 233C) and the NIC for replication (232D, 233D) among the storage traffic NICs of the bare metal server (230) and virtualizing the bridge of the corresponding bonding device.
[0092] Next, the user registers the host information, host name, Internet Protocol address of the management NIC, Internet Protocol address of the storage server NIC, and Internet Protocol address of the storage replication NIC (step 805). The host information and host name information are set using the host file created in Fig. 6, which prepares the cluster configuration that constitutes the HCI above.
[0093] Afterwards, the encrypted shell execution key information to be used for connection and command line execution of the storage controller virtual machine is registered (step 806). The encrypted shell execution key information uses the encrypted shell execution key file created in Fig. 6, which prepares the cluster configuration in advance.
[0094] After registering all information, the storage controller virtual machine is provisioned and the set information is orchestrated (step 807). The storage controller virtual machine replicates the controller virtual machine image and then runs the virtual machine by applying the registered CPU, memory, disk, and NIC device information. When the virtual machine is run, the storage controller virtual machine orchestration unit (503) uses the registered information in this embodiment to create an initialization script and transfers the initialization script to the virtual machine. The virtual machine's Internet Protocol address, encrypted shell execution key, and host information are all set by the initialization script. When the storage controller virtual machine is run and orchestration is completed, the storage controller virtual machine orchestration unit (503) is terminated (step 808).
[0095] Once the storage controller virtual machine is provisioned and running, all storage controller virtual machines running on the bare metal server (230) are linked to form a storage cluster. The storage cluster is a process of integrating the disks of the bare metal server (230) into a single storage and providing service.
[0096] FIG. 9 illustrates a procedure for provisioning a storage cluster to provision Ceph-based HCI according to an embodiment of the present invention.
[0097] Referring to FIG. 9, the user first executes the Ceph orchestration unit (701) of the storage controller virtual machine (step 901). The Ceph orchestration unit (701) is a device that manages various services for configuring a Ceph storage pool and manages nodes that constitute a cluster, i.e., each storage controller virtual machine.
[0098] The Ceph orchestration unit (701) executes a Ceph storage server (702) monitor for each storage controller virtual machine (step 902). The storage controller virtual machine is registered as a storage cluster node through the Ceph orchestration unit (701) (step 903). The Ceph orchestration unit (701) executes a Ceph storage server (702) within the storage controller virtual machine. The Ceph storage server (702) identifies the inventory of the storage disk of the storage controller virtual machine (step 904). Then, the device for the object storage daemon is selected from the identified disk inventory (step 905), and the object storage daemon is executed through linkage between the selected object storage daemon device and the Ceph disk controller (703) (step 906), and a basic Ceph storage pool is created through linkage between the storage server of the storage controller virtual machine and the object storage daemon (step 907).
[0099] After creating a Ceph storage pool through a storage controller virtual machine, a cloud management virtual machine is provisioned through linkage with the Ceph storage pool, controller virtual machine image (501), and cloud management virtual machine orchestration unit (504).
[0100] FIG. 10 illustrates a procedure for provisioning a cloud management virtual machine to provision a Ceph-based HCI according to an embodiment of the present invention.
[0101] Referring to FIG. 10, first, the cloud management virtual machine orchestration unit (504) is executed (step 1001). The user sets up failover cluster information for providing high availability of the cloud management virtual machine (step 1002). The failover cluster executes the cloud management virtual machine on one of all bare metal servers (230) that constitute the HCI, and provides a high availability function for executing the cloud management virtual machine on another server in the event of a failure.
[0102] Next, the user sets the compute resources of the cloud management virtual machine (step 1003).
[0103] Next, configure the network of the cloud management virtual machine and select the network NIC to be assigned (step 1004). Then, set the host information, host name, management NIC Internet Protocol address, and service NIC Internet Protocol address information (step 1005).
[0104] Next, register the key information for executing the encrypted shell (step 1006). After registering all information, provision and orchestrate the cloud management virtual machine (step 1007).
[0105] Provisioning of virtual machines utilizes registered compute resources and NIC resource information. Virtual machine orchestration executes a cloud management database (705) service and initiates a cloud management service in conjunction with a cloud management platform (706). Finally, the cloud management virtual machine orchestration unit (504) is terminated.
[0106] After provisioning a cloud management virtual machine and starting the cloud management service, provision the monitoring center. The monitoring center runs as a service within the cloud management virtual machine.
[0107] FIG. 11 illustrates a procedure diagram for provisioning a monitoring center within a cloud management virtual machine to provision a Ceph-based HCI according to an embodiment of the present invention.
[0108] Referring to Fig. 11, the monitoring center orchestration unit (505) is first executed (1101). The user inputs the Internet Protocol addresses of the target hosts and virtual machines to be monitored through the orchestration unit (505) (step 1102). The host is a bare metal server (230), and the virtual machines include both a storage controller virtual machine and a cloud management virtual machine.
[0109] Next, enter the SMTP information that the monitoring center will use to send alarms (step 1103).
[0110] Next, the monitoring center is provisioned and orchestrated (step 1104) using the entered information. The monitoring center provisioning is to execute the monitoring unit (707) of the controller virtual machine image (501), and the orchestration is to collect and visualize monitoring information in conjunction with the bare metal server (230) and the controller virtual machine, and to set up and execute the alarm unit (708) using SMTP information. Finally, the monitoring center orchestration unit (505) is terminated.
[0111] As described above, the system and method for provisioning hyper-converged infrastructure (HCI) according to an embodiment of the present invention to a bare metal server,
[0112] By connecting a user computer to a bare metal server via an IPMI network, mounting a Ceph-based operating system image existing on the user computer to each connected bare metal server to install the operating system, and sequentially provisioning a storage controller virtual machine, a cloud management virtual machine, and a monitoring center on the bare metal server through linking the controller virtual machine image included in the operating system and multiple virtual machine orchestration units, there is an advantage in that human and time waste can be minimized compared to the previous method where all processes had to be manually reworked when some of the bare metal servers needed to be replaced or some needed to be additionally expanded.
[0113] In addition, the present invention has the advantage of ensuring high performance and high availability when provisioning HCI to a bare metal server using Ceph by adopting a structure in which network interface cards provided in each of high-speed network switches and bare metal servers are cross-paired with each other.
[0114] While the present invention has been described with reference to the embodiments illustrated in the drawings, these are merely exemplary, and those skilled in the art will appreciate that various modifications and equivalent alternative embodiments are possible. Therefore, the true technical protection scope of the present invention should be defined solely by the appended claims.
[0115] The present invention relates to hyper-converged infrastructure (HCI), and is an invention that can be used in the storage industry within a data center.
Claims
1. Including at least two high-speed network switches and three or more bare metal servers, Each of the high-speed network switches has a NIC connection port for connecting to different network interface cards (NICs) existing in each of the bare metal servers and an external network connection port for connecting to an external network, and one of the high-speed network switches further has an IPMI connection port for connecting to an IPMI port of each of the bare metal servers, so that the NICs provided in each of the high-speed network switches and the bare metal servers are cross-paired, and each of the bare metal servers has: A system for provisioning HCI on bare metal servers, characterized by having a Ceph-based operating system installed and running, including components for provisioning HCI.
2. In claim 1, each of the bare metal servers, IPMI traffic is processed through the IPMI port connected to the above IMPI connection port and the cable, Management traffic is processed through the management traffic port on the NIC side, which is connected by a cable to the management traffic port among the ports for the above NIC connection. Among the ports for connecting the above NIC, the virtual machine traffic is processed through the port for virtual machine traffic on the NIC side that is connected by a cable. Storage service traffic is processed through the storage service traffic port on the NIC side, which is connected by a cable to the storage service traffic port among the ports for the above NIC connection, A system for provisioning HCI on a bare metal server, characterized in that data replication traffic is processed through a port for data replication traffic on the NIC side that is connected by a cable to a port for data replication traffic among the ports for the above NIC connection.
3. A system for provisioning HCI on a bare metal server, characterized in that, in claim 1, the Ceph-based operating system comprises a controller virtual machine image, a component for preparing a cluster configuration, a component for orchestrating a storage controller virtual machine, a component for orchestrating a cloud management virtual machine, and a component for orchestrating a monitoring center as components for provisioning HCI.
4. In claim 1 or claim 3, the Ceph-based operating system, A system for provisioning HCI to a bare metal server, characterized in that it is downloaded and installed to each bare metal server from a user computer that can be connected through an external network connection port of the high-speed network switch.
5. A system for provisioning an HCI to a bare metal server, characterized in that it further includes a recording medium on which Ceph-based operating system program data including components for provisioning the HCI according to claim 1 or claim 3 is recorded.
6. A method for provisioning hyper-converged infrastructure (HCI) to each bare metal server having a plurality of high-speed network switches and a plurality of network interface cards (NICs) cross-paired with the plurality of high-speed network switches, Steps of connecting a user computer to a bare metal server via an IPMI network and mounting a Ceph-based operating system image existing on the user computer to the connected bare metal server; A step of booting a connected bare metal server to install a Ceph-based operating system, selecting and setting a management network NIC and IP address to install the operating system; A step of executing the cluster configuration preparation part included in the operating system installed on each of the above bare metal servers to prepare preliminary information for provisioning the HCI; A method for provisioning an HCI on a bare metal server, characterized in that it includes a step of sequentially executing the virtual machine orchestration units so as to provision a storage controller virtual machine, a cloud management virtual machine, and a monitoring center on the bare metal server by linking the controller virtual machine image included in the above operating system and the plurality of virtual machine orchestration units.
7. In claim 6, in the step of executing the preparation of the prior information, A method for provisioning an HCI on a bare metal server, characterized in that a key file and a host file for encrypted shell execution are generated by execution of the above cluster configuration preparation part, and time server information for synchronizing the time of a bare metal server and a controller virtual machine configuring the HCI is set.
8. In claim 6, in the step of executing the storage controller virtual machine orchestration unit, which is one of the virtual machine orchestration units, A method for provisioning an HCI on a bare metal server, characterized in that the size of compute resources among the devices to be configured in the virtual machine is set, the configuration method and configuration target device of the disk to be used for Ceph storage configuration among the virtual machine devices are selected, the NIC for management traffic, the NIC for the server to be used as the storage traffic NIC, and the NIC for replication are selected among the virtual machine devices, and then the host information, host name, IP address of the management NIC, IP address of the storage server NIC, and IP address information of the storage replication NIC are registered, and the encrypted shell execution key information to be used for connection to the storage controller virtual machine and command line execution is registered, and then the storage controller virtual machine is provisioned and the registered settings are orchestrated.
9. In claim 8, the storage controller virtual machine orchestration unit, A method for provisioning an HCI on a bare metal server, characterized in that an initialization script is generated using the registered information above and transmitted to the virtual machine so that the IP address, encrypted shell execution key, and host information of the virtual machine are set.
10. A method for provisioning an HCI on a bare metal server according to claim 8, wherein a Ceph orchestration unit configuring the provisioned storage controller virtual machine registers the storage controller virtual machine as a storage cluster node, the Ceph orchestration unit executes a Ceph storage server within the provisioned storage controller virtual machine, the Ceph storage server identifies an inventory of storage disks of the provisioned storage controller virtual machine, selects a device for an object storage daemon from among the identified disk inventories, and executes an object storage daemon in conjunction with the selected object storage daemon device and a Ceph disk controller to create a Ceph storage pool.
11. In claim 8 or claim 10, in the step of executing the cloud management virtual machine orchestration unit, which is one of the virtual machine orchestration units, A method for provisioning an HCI on a bare metal server, characterized in that the method comprises: setting failover cluster information for providing high availability of a cloud management virtual machine and compute resources of the cloud management virtual machine, selecting a network configuration and a network NIC to be allocated for the cloud management virtual machine, setting host information, a host name, a management NIC IP address, and a service NIC IP address, and registering key information for executing the encrypted shell, and then provisioning and orchestrating the cloud management virtual machine.
12. A method for provisioning an HCI on a bare metal server, characterized in that, in claim 11, provisioning of the cloud management virtual machine uses the registered compute resources and NIC resource information, and orchestration of the cloud management virtual machine executes a cloud management database service included in the controller virtual machine image and starts a cloud management service in conjunction with a cloud management platform.
13. In claim 11, in the step of executing the monitoring center orchestration unit, which is one of the virtual machine orchestration units, A method for provisioning HCI on a bare metal server, characterized by entering the IP addresses of the target hosts and virtual machines to be monitored and the SMTP information to be used by the monitoring center when sending alarms, and then provisioning and orchestrating the monitoring center using the entered information.
14. A method for provisioning an HCI on a bare metal server, characterized in that the monitoring center according to claim 13 is executed in the form of a service within a cloud management virtual machine provisioned by a cloud management virtual machine orchestration unit.
Citation Information
Patent Citations
Provisioning the hyper-converged infrastructure bare metal systems from the top of the rack switch
KR1020180000280A
Speech coding method and apparatus for performing the same
KR1020240060425A
Storage management in a data management platform for cloud-native workloads
US20200136943A1
Discovery of hyper-converged infrastructure devices
US20200204449A1
Systems, computing devices, and methods for authenticating privileged subsystem access by policy and by use of a security key generated at boot
US20200285750A1
