yexinru/doc-exports
1
0
Fork 0
forked from docs/doc-exports
Code Pull Requests Activity
Files
8f01f116a5ad22cc6ad22bdf6cdc797063a2f5f1
doc-exports/docs/ecs/umn/en-us_topic_0094118976.html
wanghuijuan738 8f01f116a5 ECS UMN 20260514 version. Added X1m instance specifications. 
Reviewed-by: Pristromskaia, Margarita <margarita.pristromskaia@t-systems.com>
Co-authored-by: wanghuijuan738 <wanghuijuan738@huawei.com>
Co-committed-by: wanghuijuan738 <wanghuijuan738@huawei.com>
2026-05-18 09:03:42 +00:00

72 KiB
Raw Blame History

Ultra-high I/O ECSs

Overview

Ultra-high I/O ECSs use high-performance local NVMe SSDs to provide high storage input/output operations per second (IOPS) and low read/write latency. You can create such ECSs on the management console.

Scenarios

  • High-performance relational databases
  • NoSQL databases (such as Cassandra and MongoDB) and ElasticSearch

Specifications

Table 1 I3m ECS specifications

Flavor

vCPUs

Memory

(GiB)

Max./Assured Network Bandwidth

(Gbit/s)

Max. Network PPS

(10,000)

Max. NIC Queues

Local Disks

(GiB)

Max. NICs

Virtualization

i3m.2xlarge.8

8

64

2.5/2.5

100

4

1 × 1,600 GiB NVMe

4

KVM

i3m.4xlarge.8

16

128

5/5

150

4

2 × 1,600 GiB NVMe

8

KVM

i3m.8xlarge.8

32

256

10/10

200

8

4 × 1,600 GiB NVMe

8

KVM

i3m.12xlarge.8

48

384

15/15

240

8

6 × 1,600 GiB NVMe

8

KVM

i3m.15xlarge.8

60

512

25/25

500

16

7 × 1,600 GiB NVMe

8

KVM

i3m.16xlarge.8

64

512

25/25

500

16

8 × 1,600 GiB NVMe

8

KVM
Table 2 I3 ECS specifications

Flavor

vCPUs

Memory

(GiB)

Max./Assured Network Bandwidth

(Gbit/s)

Max. Network PPS

(10,000)

Max. NIC Queues

Max. NICs

Local Disks

(GiB)

Virtualization

i3.2xlarge.4

8

32

15/4.5

150

4

4

1 × 3,200 GiB NVMe

KVM

i3.4xlarge.4

16

64

20/9

280

8

8

2 × 3,200 GiB NVMe

KVM

i3.8xlarge.4

32

128

30/18

550

16

8

4 × 3,200 GiB NVMe

KVM

i3.12xlarge.4

48

192

35/27

750

16

8

6 × 3,200 GiB NVMe

KVM

i3.16xlarge.4

64

256

40/32

1,000

32

8

8 × 3,200 GiB NVMe

KVM

i3.2xlarge.8

8

64

15/4.5

150

4

4

1 × 3,200 GiB NVMe

KVM

i3.4xlarge.8

16

128

20/9

280

8

8

2 × 3,200 GiB NVMe

KVM

i3.8xlarge.8

32

256

30/18

550

16

8

4 × 3,200 GiB NVMe

KVM

i3.12xlarge.8

48

384

35/27

750

16

8

6 × 3,200 GiB NVMe

KVM

i3.16xlarge.8

64

512

40/32

1,000

32

8

8 × 3,200 GiB NVMe

KVM

Local Disk Performance

Table 3 lists the local disk IOPS of I3m instance flavors. Table 5 lists the performance of a single local disk used by an I3m ECS.

Table 4 lists the local disk IOPS of I3 instance flavors. Table 6 lists the performance of a single local disk used by an I3 ECS.

Table 3 Local disk IOPS of I3m instance flavors

Flavor

Maximum IOPS for Random 4 KB Read

i3m.2xlarge.8

750,000

i3m.4xlarge.8

1,500,000

i3m.8xlarge.8

3,000,000

i3m.12xlarge.8

4,500,000

i3m.15xlarge.8

5,250,000

i3m.16xlarge.8

6,000,000
Table 4 Local disk IOPS of I3 instance flavors

Flavor

Maximum IOPS for Random 4 KB Read

i3.2xlarge.4

750,000

i3.4xlarge.4

1,500,000

i3.8xlarge.4

3,000,000

i3.12xlarge.4

4,500,000

i3.16xlarge.4

6,000,000

i3.2xlarge.8

750,000

i3.4xlarge.8

1,500,000

i3.8xlarge.8

3,000,000

i3.12xlarge.8

4,500,000

i3.16xlarge.8

6,000,000
Table 5 Performance of a single local disk used by an I3m ECS

Metric

Performance

Disk capacity

1.6 TB

IOPS for random 4 KB read

750,000

IOPS for random 4 KB write

200,000

Read throughput

2.9 GiB/s

Write throughput

1.9 GiB/s

Access latency

Within microseconds
Table 6 Performance of a single local disk used by an I3 ECS

Metric

Performance

Disk capacity

1.6 TB

IOPS for random 4 KB read

750,000

IOPS for random 4 KB write

200,000

Read throughput

2.9 GiB/s

Write throughput

1.9 GiB/s

Access latency

Within microseconds

Notes

  • Table 7 lists the OSs supported by ultra-high I/O ECSs.
    Table 7 Supported OS versions

    OS

    Version

    Alma

    Alma 8 64bit

    CentOS
    • CentOS Stream 9.6 64bit
    • CentOS Stream 8.6 64bit
    • CentOS 8.5 64bit
    • CentOS 8.4 64bit
    • CentOS 8.3 64bit
    • CentOS 8.2 64bit
    • CentOS 8.1 64bit
    • CentOS 7.9 64bit
    • CentOS 7.7 64bit

    Debian
    • Debian GNU/Linux 12.0.0 64bit
    • Debian GNU/Linux 11.7.0 64bit
    • Debian GNU/Linux 11.6.0 64bit
    • Debian GNU/Linux 11.5.0 64bit
    • Debian GNU/Linux 11.4.0 64bit
    • Debian GNU/Linux 11.3.0 64bit
    • Debian GNU/Linux 11.2.0 64bit
    • Debian GNU/Linux 11.1.0 64bit
    • Debian GNU/Linux 11.0.0 64bit
    • Debian GNU/Linux 10.13.0 64bit
    • Debian GNU/Linux 10.12.0 64bit
    • Debian GNU/Linux 10.11.0 64bit
    • Debian GNU/Linux 10.10.0 64bit
    • Debian GNU/Linux 10.9.0 64bit
    • Debian GNU/Linux 10.8.0 64bit
    • Debian GNU/Linux 10.7.0 64bit
    • Debian GNU/Linux 10.6.0 64bit
    • Debian GNU/Linux 10.5.0 64bit
    • Debian GNU/Linux 10.4.0 64bit
    • Debian GNU/Linux 10.3.0 64bit
    • Debian GNU/Linux 10.2.0 64bit
    • Debian GNU/Linux 10.1.0 64bit
    • Debian GNU/Linux 10 64bit

    EulerOS
    • EulerOS 2.12 64bit
    • EulerOS 2.11 64bit
    • EulerOS 2.10 64bit
    • EulerOS 2.8 64bit
    • EulerOS 2.7 64bit
    • EulerOS 2.5 64bit

    Fedora
    • Fedora 39 64bit
    • Fedora 38 64bit
    • Fedora 37 64bit
    • Fedora 36 64bit
    • Fedora 35 64bit
    • Fedora 34 64bit
    • Fedora 33 64bit
    • Fedora 32 64bit
    • Fedora 31 64bit

    OpenSUSE
    • OpenSUSE 15.5 64bit
    • OpenSUSE 15.4 64bit
    • OpenSUSE 15.3 64bit
    • OpenSUSE 15.2 64bit

    Oracle Linux
    • Oracle Linux Server release 8.4 64bit
    • Oracle Linux Server release 7.6 64bit

    Red Hat
    • Red Hat Enterprise Linux 9.1 64bit
    • Red Hat Enterprise Linux 9.0 64bit
    • Red Hat Enterprise Linux 8.7 64bit
    • Red Hat Enterprise Linux 8.6 64bit
    • Red Hat Enterprise Linux 8.5 64bit
    • Red Hat Enterprise Linux 8.4 64bit
    • Red Hat Enterprise Linux 8.3 64bit
    • Red Hat Enterprise Linux 8.2 64bit
    • Red Hat Enterprise Linux 8.1 64bit
    • Red Hat Enterprise Linux 7.9 64bit
    • Red Hat Enterprise Linux 6.10 64bit

    Rocky
    • 9.2 64bit
    • 9.1 64bit
    • 9.0 64bit
    • 8.8 64bit
    • 8.7 64bit
    • 8.6 64bit
    • 8.5 64bit
    • 8.4 64bit
    • 8.3 64bit
    • 8 64bit

    SUSE
    • Novell SUSE Linux Enterprise Server 15 SP5 64bit
    • Novell SUSE Linux Enterprise Server 15 SP4 64bit
    • Novell SUSE Linux Enterprise Server 15 SP3 64bit
    • Novell SUSE Linux Enterprise Server 15 SP2 64bit
    • Novell SUSE Linux Enterprise Server 15 SP1 64bit
    • Novell SUSE Linux Enterprise Server 15 64bit
    • Novell SUSE Linux Enterprise Server 12 SP5 64bit
    • Novell SUSE Linux Enterprise Server 12 SP4 64bit
    • Novell SUSE Linux Enterprise Server 12 SP3 64bit

    Ubuntu
    • Ubuntu 24.04 server 64bit
    • Ubuntu 22.04 Server 64bit
    • Ubuntu 20.04 server 64bit
    • Ubuntu 18.04 server 64bit

    Windows
    • Windows Server 2025 Standard 64bit
    • Windows Server 2022 Standard 64bit
    • Windows Server 2022 Datacenter 64bit
    • Windows Server 2019 Datacenter 64bit
    • Windows Server 2019 Standard 64bit
    • Windows Server 2016 Standard 64bit
    • Windows Server 2012 R2 Standard 64bit

    openEuler
    • openEuler 22.03 SP1 64bit
    • openEuler 22.03 64bit
    • openEuler 20.03 SP3 64bit
    • openEuler 20.03 SP2 64bit
    • openEuler 20.03 SP1 64bit
    • openEuler 20.03 64bit
  • If the host where an ultra-high I/O ECS is deployed is faulty, the ECS cannot be restored through live migration.
    • If the host is faulty or subhealthy, you need to stop the ECS for hardware repair.
    • In case of system maintenance or hardware faults, the ECS will be redeployed (to ensure HA) and cold migrated to another host. The local disk data of the ECS will not be retained.
  • Ultra-high I/O ECSs do not support specification modification.
  • Ultra-high I/O ECSs do not support snapshots or backups of local disks.
  • Ultra-high I/O ECSs can use local disks, and can also have EVS disks attached to provide a larger storage size. Note the following when using the two types of storage media:
    • Only an EVS disk, not a local disk, can be used as the system disk of an ultra-high I/O ECS.
    • Both EVS disks and local disks can be used as data disks of an ultra-high I/O ECS.
    • An ultra-high I/O ECS can have a maximum of 60 attached disks (including VBD, SCSI, and local disks). For details about constraints, see Can I Attach Multiple Disks to an ECS?
  • You can configure the fstab file to set automatic disk mounting at ECS startup. For details, see Initializing EVS Data Disks.
  • The local disk data of an ultra-high I/O ECS may be lost if an exception occurs, such as physical server breakdown or local disk damage. If your application does not use the data reliability architecture, it is a good practice to use EVS disks to build your ECS.
  • When an ultra-high I/O ECS is deleted, the data on local NVMe SSDs will also be automatically deleted. Back up data in advance. Deleting local disk data will increase the time needed to release resources compared with ECSs without local disks.
  • The reliability of local disk data depends on the reliability of physical servers and hard disks, which are SPOF-prone. It is a good practice to use data redundancy at the application layer to ensure data availability. Use EVS disks to store service data that needs to be stored for a long time.
  • The device name of a local disk attached to an ultra-high I/O ECS is /dev/nvme0n1 or /dev/nvme0n2.
  • The basic resources, including vCPUs, memory, and the image of an ultra-high I/O ECS will continue to be billed after the ECS is stopped. To stop the ECS from being billed, delete it and its associated resources.

Handling Damaged Local Disks Used by I-Series ECSs

If a local disk attached to an ECS is damaged, perform the following operations to handle this issue:

For a Linux ECS:

  1. Detach the faulty local disk.
    1. Run the following command to query the mount point of the faulty disk:

      df -Th

      Figure 1 Querying the mount point
    2. Run the following command to detach the faulty local disk:

      umount mount-point

      In the example shown in Figure 1, the mount point of /dev/nvme0n1 is /mnt/nvme0. Run the following command:

      umount /mnt/nvme0

  2. Check whether the mount point of the faulty disk is configured in /etc/fstab of the ECS. If yes, comment out the mount point to prevent the ECS from entering the maintenance mode upon ECS startup after the faulty disk is replaced.
    1. Run the following command to obtain the partition UUID:

      blkid disk-partition

      In this example, run the following command to obtain the UUID of the /dev/nvme0n1 partition:

      blkid /dev/nvme0n1

      Information similar to the following is displayed:

      /dev/nvme0n1: UUID="b9a07b7b-9322-4e05-ab9b-14b8050cd8cc" TYPE="ext4"
    2. Run the following command to check whether /etc/fstab contains the automatic mounting information about the disk partition:

      cat /etc/fstab

      Information similar to the following is displayed:

      UUID=b9a07b7b-9322-4e05-ab9b-14b8050cd8cc    /mnt   ext4    defaults        0 0
    3. If the mounting information exists, perform the following steps to delete it.
      1. Run the following command to edit /etc/fstab:

        vi /etc/fstab

        Use the UUID obtained in 2.a to check whether the mounting information of the local disk is contained in /etc/fstab. If yes, comment out the information. This prevents the ECS from entering the maintenance mode upon ECS startup after the local disk is replaced.

      2. Press i to enter editing mode.
      3. Delete or comment out the automatic mounting information of the disk partition.

        For example, add a pound sign (#) at the beginning of the following command line to comment out the automatic mounting information:

        # UUID=b9a07b7b-9322-4e05-ab9b-14b8050cd8cc    /mnt   ext4    defaults        0 0
      4. Press Esc to exit editing mode. Enter :wq and press Enter to save the settings and exit.
  3. Run the following command to obtain the SN of the local disk:

    For example, if the nvme0n1 disk is faulty, obtain the serial number of the nvme0n1 disk.

    ll /dev/disk/by-id/

    Figure 2 Querying the serial number of the faulty local disk
  4. Stop the ECS and provide the serial number of the faulty disk to technical support personnel to replace the local disk.

    After the local disk is replaced, restart the ECS to synchronize the new local disk information to the virtualization layer.

For a Windows ECS:

  1. Open Computer Management, choose Computer Management (Local) > Storage > Disk Management, and view the disk ID, for example, Disk 1.
  2. Open Windows PowerShell as an administrator and run the following command to query the disk on which the logical disk is created:

    Get-CimInstance -ClassName Win32_LogicalDiskToPartition |select Antecedent, Dependent | fl

    Figure 3 Querying the disk on which the logical disk is created
  3. Run the following command to obtain the serial number of the faulty disk according to the mapping between the disk ID and serial number:

    Get-Disk | select Number, SerialNumber

    Figure 4 Querying the mapping between the disk ID and serial number

    If the serial number cannot be obtained by running the preceding command, see Using a Serial Number to Obtain the Disk Device Name (Windows).

  4. Stop the ECS and provide the serial number of the faulty disk to technical support personnel to replace the local disk.

    After the local disk is replaced, restart the ECS to synchronize the new local disk information to the virtualization layer.

Parent topic: ECS Specifications
<script language="JavaScript"> </script>