Block Storage FAQ_ ECS - Alibaba Cloud Help Center

What is a storage capacity unit package SCU?

SCU (Storage Capacity Unit) is a prepaid storage sales form, which can offset the pay as you go bill of some storage products such as cloud disks. Compared with the mode of purchasing cloud disks with monthly and yearly ECS instances, the combination of SCU and pay as you go cloud disks has both cost performance and flexibility in resource use. For more information, see Storage Capacity Unit Package Overview 。

Which block storage products can use the storage capacity unit package SCU?

The storage capacity unit package SCU can deduct the pay as you go bill for some storage products. The specific deduction products are as follows:

Block storage products that support deduction include ESSD cloud disk, SSD cloud disk, efficient cloud disk and ordinary cloud disk; Local disks cannot be deducted.
NAS products supporting deduction include capacity NAS and performance NAS; It does not support deduction of high-speed NAS and low-frequency media.
Supports deduction of snapshots.
OSS products supporting deduction include standard OSS, low-frequency OSS and archive OSS.

Can the storage capacity unit SCU be used separately?

may not. The SCU must match the pay as you go cloud disk and deduct the pay as you go bill before it can work.

What are the deduction rules for storage capacity unit package SCUs?

The SCU deducts the cloud disk pay as you go bill by a certain discount factor. For more information, see Storage capacity unit package deduction rules 。

How is the SCU charged?

SCUs are priced according to capacity, and prices vary in different regions. For the price list of SCUs in different regions, see Cloud product price page 。

Can I get a refund from SCU?

You can apply for a refund if:

Within five days of purchasing an SCU, you can apply for a refund without reason.
explain
Each AliCloud account has only one chance to refund SCUs for five days without reason every year, that is, each account can refund SCUs at most once a year, and the maximum amount of refundable is one storage capacity unit package.
For other information, please Submit work order Contact Alibaba Cloud customer service for processing.

When refunding, the amount already spent will be deducted, and the used coupons and other discounts will not be returned. For more information, see Refund rules and procedures 。

What is an ESSD cloud disk?

The ESSD (Enterprise SSD) cloud disk is an ultra-high performance cloud disk product launched by Alibaba Cloud. The ESSD cloud disk combines the 25GE network and RDMA technology to provide you with a single disk's random read/write capability of up to 1 million and lower one-way delay capability. For more information, see ESSD cloud disk 。

What are the product specifications provided by the ESSD cloud disk?

The ESSD cloud disks are divided into different specifications according to their performance levels. For the latest ESSD cloud disk performance details, see ESSD cloud disk 。

For storage devices, different capacities can achieve different performance. The larger the capacity, the stronger the data processing capacity of the storage device. The I/O performance per unit capacity of all ESSD cloud disk performance levels is the same, but the cloud disk performance increases linearly with the capacity growth until it reaches the upper limit of the single disk performance of this performance level.

Performance level	ESS Cloud Disk Capacity Range (GiB)	Maximum IOPS	Maximum throughput (MB/s)
PL0	1~65,536	10,000	one hundred and eighty
PL1	20~65,536	50,000	three hundred and fifty
PL2	461~65,536	100,000	seven hundred and fifty
PL3	1,261~65,536	1,000,000	4,000

What are the differences and similarities between an ESS cloud disk and an SSD cloud disk or an efficient cloud disk?

Common point: the three cloud disks are cloud disk products based on distributed block storage architecture, with features such as high reliability and elastic capacity expansion, and support data functions such as snapshot and encryption.
Difference: Compared with SSD cloud disks or efficient cloud disks, the performance of ESS cloud disks has been greatly improved. For more information, see ESSD cloud disk and Block storage performance 。

How are the performance indicators and product specifications of the ESSD cloud disk calculated?

The performance of the ESS cloud disk is linearly related to the capacity. The larger the capacity, the higher the performance. Compared with the SSD cloud disk, the performance of the ESS cloud disk is significantly improved. For more information, see ESSD cloud disk 。

How to test the performance of the ESSD cloud disk?

You can use the FIO tool to pressure test the ESSD cloud disk performance. See Test the performance of ESSD cloud disk IOPS 。

What is the relationship between ESSD cloud disk performance and instance performance?

The storage I/O performance of some instance dimensions is linearly proportional to the instance type. For example, the storage I/O performance of ECS storage enhanced instance g7se is linearly related to the instance type. The higher the instance type, the higher the storage IOPS and throughput.

When you create a g7se instance of a certain specification and mount the ESSD cloud disk:

If the total performance of the ESSD cloud disk does not exceed the storage I/O capacity corresponding to the instance type family, the actual storage performance is subject to the performance of the ESSD cloud disk.
If the total performance of the ESSD cloud disk exceeds the storage I/O capacity corresponding to the instance type family, the actual storage performance is subject to the storage I/O capacity corresponding to the instance type family.
For example, when you create an ecs.g7se.xlarge 16GiB instance, the maximum storage IOPS of the instance is 60000. If an ESS cloud disk with a storage I/O capacity of 2 TiB is attached (the IOPS of a single disk is 101800), the maximum storage IOPS of this instance can only be 60000, and cannot reach the 101800 IOPS of the 2 TiB ESS cloud disk.

For the specifications and performance of ECS storage enhanced instance g7se, see Instance Type Family 。

How is the ESSD cloud disk charged?

Support monthly package and pay as you go. Please refer to Block Storage Price 。

What instance type families can an ESSD cloud disk be attached to?

For instance type families supported by ESSD cloud disk, see Instance Type Family 。

What tools can be used to test block storage performance?

See Test block storage performance 。

Why does FIO cause instance downtime when testing performance?

FIO test tool supports raw disk partition and file system to test I/O performance. If you test the raw disk partition directly, the file system metadata in the raw disk partition will be damaged, resulting in the inability to access the files in the raw disk partition, or even the instance downtime. The above problem does not exist when using FIO's file system to test the performance mode.

How to select a zone according to application requirements?

Volume based cloud disks can only be attached to ECS instances in the same zone.

For high availability applications, it is recommended that you create multiple data disks in different zones and attach them to ECS instances in the corresponding zones.
For applications with low network latency, it is recommended that you create data disks in the same zone and attach them to ECS instances in the zone.

What are the common operations on cloud disks?

For common operations of cloud disk, see Overview of cloud disk The relevant operation section of.

How can I view the usage and remaining space of block storage?

You can log in to the ECS instance and view the usage and remaining space of block storage inside the instance. The ECS management console and ECS API cannot view this type of information temporarily.

Can I compress the capacity of cloud disk?

Alibaba Cloud does not support shrinking cloud disk space. If the cloud disk space you purchased is large and needs to be reduced, it is recommended that you first create and mount a new cloud disk of appropriate capacity, copy the data required by the old disk to the new disk, and then release the old disk.

How to release monthly cloud disks in advance?

Alibaba Cloud does not support the release of monthly data disks. You can change the billing method to pay as you go before releasing the cloud disk. Before releasing the cloud disk, please confirm that you have backed up important data. For details, see Convert cloud disk billing method and Release Cloud Disk 。

explain

After the monthly subscription data disk is converted to pay as you go, the billing bill will be billed on an hourly basis. After the cloud disk is released, there will be no further deduction for one hour. After the cloud disk billing method is converted, the actual refund amount is subject to the page display, and the used vouchers will not be returned.

What is I/O optimization? Can I upgrade an existing ECS instance to an I/O optimized instance?

I/O optimization provides better network capability between the instance and the cloud disk, and can ensure the performance of SSD cloud disk storage. For I/O optimized instances, when you mount an SSD cloud disk, you can obtain all the storage performance of the SSD cloud disk.

Purchased stock ECS instances can be passed ModifyInstanceSpec and ModifyPrepayInstanceSpec Convert non I/O optimized instances to I/O optimized instances.

How does Alibaba Cloud handle the contention of block storage resources?

Alibaba Cloud Block Storage is a multi tenant storage service. We define performance standards according to different instance specifications and cloud disk specifications. For example, cloud disk performance specifications can be divided into different specifications such as ESS cloud disk, SSD cloud disk, or efficient cloud disk. Users can obtain different performance by purchasing different cloud disk specifications. At the same time, you can also change the performance specifications of cloud disks by adjusting their size or changing their type.

Users can monitor cloud disk performance data through cloud monitoring or CloudLens for EBS to determine whether the type of cloud disk currently purchased meets business needs. For more information about cloud monitoring and CloudLens for EBS, see Cloud product monitoring and What is Block Storage Data Insight 。

Alibaba Cloud storage services will provide support, avoid resource contention, and continuously monitor by allocating appropriate infrastructure. When hardware failure or contention is monitored, Alibaba Cloud will dispatch or repair resources (depending on the extent of infrastructure damage).

What I/O performance does SSD cloud disk have?

See Block storage performance 。

What application scenarios are SSD cloud disks applicable to?

SSD cloud disk has the characteristics of high performance and reliability, and is suitable for I/O-intensive applications with high data reliability requirements, such as MySQL, SQL Server, Oracle, PostgreSQL and other small and medium-sized relational database scenarios. At the same time, it is suitable for small and medium-sized development and test environments that require high data reliability.

Do you support replacing an original ordinary cloud disk with an SSD cloud disk?

Since the SSD cloud disk is implemented with a full SSD storage medium, it is not supported to directly replace the existing ordinary cloud disk with an SSD cloud disk.

How to purchase SSD cloud disks, I/O optimized instances and what is the price of SSD cloud disks?

See SSD cloud disk and I/O optimization instance Price Details Page 。

Is upgrading supported after purchasing SSD cloud disks?

Support upgrading and capacity expansion. For details, see Cloud Disk Expansion Guidelines 。

After I/O optimization instance and SSD cloud disk are used, why does the Linux system report an error when the partition is mounted?

The mount point of the SSD cloud disk in the Linux system is displayed as/dev/vd *, which is different from the common mount point/dev/xvd *. It is recommended that you fill in the mount command in the format of/dev/vd * mount point.

What should we pay attention to when adding mount information to ordinary cloud disks and SSD cloud disks?

When adding data disks to Linux systems, when formatting data disks and editing partition information, it is important to note that the device names of data disks for I/O optimization instances and non I/O optimization instances are different. The device name of the data disk is assigned by the system by default. The naming rules are as follows:

I/O optimization instance:
- The name of the data disk device attached through the NVMe protocol is /dev/nvmeXn1 , for example /dev/nvme1n1 、 /dev/nvme2n1 、 /dev/nvme3n1 , and so on. For more information about cloud disks that support the NVMe protocol, see Overview of NVMe Cloud Disk 。
- The name of the data disk device that is not attached to the NVMe protocol is /dev/vd[b-z] , for example /dev/vdb 、 /dev/vdc and /dev/vdd Etc.
Non I/O optimization instance: the data disk device name is /dev/xvd[b-z] , for example /dev/xvdb 、 /dev/xvdc and /dev/xvdd Etc.

important

It is recommended that you /etc/fstab The globally unique identifier UUID is used in to reference the new partition. The device name of other cloud disks may change due to operations such as releasing cloud disks /etc/fstab The device name is directly used in. When the device name changes, your storage data may be affected.

If incorrect information is added, use mount -a The command will not mount. The solution is as follows:

function fdisk -l View specific data disk information.
see /etc/fstab Whether the information added in it corresponds.
explain
Do not add mount information repeatedly, or the instance will be powered on and repeatedly mounted, and the system will fail to start.
use vim modify /etc/fstab File.
Annotate or delete the wrong information, and add the corresponding mount information.
function mount -a Test whether the mount is successful.

Refer to Initialize TiB data disks less than or equal to 2 (Linux) 。

What is a device name (mount point)?

The device name (mount point) is the location of the cloud disk on the ECS on the disk controller bus. The selected device name corresponds to the disk device number in the Linux operating system, and the disk order in the Disk Manager is consistent in the Windows operating system.

What is an independent cloud disk?

An independent cloud disk is a pay as you go data disk purchased separately, not a data disk created with an instance. It can be mounted and uninstalled freely between different ECS instances in the same zone, but it cannot be mounted across regions and zones. The independent cloud disk must be attached to the instance and formatted in partitions before it can be used. For more information, see Create cloud disk 。

Can a cloud disk be attached to multiple ECS instances?

No. A cloud disk can only be attached to one instance in the same zone. Cross zone mounting is not supported.

After purchasing pay as you go cloud disks and attaching them to ECS instances, do you still need to attach partitions?

After purchasing a pay as you go data disk separately, you need to attach the data disk to the instance and format it by partition. For more information, see Initialize TiB data disks less than or equal to 2 (Linux) and Initialize TiB data disks less than or equal to 2 (Windows) 。

What should I do if I buy a data disk for the instance but cannot see it in the system?

Linux system
If you purchase a pay as you go data disk separately, you need to format and mount the partition before you can use and see the space. For more information, see Initialize TiB data disks less than or equal to 2 (Linux) and Attach data disk 。
Windows system
For reasons and solutions, see Resolve the problem that the data disk cannot be viewed in the Windows instance 。

How many cloud disks can an instance mount?

An ECS instance can mount up to 64 data disks. The maximum number of cloud disks supported by different instance types varies. For more information, see Instance Type Family 。

explain

When creating an instance, you can mount up to 1 system disk and 16 data disks. If the instance needs more data disks, you need to continue to mount after creating the instance. The maximum number of cloud disks supported by different instance types varies. You can use DescribeInstanceTypes The interface queries the number of cloud disks that an instance can attach.

Why can't I find the instance I want to mount when I mount the cloud disk?

Please ensure that your instance has not been released and that the instance and cloud disk are in the same zone in the same region.

Can cloud disks and instances be mounted in different zones?

may not. You can only freely mount and uninstall pay as you go cloud disks between different ECS instances in the same zone.

Will the cloud disk data be lost when the cloud disk (data disk) is uninstalled?

Under the Windows operating system, in order to ensure data integrity, it is strongly recommended that you suspend the read and write operations on all file systems of the cloud disk, otherwise the data that has not completed the read and write operations will be lost.
Under Linux operating system, you need to log in to the instance to execute umount Command. After the command is successfully executed, enter the console to uninstall the cloud disk.

Can the system disk be uninstalled?

Uninstall is supported. See Uninstall or mount the system disk 。

How is the pay as you go data disk purchased separately charged?

Pay as you go data disks are billed by the hour. If your account balance is insufficient, the service of pay as you go data disks will be stopped.

Why is the cloud disk I created separately released together with my instance?

The cloud disk can be set to automatically release with the instance. You can turn the release with instance function on or off through the console or API. For more information, see Release Cloud Disk 。

Can a pay as you go data disk purchased separately be attached to a subscription ECS instance?

sure.

Can the data disks on the subscription ECS instance be uninstalled?

The yearly/monthly ECS instance does not support direct uninstallation of data disks. The expiration time of the data disk is the same as that of the instance, and it is released with the instance. If you want to release a data disk, you can first convert the monthly subscription data disk to a pay as you go data disk, and then uninstall and release the data disk. For how to convert the cloud disk billing method, see Convert cloud disk billing method 。

I have already operated renewal and configuration change. Can I change the subscription cloud disk to pay as you go cloud disk during the renewal and configuration change period?

No. You can wait until the renewal and configuration change period is over, and then select renewal and configuration change to convert to pay as you go cloud disks.

When deleting cloud disks, will snapshots be retained?

Manual snapshots are not deleted. Whether the automatic snapshot will be deleted depends on whether you have set Automatic snapshot release with disk 。 For more information, see Set automatic snapshot release with cloud disk 。

Why doesn't my cloud disk have an automatic snapshot?

When the number of snapshots reaches the snapshot quota, the system will automatically delete the earliest created automatic snapshot, and the manual snapshot will not be affected.

explain

The automatic snapshot policy will only be executed when the cloud disk is attached to an instance.

Can I use snapshots to create cloud disks separately?

sure. When creating pay as you go cloud disks separately, select the existing snapshot creation. For more information, see Create a cloud disk using snapshots 。

How to troubleshoot the data disk mounting problem that makes the data inaccessible?

If you cannot access the data of the Linux data disk, you can refer to the following troubleshooting ideas.

Find the cloud disk where the data resides, and confirm whether the data disk is attached to the corresponding ECS instance.
- View on the console. See View cloud disk monitoring information 。
- Inside the instance, use fdisk -l Command to check whether the partition information of the data disk is normal, and use the df -h and mount | grep "<devpath>" Command to view the mount status.
use cat Command View /etc/fstab File, whether to mount the two cloud disks to the same directory.
- If the cloud disk is mounted in the same directory, the cloud disk that was mounted first will be replaced by the cloud disk that was mounted later, resulting in inaccessible data. It is recommended to mount one cloud disk to another directory.
- If it is not mounted in the same directory, but the mounting information is still displayed in the same directory, use ll Command to check whether there is a connection between the two attached directories. If there is a connection, it is recommended to use mkdir Command to create a new directory to mount and test whether the data can be accessed.

What about data loss after restarting the ECS instance of the Linux system?

Symptom: A directory is found after restarting the instance (for example /alidata ）All data is lost.
Cause: Operation df -h It is found that the data disk partition is not attached.
Solution: This method uses an I/O optimization instance to mount data disks that are not NVMe protocols /dev/vdb1 Zoning is an example. If it is a data disk attached to the I/O optimization instance NVMe protocol, the cloud disk device is named /dev/nvmeXn1 ； If it is a non I/O optimization instance, the cloud disk device name is /dev/xvd[b-z] 。 Please operate according to the actual situation.
1. function fdisk -l Command to view the unmounted data disk partition.
2. function mount /dev/vdb1 /alidata Mount the data disk partition.
3. function df -h Check whether the data disk partition is successfully mounted.
4. (Optional) On /etc/fstab To avoid the same problem next time, set auto attach when starting up.

Will my snapshot be lost when the cloud disk is reinitialized?

No, neither manual snapshot nor automatic snapshot will be lost.

What if the data disk disappears after restarting the ECS instance of the Linux system or reinitializing the system disk?

Symptom: After the Linux instance restarts or initializes the system, log in to the instance to run df -h Check the cloud disk mount and find that the data disk is missing.
Cause:
- Restart the instance: If you do not write the mount information to the /etc/fstab After restarting the instance, the instance will not automatically mount the data disk.
- Reinitialize the system disk: Initializing the system disk will reset /etc/fstab The instance will not automatically mount the data disk after reinitializing the system disk.
Solution:
This method uses a non I/O optimization instance /dev/xvdb1 Zoning is an example. If the I/O optimization instance is a data disk attached to a non NVMe protocol, the cloud disk device name is /dev/vd[b-z] ； If it is a data disk attached to the I/O optimization instance NVMe protocol, the cloud disk device is named /dev/nvmeXn1 。 Please operate according to the actual situation.
1. function mount /dev/xvdb1 Command to remount the data disk.
2. function mount Command View Data Disk Partition /dev/xvdb1 File format.
3. Suppose you see /dev/xvdb1 The partition uses the ext3 file system. Execute the following command to write the data disk mount information to the /etc/fstab File.
```
 echo '/dev/xvdb1 /data ext3 defaults 0 0' >> /etc/fstab
```
4. Restart the ECS instance on the console.

How to re mount the data disk after the Linux ECS instance reinitializes the system disk?

In the Linux instance, reinitializing the system disk will not change the contents of the data disk, but the mount information of the data disk will be lost. Assume that before reinitializing the system disk, the partition name of the data disk attached to the instance is /dev/vdb1 , the mount point name is /InitTest 。 After restarting the Linux instance, see the following steps to create new mount point information and mount the data disk partition.

function mount View the data disk mount information.
None of the returned results /dev/vdb1 Information for.
function fdisk -l View data disk partition information.
function cat /etc/fstab View data disk partition /dev/vdb1 The original mount point name.
function mkdir /InitTest Re create the mount point of the data disk partition.
The name of the new mount point must be the same as that of the reinitialized system disk /dev/vdb1 The mount point names of are consistent.
function mount /dev/vdb1 /InitTest Re mount the data disk partition.
function df -h View the mount results.
Test the data disk as follows /dev/vdb1 Whether it can be automatically attached.
1. function umount /dev/vdb1 uninstall /dev/vdb1 。
2. function mount View the mount information.
  If the uninstallation is successful, there will be no /dev/vdb1 Information for.
3. function mount -a Auto Mount /dev/vdb1 。
4. function mount View the mount information.
  If the auto mount is successful, the returned result will contain /dev/vdb1 Information for.

What is the difference between reinitializing the system disk and replacing the system disk (operating system)?

The main differences are shown in the table below:

Variance Item	Reinitialize the system disk	Replace the system disk (operating system)
Functional differences	Initialization will restore the initialization status of the ECS, and the operating system will remain unchanged before and after initialization.	Switch the existing operating system type to another type of operating system.
Impact on system disk	The old system disk will not be released. The system disk data will be restored to the state when it was just created. After the instance is created, the applications installed and the data generated will be cleared. It is recommended that you backup the data. The system disk ID, cloud disk type, and ECS instance IP address remain unchanged.	After replacing the operating system, the old system disk will be released. All data on the system disk will be cleared. It is recommended that you backup the data. The system disk ID will change, but the cloud disk type, ECS instance IP address, and elastic network card MAC address will remain unchanged.
Impact on data disk	The data disk will not be affected.	The data disk will not be affected.
Impact on snapshots	Snapshots created based on system disks can be used to roll back cloud disks. Both manual and automatic snapshots created based on the system disk will be retained. The automatic snapshot policy settings are still valid and do not need to be reset.	Snapshots created based on the old system disk cannot be used to roll back the new system disk, but can be used to create custom images. Snapshots created manually based on old disks will not be released. Automatically created snapshot, if the old system disk is enabled Auto Snapshot Released with Cloud Disk , the automatic snapshot will be deleted automatically; If not turned on Auto Snapshot Released with Cloud Disk , it will be automatically released upon expiration. The automatic snapshot policy of the old system disk will become invalid and needs to be reset.
Charging Case	There is no charge for initializing the system disk. The operating system remains unchanged before and after initialization, and the billing items do not change.	There is no charge for replacing the operating system itself, and the following cases will charge a fee: If the replaced image is a charged image, the image fee will be charged. For more information, see Image billing 。 If the cloud disk capacity of the system disk is expanded during the replacement, a fee will be charged for the new capacity. For more information, see Block storage charging 。

Will my snapshot be lost when I replace the system disk?

The manual snapshot of the historical system disk is still retained. If the automatic snapshot is set, the automatic snapshot will be deleted if it is released with the disk.

explain

The cloud disk ID will change after the system disk is replaced, and the snapshot of the historical system disk cannot be used to roll back the new system disk.

What should I pay attention to when replacing the system disk?

Before replacing the system disk, it is recommended that you create a snapshot backup data, and you need to ensure that the system disk has more than 1GB of reserved space. Otherwise, the instance may not start normally after the system disk is replaced.

For the system disk price, see Block Storage Price 。

How to expand the system disk?

You can use the console or call ResizeDisk Expand the system disk.

Does cloud disk support resizing after resizing?

I won't support it. The new capacity of the cloud disk after expansion cannot be reduced after it takes effect. It is recommended that you plan the storage space reasonably. If you need to shrink the cloud disk, you can do so through AliCloud Server Migration Center SMC. See Cloud Disk Shrinkage 。

Does all block storage support system disk expansion? Are there any geographical restrictions?

ESSD AutoPL cloud disk ESSD PL-X cloud disk Efficient cloud disk, SSD cloud disk, and ESSD cloud disk support system disk expansion. All regions support system disk expansion.

Do subscription and pay as you go ECS instances support system disk expansion?

Both subscription and pay as you go ECS instances support system disk expansion.

What is the capacity range of the system disk? What is the maximum capacity?

The capacity range varies with the image used by the instance. For more information, see Cloud Disk Expansion Guidelines 。

After ECS renewal and configuration change, does it not support specifying the system disk capacity when replacing the system disk?

After the subscription ECS instance completes the renewal and configuration reduction operation, the system disk can only be expanded after entering a new billing cycle. The current billing cycle does not support system disk expansion.

How to use snapshots to create cloud disks to achieve lossless expansion of data disks?

When expanding a data disk, if the cloud disk cannot be expanded losslessly due to cloud disk reasons, you can temporarily create a pay as you go cloud disk to store data, and then completely format the data disk. See the following steps.

Create a snapshot of the current source data disk. See Create a cloud disk snapshot 。
go to Cloud Disk Purchase Page , select the same region and availability zone as the ECS instance, and select Create disk with snapshot , select the snapshot generated in the previous step.
On the ECS console, mount the new data disk to the same ECS instance.
Log in to the ECS instance and run mount Command to mount the new cloud disk. For how to mount the cloud disk created from the snapshot, see Create a cloud disk using snapshots 。
Spot check to see whether the files in the newly created data disk are consistent with the source data disk.
After confirming the data consistency, use fdisk Delete the original partition table and then use it fdisk 、 mkfs.ext3 Wait for the command to perform the partition formatting operation on the data disk again, so that the free space of the new data disk becomes the size after expansion. See Offline capacity expansion partition and file system (Linux data disk) 。
function cp -R The command copies all the data in the newly created data disk back to the source data disk.
You can add --preserve=all Parameter to retain the attributes of the file.
function umount Command to unload the new data disk.
Uninstall and release the new data disk from the instance on the ECS console.

Why is Bad magic number in super block while trying to open/dev/vdb1 prompted by ECS instance expansion cloud disk of Linux system?

Problem phenomenon: e2fsck -f /dev/vdb Prompt error when resizing and formatting cloud disk Bad magic number in super-block while trying to open /dev/vdb1
Cause of the problem: The starting sector is incorrect. For example, through fdisk -lu The command shows that the operation record shows that the starting sector position before resizing is 63, while the default position used when rebuilding partitions is 2048.
Solution:
1. function fdisk -lu /dev/vdb Command to view the starting position of the partition and record it.
2. After creating a new partition, run fdisk /dev/vdb Command to modify the starting position of the partition, and then run r Command to return to the main menu, and then run wq Save the command and exit.
3. Run separately e2fsck -f /dev/vdb and resize2fs /dev/vdb Command resizing.
4. function mount Command to remount the cloud disk.

What if you do not want to restart the instance after offline capacity expansion?

If you have done offline capacity expansion but do not want to restart the instance, you can make the previous offline capacity expansion effect effective by online capacity expansion again while maintaining the normal operation of the instance. However, there will be a certain amount of expense for the second expansion. Please complete the expansion according to the actual business needs.

For example, you have expanded the cloud disk offline to 60 GiB on the ECS console, but at this time, because of business needs and do not want to restart the instance, the cloud disk expansion has not actually taken effect. Then you can perform online capacity expansion on the ECS console again, such as online capacity expansion to 61 GiB. After the expansion, you can see that the disk capacity in the operating system has changed to 61 GiB.

Can the data disk be subdivided into a partition for data storage?

support. You can divide the data disk into multiple partitions as required. It is recommended that you use the tools provided with the system to partition.

For a cloud disk with multiple partitions, is the snapshot created for a single partition or all partitions?

All partitions. The snapshot backs up the data of the whole cloud disk, rather than some partitions of the cloud disk.

What are the precautions for cloud disk secondary partitions?

For the sake of data security, you need to create a snapshot to back up the cloud disk data in advance, and you can roll back the cloud disk in time when an operation error occurs. See Create Snapshot and Rollback cloud disk using snapshot 。

What is the relationship between writing data and partition formatting?

A new cloud disk or new cloud disk partition must be initialized and the data structure recorded on the cloud disk before normal use. The purpose of formatting is to establish a file system. Therefore, establishing a file system on a cloud disk means that there are write operations on the cloud disk. Different file systems write different file sizes when formatting:

Windows system formatting (divided into fast and normal formatting):
- Quick format: just allocate the file system for the partition and rewrite the directory table. Quick formatting takes up relatively little real space.
- Normal format: not only completes the quick format, but also scans the partition sector by sector to identify and mark bad sectors, and fills in empty blocks of the cloud disk, which is equivalent to the amount of data written to the entire cloud disk. At this point, the first full snapshot will approximate the cloud disk capacity.
Linux system formatting: After the cloud disk is formatted, before the instance writes business data, the capacity of the first snapshot is related to the cloud disk file system format.

After the secondary partition of the data disk, how many partitions is the data disk after the snapshot is rolled back?

Snapshot rollback is to rollback to the snapshot time point. If the time point has not been partitioned for the second time, there will be no multiple partitions.

Why does the error "The instance needs to be stopped to rollback the disk, and ensure that the current disk has not created a snapshot or replaced the operating system" report when rolling back the cloud disk?

Problem phenomenon: When rolling back the cloud disk, an error is reported, "To execute the rollback of the cloud disk, you need to stop the instance, and ensure that the current cloud disk has not created a snapshot or changed the operating system."
Cause of the problem: It is usually caused by the cloud disk attribute or cloud disk status inconsistency.
Solution: You can troubleshoot according to instance status and related snapshot tasks.
- Check whether the instance has stopped normally.
  Only stopped instances can perform cloud disk rollback operations. You can view the instance status on the instance list page of the console.
- Check whether the instance associated with the relevant snapshot has replaced the system disk.
  If the operating system is replaced, the system disk of the instance will be automatically recreated with the new image, and the system disk ID will change. Therefore, the original snapshot of the instance can no longer be used for rollback. However, you can create a custom image based on the relevant snapshot, and then specify the custom image by replacing the system disk to switch the instance to the corresponding snapshot state. See Create a custom image using snapshots and Replace the operating system (non-public image) 。
- Check whether the cloud disk to which the relevant snapshot belongs is creating a snapshot.
  To ensure data consistency, if there is a snapshot task in progress on the cloud disk to which the snapshot belongs, the rollback operation is not allowed. You can visit Instance Details Of page snapshot In the tab, check whether it exists speed of progress no 100% ， state be in hang in the air Snapshot task for.
  If you need to forcibly terminate the relevant snapshot task to roll back the cloud disk as soon as possible, you can check the corresponding snapshot and click Delete Snapshot Terminate the snapshot task.

How to transfer the data in the system disk of the Linux instance?

When purchasing ECS of Linux system, the data disk is not attached. After a period of use, the space utilization rate of the system disk is close to 100%, which cannot meet the business requirements. At this time, you can purchase and attach a new data disk, and use mv Command to transfer the data in the system disk to the newly purchased data disk.

How to copy data across ECS instances?

According to different operating systems, you can refer to the following methods to copy data across ECS instances:

Data copy between Linux instances
- Lrzsz mode
  Log in to the instance, install lrzsz, and then execute rz Upload, sz Download.
  You can also use it first sz Download to local, and then rz Upload to another instance.
- FTP mode
  When using the sftp tool, it is recommended to log in with the root account for downloading and uploading.
- Wget command
  Compress the file or folder and put it in the Web directory to generate a download URL, which can be used on another instance wget Command to download.
- SCP mode
  Log in to the instance and use the SCP+source SSH account password to directly copy data remotely. For details, see the Linux scp command.
Data copy between Linux and Windows
It is recommended to use sftp on the Linux instance, download it locally, and then upload it to the Windows instance using FTP.
Data copy between Windows
- How to use FTP
  For a tutorial on installing FTP services on Windows, see Manually set up an FTP site (Windows) 。
- Both instances log in to Alibaba Wangwang for transmission

Problems related to the previous generation cloud disk local SSD disk

The local SSD disk has stopped selling. If you are still using a local SSD disk, see Introduction to Local SSD Disk and FAQ about local SSD disk 。