What are the limitations of physical servers?
l Users cannot directly load external hardware devices, such as USB flash drives, U keys of banking services, external hard disks, and dongles.
l Out-of-band management is not supported. Your physical server resources are managed and maintained by eSurfing Cloud.
l Hot migration is not supported. The failure of a physical server may adversely affect services. Therefore, we recommend that you deploy services in cluster or active/standby mode to achieve high service availability.
l You cannot create a raw device without OS. That is, a physical server must be pre-installed with an OS.
l You cannot customize configurations of a physical server such as CPU and memory nor expand the capacity of CPU, memory or local disk. You can only expand the capacity of an EVS.
l You cannot mount an EVS to some physical server specifications or images because they are not equipped with smart NICs or for some other reasons.
l You cannot configure a bridged NIC for physical servers. Otherwise, network exceptions may occur.
l Do not upgrade the built-in kernel version of the OS. Otherwise, the hardware drive of the physical server may not be completely compatible and thus the reliability may be affected.
What is the main difference between physical servers and elastic cloud servers?
The main difference between physical servers and elastic cloud servers is how resources are shared. An elastic cloud server is a virtual machine (VM) where multiple tenants share the same physical resource while the physical server resource is exclusively available to users.
Physical servers are recommended for critical applications or services with high performance requirements, such as big data clusters or enterprise middleware systems. This is because physical servers provide a secure and reliable operating environment, and users can exclusively use physical resources to obtain better performance and stability.
What is the relationship between physical servers and elastic bare metal?
A physical server exists in two forms: standard bare metal and elastic bare metal.
Standard bare metal features excellent computing and storage performance to meet core applications for high performance and stability. In addition, it can be networked with elastic cloud servers to provide users with flexible service deployment solutions.
Elastic bare metal combines the resource isolation of a physical server with the flexibility of a cloud server. By unloading network and storage functions onto a DPU card, a universal physical server can be upgraded to a high-performance elastic computing product featuring hardware acceleration, EVS mounting and startup, and support for VPC network.
What are the differences between standard bare metal and elastic bare metal?
Elastic bare metal possesses the elastic and flexible attributes of cloud servers and features hardware acceleration, EVS mounting, security group isolation, and multiple elastic NICs. Standard bare metal does not have the functions of EVS mounting, security group isolation, and multiple elastic NICs.
What are the differences between physical servers and traditional physical servers?
The physical servers, on the basis of retaining the features and advantages of traditional physical servers, provide a more flexible, efficient, and scalable deployment and use mode by automatically managing the cloud platform and integrating other cloud resources.
Automated management: Physical servers provide automatic provisioning and O&M capabilities of cloud platforms. Compared with traditional physical servers, a physical server provides rapid provisioning, configuration, and management through cloud platform without manual hardware deployment and maintenance.
Flexibility: A physical server can be flexibly provisioned and used like VMs. You can quickly adjust the configuration and scale of a physical server to flexibly allocate resources to meet business needs.
Direct access to hardware resources: A physical server allows your application to directly access the processor and memory of the physical server without going through the virtualization layer, thereby reducing the performance overhead of virtualization. This is important for applications with high performance requirements.
Virtual private cloud (VPC) interconnection: A physical server can connect to other resources in the cloud platform over a virtual private cloud. This means that you can seamlessly connect and communicate physical servers with other virtual machines, cloud storage, and other resources in a cloud environment.
Support for file storage: A physical server can connect to a scalable file service to share and access data. Compared with traditional physical servers, a physical server can be more easily integrated with a file storage system to provide more efficient data sharing and storage.
How does a physical server ensure data security?
A physical server features excellent performance, resource isolation, support for local disk, and volume backup service to ensure data security and reliability. A physical server provides high-level data protection with resource isolation, redundant data storage, and backup and restoration to meet users' requirements for data security.
Performance and resource isolation: A physical server is a computing resource exclusively available to users and is not shared with other tenants. This ensures data processing performance and resource isolation. This prevents other tenants from accessing or tampering with your data.
No virtualization overhead: A physical server has no virtualization layer, and your applications can directly access the processor and memory of the physical server without the performance overhead of virtualization. This can guarantee higher performance and lower latency while reducing the security risks associated with virtualization.
Support for local disk: A physical server with local disk supports RAID constituting local disks. The RAID technology can store data redundantly among multiple disks, improving the fault tolerance and reliability of data. Even if one of the disks fails, the data can still be recovered or accessed, thereby guaranteeing the security of the data.
EVS backup and restoration: If a physical server does not have a local disk, the EVS can be used as a system disk and the volume backup service is used to protect data. The volume backup service supports the consistency snapshot technology based on multiple EVS disks to periodically back up data on the physical server and thus ensure data security and accuracy. When data is lost or corrupted, data can be restored from backup data, ensuring service continuity and data security to the maximum extent.
Can a physical server be accessed using APIs?
Yes.
A physical server can be accessed and managed using APIs. A physical server provides a set of APIs that you can use to manage and manipulate your physical server instances. The APIs allow you to create, start, stop, and restart physical servers, obtain physical server status information, manage network configurations, and adjust resource configurations. With APIs, you can automate and integrate the management process of physical servers to improve efficiency and flexibility.
What system images do physical servers support?
Standard bare metal has been adapted to CPUs including X86, Kunpeng, Hygon, and Phytium, and supports system images including CTyunOS, CentOS, and Ubuntu.
Elastic IP bare metal has been adapted to CPUs including X86, Kunpeng, and Hygon, and supports system images including CTyunOS, CentOS, and Ubuntu.
For the support list, see Physical Server Type and Supported OS Version.
Does a dedicated server support hyper-threading?
Yes.
The eSurfing Cloud physical server service supports hyper-threading. Hyper-threading is enabled by default. To cancel or configure hyper-threading, modify the bios and restart. Services will be affected.
Enable hyper-threading: Applicable to scenarios that require the CPU kernel to process a lot of information in parallel or background tasks at a time. In these scenarios, hyper-threading can improve computing efficiency.
Disable hyper-threading: Applicable to computing-intensive scenarios where the performance is better when hyper-threading is disabled than when hyper-threading is enabled, such as HPC scenarios such as material computing.
How do I apply for a resource quota increase for a dedicated server?
You can apply for a quota increase in the following steps:
1. Log in to the console.
2. Click Location on the upper part of the console and select your region.
3. In the left-side navigation pane, click Service List, and go to Computing > Physical Server.
4. On the physical server list page, click View Quota Details.
5. In the upper right corner, click Apply for Quota Increase.
6. On the Create Ticket page, enter related parameters as needed.
7. Among them, enter the content to adjust and the reason for application in the Issue Description field.
8. Then, check the Agreement and click Submit.
How do I synchronize time on a physical server?
Use Network Time Protocol (NTP): NTP is a network protocol used to synchronize time on computer systems. You can configure an NTP server for physical servers to get accurate time information. You can specify the IP address of the NTP server to use and the physical server will periodically synchronize time with the NTP server.
Use the time synchronization tool of the operating system: A time synchronization tool is available in different operating systems, such as the Windows Time Service on the Windows system and the NTP client on the Linux system. You can configure these tools to synchronize with external time sources to ensure that the physical server's time is consistent with standard time.
Manually set time: You can also manually set the time of the physical server, but this method is not accurate and automated enough for large-scale deployments.
In ELB, can a physical server be selected as a backend instance?
Yes.
In ELB, a physical server (elastic bare metal) can be selected as a backend instance. Currently, only elastic bare metal can be used as a backend instance for elastic load balancing.
What RAID levels are supported by a physical server?
By default, NO RAID, Raid0, Raid1, Raid5, and Raid10 are provided. Users can set RAID levels for local disks, but not for EVS disks.