FAQ about Elastic Block Storage - Elastic Compute Service

FAQ about Enhanced SSDs (ESSDs)
Common FAQ
FAQ about performance testing
FAQ about standard SSDs
FAQ about attaching and detaching disks
FAQ about independent disks
FAQ about disk snapshots
- When I delete a disk, will its snapshots also be deleted?
- Why are some automatic snapshots on my disk missing?
FAQ about re-initializing disks
FAQ about resizing disks
FAQ about partitions
FAQ about rolling back disks
- I rolled back a data disk by using a snapshot after I re-partitioned the disk. How many partitions are available in the disk?
- An error message similar to the following one is returned when I want to roll back a disk: "A disk can be rolled back only when the instance to which the disk is attached has been stopped and the disk has no snapshots being created. If the operating system of the current ECS instance has been replaced, the snapshot taken before the operating system is replaced cannot be used to roll back the new system disk." What do I do?
Other issues
- How do I migrate data from the system disk of a Linux instance?
- How do I copy data across instances?
FAQ about low disk space
- How do I resolve the issue of insufficient disk space in a Windows instance?
- How do I resolve the issue of insufficient disk space in a Linux instance?

What specifications do ESSDs have?

ESSDs offer varying specifications corresponding to their performance levels. For the most up-to-date performance specifications of ESSDs, please refer to ESSD.

The performance of a storage device is closely related to the capacity of the device. A storage device that has a larger capacity provides higher data processing capabilities. All ESSDs have the same I/O performance per unit of capacity. However, the performance of ESSDs increases linearly with its capacity until the maximum performance per disk at the PL is reached.

Performance level	ESSD capacity range (GiB)	Maximum IOPS	Maximum I/O throughput (MB/s)
PL0	1~65,536	10,000	180
PL1	20~65,536	50,000	350
PL2	461~65,536	100,000	750
PL3	1,261~65,536	1,000,000	4,000

What are the similarities and differences between ESSDs, standard SSDs, and ultra disks?

Similarities: ESSDs, standard SSDs, and ultra disks all leverage a distributed Elastic Block Storage architecture, ensuring high reliability and scalability. Additionally, they support snapshot creation and data encryption.
Differences: Of the three, ESSDs offer superior performance. For more information, see ESSD and EBS performance.

How are ESSDs billed?

You can select from two billing methods: subscription or pay-as-you-go. For detailed pricing information, visit the or the ECS pricing page.

How is the performance level of an ESSD measured?

The performance level of an ESSD is directly linked to its storage capacity; larger capacity ESSDs offer enhanced performance. ESSDs outperform standard SSDs in terms of performance. For more information, see ESSD.

What is the relationship between the storage performance of an ESSD and the storage performance of the instance to which the ESSD is attached?

The storage I/O performance of certain instances correlates directly with the specifications of their instance types. For instance, the storage I/O performance of the g7se storage-enhanced general-purpose instance family scales with the specifications of the instance types. The more advanced the specifications of an instance type, the greater the IOPS and throughput it can achieve.

When you create an instance using the g7se instance type and attach ESSDs to it:

If the combined performance of the ESSDs does not surpass the storage I/O performance of the instance type, then the ESSDs' performance takes precedence.
If the overall performance of ESSDs surpasses that of the instance type's storage I/O capabilities, the latter will take precedence.
If the combined performance of the ESSDs exceeds that of the instance type, the instance type's storage I/O performance becomes the limiting factor. For example, if you create a 16 GiB instance of the ecs.g7se.xlarge instance type with a maximum of 60,000 IOPS, and you attach an ESSD offering 2 TiB of storage and up to 101,800 IOPS, the instance's maximum IOPS will be capped at 60,000, not 101,800.

For more information about the specifications and performance of the g7se storage-enhanced general-purpose instance family, see instance family.

What tools can I use to test the performance of EBS devices?

You can use the open-source tool fio to assess key performance indicators of raw disks, such as IOPS, throughput, and latency. For more information, see testing the performance of EBS devices.

Why does my instance go down when I use fio to test the I/O performance of the instance?

Using fio to test the I/O performance of an instance can be done on either raw disk partitions or file systems. Testing on raw disk partitions may corrupt the file system metadata, leading to inaccessible files within those partitions and potentially causing the instance to crash. However, this issue does not arise when testing is conducted on file systems using fio.

How do I test the performance of an ESSD?

You can utilize the flexible IO tester (fio) to conduct a stress test on an ESSD. For more information, see testing the IOPS performance of an ESSD.

How do I select a zone for a disk based on application requirements?

A pay-as-you-go disk can only be attached to an ECS instance within the same zone.

For high-availability applications, it is recommended to create data disks in multiple zones and attach them to ECS instances located in those respective zones.
For applications requiring low latency, create data disks in the same zone as your ECS instances and attach them accordingly.

How do I query the usage and free space of EBS devices?

You can sign in to an ECS instance to check the usage and available space on EBS devices. However, it's not possible to obtain this information through the ECS console or by invoking ECS API operations.

Can I shrink the capacity of a disk ?

Alibaba Cloud does not support the reduction of disk capacity. However, Alibaba Cloud offers the following solutions:

Solution 1: Use Alibaba Cloud Server Migration Center (SMC)

You can reduce a disk's size using Alibaba Cloud Server Migration Center (SMC). For more information, see shrink a disk.
To ensure data security, it is recommended to create a snapshot of the disk before shrinking it. For more information, see create a snapshot.

Solution 2: Replace the data disk and migrate data

Purchase a new data disk with the desired capacity and attach it to your ECS instance. For more information, see create an empty data disk and attach a data disk.
Migrate data from the original data disk to the new one.
Switch the billing method for the original data disk from subscription to pay-as-you-go, then detach and release the disk. For more information, see detach a data disk and release a disk.

Important
Once released, a disk cannot be restored. Ensure all critical data is backed up before releasing the disk.
After converting a subscription disk to pay-as-you-go, Alibaba Cloud will refund the balance. For more details on the pay-as-you-go billing method, see EBS billing.

Carry out the above operations as per your specific requirements.

How do I release a subscription disk that has not expired?

You cannot release Alibaba Cloud subscription data disks before their expiration date. To release a subscription data disk, you must first convert its billing method to pay-as-you-go. Ensure that you have backed up all critical data on the disk before releasing the newly converted pay-as-you-go data disk. For more information, see Switch between billing methods and release a disk.

Note

Once you change a data disk's billing method from subscription to pay-as-you-go, the disk is billed hourly. Charges for the disk cease one hour after its release, and the refund amount will be displayed in the ECS console. Please note that used coupons are not eligible for a refund.

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

I/O optimization enhances network capabilities and storage performance for both instances and disks. It allows instances to fully leverage the storage performance of standard SSDs.

You can upgrade an existing ECS instance to an I/O optimized instance by using the ModifyInstanceSpec or ModifyPrepayInstanceSpec feature.

How does Alibaba Cloud deal with EBS resource contention?

Alibaba Cloud's Elastic Block Storage (EBS) is a multitenant storage service that sets performance benchmarks according to instance types and disk specifications. Disks come in various categories, including ESSDs, standard SSDs, and ultra disks, each with distinct performance levels. To accommodate your data storage requirements, you can select from these disk categories, along with resize or reclassify disks to adjust their performance capabilities.

To monitor disk performance and gather relevant data, you can utilize CloudMonitor or EBS Lens. This data helps you assess whether the disk performance aligns with your business needs. For additional details on CloudMonitor and EBS Lens, see cloud service monitoring and EBS Lens.

To mitigate resource contention, Alibaba Cloud strategically allocates infrastructure resources to ensure consistent storage availability. In the event of hardware malfunctions or resource conflicts, Alibaba Cloud promptly reallocates resources or conducts repairs, depending on the severity of the infrastructure damage.

What is the I/O performance of a standard SSD?

For more information, please refer to EBS Performance.

What scenarios are standard SSDs ideal for?

Standard SSDs deliver robust performance and reliability, making them perfectly suited for I/O-intensive applications that demand high data reliability, including relational databases like MySQL, SQL Server, Oracle, and PostgreSQL. They are also the preferred choice for small to medium-sized development and testing environments where data reliability is a critical requirement.

Can I replace a basic disk with a standard SSD?

No, you cannot replace a basic disk with a standard SSD, as the latter utilizes SSDs for its physical storage medium.

How do I purchase a standard SSD? What are the pricing options for I/O optimized instances and standard SSDs?

For detailed pricing information on standard SSDs and I/O optimized instances, please visit the pricing details page.

Can I upgrade a standard SSD after I purchase it?

Yes, you can upgrade or resize your standard SSD. For more details, see the guidelines for resizing disks.

Why is an error returned when I attempt to mount the partitions of a standard SSD to an I/O optimized Linux instance?

In the Linux operating system, standard SSD mount points follow the /dev/vd* format, while basic disks use the /dev/xvd* format. To mount a standard SSD partition successfully, ensure you use the /dev/vd* format in your mount command.

What must I be aware of before I add mount information about basic disks or standard SSDs to Linux instances?

Before attaching and preparing a data disk for a Linux instance, note that device names differ between I/O optimized and non-I/O optimized instances. By default, system-assigned device names follow these conventions:

I/O optimized instances:
- For data disks using the NVMe protocol, device names adhere to the /dev/nvmeXn1 format, such as /dev/nvme1n1, /dev/nvme2n1, and /dev/nvme3n1. For details on NVMe protocol-supported disks, see overview of NVMe disks.
- For data disks not using the NVMe protocol, device names are in the /dev/vd[b-z] format, for instance, /dev/vdb, /dev/vdc, and /dev/vdd.
Non-I/O optimized instances: Device names for data disks are in the /dev/xvd[b-z] format, such as /dev/xvdb, /dev/xvdc, and /dev/xvdd.

Important

It is recommended to use UUIDs to reference new partitions in the /etc/fstab file, as disk device names may change upon release. Specifying device names in the /etc/fstab file could lead to data storage issues due to these changes.

Adding incorrect mount information prevents the mount -a command from mounting the disk. Resolve this by:

Using the fdisk -l command to review data disk details.
Verifying the validity of the /etc/fstab file contents.

Note
Avoid adding duplicate mount information to prevent system startup issues.
Use the vim command to edit the /etc/fstab file.
Remove or comment out invalid entries and insert correct mount information.
Test if the disk mounts successfully with the mount -a command.

For the standard initialization procedure, refer to initialize a data disk (Linux).

What is a device name (mount point)?

A device name, also known as a mount point, refers to the specific location of an ECS disk on the disk controller bus. In Linux, the device name corresponds to the disk device number, while in Windows, it aligns with the disk sequence number within the disk manager.

What is an independent disk?

An independent disk is a pay-as-you-go data storage device that you create separately, rather than one that is automatically generated alongside an ECS instance. This type of disk can be attached to or detached from an ECS instance within the same zone. However, it cannot be connected to an ECS instance in a different zone. Before utilizing an independent disk, you must attach it to an instance and then partition and format it. For more information, see create an empty data disk.

Can I attach a disk to multiple ECS instances?

No, a disk can only be attached to one ECS instance within the same zone.

Do I need to partition and format a pay-as-you-go disk after I purchase and attach it to an ECS instance?

After purchasing and attaching a pay-as-you-go disk to your ECS instance, you must partition and format the disk before it can be utilized. For detailed instructions, see initialize a data disk (Linux) or initialize a data disk (Windows).

Why am I unable to find the data disk that I purchased for an ECS instance?

Linux operating system

If you have purchased a pay-as-you-go data disk separately for a Linux instance, you need to attach the disk to the instance and then partition and format it before the storage space becomes visible and usable. For more information, see initialize a data disk (Linux) and attach a data disk.
Windows operating system

For details on potential causes and their solutions, see troubleshoot the issue of data disks that cannot be found in a Windows instance.

How many disks can be attached to one ECS instance?

An ECS instance can support up to 64 data disks. The specific limit on the number of attachable disks depends on the instance type. For more information, see instance family.

Note

When creating an instance, you can attach one system disk and up to 16 data disks. Should you need additional data disks, they can be attached after the instance has been created. The maximum number of attachable disks is determined by the instance type. To find out the maximum number of disks your instance type can support, use the DescribeInstanceTypes operation.

Why am I unable to find the desired ECS instance when I attempt to attach a disk to it?

Verify if the ECS instance has been released. If it hasn't, ensure that the disk and the ECS instance are in the same zone.

Can I attach a disk to an ECS instance that resides within a different zone?

No, a pay-as-you-go disk can only be attached to an ECS instance that is in the same zone.

Will data in a data disk be lost when I detach the disk?

For Windows systems, it's recommended to halt all read and write operations on the disk's file systems to maintain data integrity. Failing to do so may result in the loss of any data currently being read or written at the time of detachment.
On Linux systems, you should log on to the instance and execute the umount command to unmount the disk. Once the disk is unmounted, you can proceed to detach it via the ECS console.

Can I detach system disks?

Yes, you can detach system disks. For more information, see how to detach or attach a system disk.

How is a separately created pay-as-you-go data disk billed?

Pay-as-you-go data disks are billed hourly. Should your account balance be insufficient, services associated with these data disks will be suspended.

I attached a separately created disk to an ECS instance. Why is the disk released when the instance is released?

This occurs because the disk is set to release simultaneously with the instance. To modify this setting, you can use the ECS console or invoke an API operation. For more information, see release a disk.

Can I attach a separately created pay-as-you-go data disk to a subscription instance?

Yes, you can attach a pay-as-you-go data disk that was created independently to a subscription instance.

Can I detach data disks from subscription instances?

No, you cannot detach data disks from subscription instances. These data disks expire concurrently with the subscription instances they are attached to and are released simultaneously. To release a data disk, you can convert its billing method from subscription to pay-as-you-go, then detach and release it. For more information on how to change a disk's billing method, see Switch between billing methods.

I changed the configurations of a disk when I renewed the disk. Can I change the billing method of the disk from subscription to pay-as-you-go within the remaining time of the current subscription period?

No, you can only switch the disk's billing method from subscription to pay-as-you-go by altering the instance configurations once the current subscription period has concluded.

When I delete a disk, will its snapshots also be deleted?

Manual snapshots will be retained.
The retention of automatic snapshots depends on the setting of the Delete Automatic Snapshots While Releasing Disk option. For more information, see delete automatic snapshots while releasing disk.

Why are some automatic snapshots on my disk missing?

If the snapshot count reaches its limit, the oldest automatic snapshots will be deleted automatically, while manual snapshots remain unaffected.

Note

An automatic snapshot policy assigned to a disk is only executed after the disk has been attached to an instance.

I cannot access the data in a Linux data disk because an error occurred when I attached the disk. What do I do?

If you're unable to access data on a Linux data disk, try the following troubleshooting steps:

Locate the data disk and verify that it is properly attached to the corresponding ECS instance.
- Check the disk status in the ECS console. For more information, see viewing disk monitoring information.
- Log on to the instance and execute the fdisk -l command to verify the partition information of the data disk. Use the df -h and mount | grep "<devpath>" commands to check the mounting details.
Use the cat command to inspect the /etc/fstab file and confirm that no two disks are mounted to the same directory.
- If you find that two disks are mounted to the same directory, the second disk will overshadow the first, rendering its data inaccessible. It's advisable to mount one disk to a different directory.
- If the disks appear to be mounted on separate directories but the mount information suggests otherwise, execute the ll command to check for any links between the two directories. Should there be a link, create a new directory using the mkdir command and mount the disk there to see if the data becomes accessible.

What do I do if data is lost after I restart a Linux instance?

Problem description: Data in a specific directory (e.g., /alidata) is missing after the instance is restarted.
Possible cause: Use the df -h command to verify if the data disk partition is mounted.
Solution: The following steps use the /dev/vdb1 partition on a data disk connected to a non-NVMe I/O optimized instance as an example. For NVMe I/O optimized instances, the device name is /dev/nvmeXn1. For non-I/O optimized instances, it is /dev/xvd[b-z]. Follow these steps according to your specific situation:
1. Execute the fdisk -l command to identify the unmounted data disk partition.
2. Use the mount /dev/vdb1 /alidata command to mount the data disk partition.
3. Verify the mount status with the df -h command.
4. (Optional) To ensure the data disk partition mounts automatically at startup, configure the /etc/fstab file, which can help prevent future data loss.

If I re-initialize a disk, are my snapshots retained?

Yes, all manual and automatic snapshots of the disk are preserved.

What do I do if data is lost after I restart a Linux instance or re-initialize the system disk?

Problem description: After restarting a Linux instance or re-initializing the system disk, the data disks may not appear when you log on to the instance and execute the df -h command to view the mount information.
Cause:
- For instance restarts: If the mount information is not saved in the /etc/fstab file before restarting, the data disks will not mount automatically post-restart.
- For system disk re-initialization: The /etc/fstab file is reset, leading to the data disks not mounting automatically after the system disk is re-initialized.
Solution:

Here, we use the /dev/xvdb1 partition of a non-I/O optimized instance as an example. For non-NVMe I/O optimized instances, the device name is /dev/vd[b-z]. For NVMe I/O optimized instances, it is /dev/nvmeXn1. Follow these steps based on your scenario:
1. Execute the mount /dev/xvdb1 command to reattach the data disk.
2. Use the mount command to verify the file format of the /dev/xvdb1 partition.
3. Assuming the /dev/xvdb1 partition is formatted with the ext3 file system, execute the following command to add the data disk's mount information to the /etc/fstab file:
```
echo '/dev/xvdb1 /data ext3 defaults 0 0' >> /etc/fstab
```
4. Restart the ECS instance through the ECS console.

How do I re-attach data disks after I re-initialize the system disk of a Linux instance?

When you re-initialize the system disk of a Linux instance, the data on the data disks remains intact, but the mount information is lost. For instance, if the data disk partition was previously named /dev/vdb1 with a mount point at /InitTest, follow these steps to establish new mount information and re-attach the data disk partition after rebooting the Linux instance:

You can use the mount command to verify the data disk's mounting details.

The returned result lacks information regarding /dev/vdb1.
Check the partition details of the data disk with the fdisk -l command.
Execute the cat /etc/fstab command to verify the original mount point for the /dev/vdb1 partition.
You can use the mkdir /InitTest command to create a new mount point for the data disk partition.

The new mount point must be named identically to the original mount point for the /dev/vdb1 partition prior to reinitializing the system disk.
Mount the data disk partition to the new mount point using the mount /dev/vdb1 /InitTest command.
Confirm the mount was successful by running the df -h command.
To ensure the /dev/vdb1 data disk mounts automatically, execute the following:
1. You can execute the umount /dev/vdb1 command to unmount the /dev/vdb1 data disk.
2. You can use the mount command to verify the mounting details.
  
  If the data disk is detached, the returned result will not include information about /dev/vdb1.
3. You can execute the mount -a command to automatically mount the /dev/vdb1 data disk.
4. You can use the mount command to verify the mounting details.
  
  If the data disk mounts automatically, the returned result will include information about /dev/vdb1.

What are the differences between a system disk re-initialization operation and a system disk replacement operation?

The following table describes the differences.

Difference	System disk re-initialization	System disk replacement
Feature	The instance reverts to its original state at creation, with the operating system remaining the same before and after initialization.	The operating system of the instance changes.
Impact on the system disk	The original system disk is retained. The system disk is reset to its original state at creation, erasing any applications installed and data generated thereafter. Data backup is recommended. The system disk ID, type, and instance IP addresses stay the same.	The original system disk is released. All data on the original system disk is erased. Data backup is recommended. The system disk ID changes, but the disk type, instance IP addresses, and MAC addresses of attached ENIs remain the same.
Impact on data disks	Data disks remain unaffected.	Data disks remain unaffected.
Impact on snapshots	Snapshots can be used to revert the system disk to a previous state. Both manual and automatic snapshots of the system disk are preserved. If an automatic snapshot policy is in place, it continues to apply to the disk.	The snapshots of the original system disk cannot be used to roll back the new system disk, but can be used to create custom images. Manual snapshots of the original system disk are not released. If you enable Delete Automatic Snapshots While Releasing Disk for the original system disk, the automatic snapshots of the original system disk are automatically released together with the disk. If you disable Delete Automatic Snapshots While Releasing Disk for the original system disk, the automatic snapshots of the original system disk are retained until the snapshots expire. The automatic snapshot policy applied to the original system disk becomes invalid, and you can configure an automatic snapshot policy for the new system disk based on your business requirements.
Charges	There is no charge for disk initialization, and billing items remain unchanged as the operating system is unaffected.	You are not charged for the operating system replacement operation, but you are charged for the resources used in the following scenarios: If you use a paid image as the replacement image, you are charged for the image. For more information, see Images. If you extend the system disk when you replace the operating system of an ECS instance, you are charged for the additional disk capacity. For more information, see Block storage devices.

Are my snapshots retained if I replace a system disk?

Manual snapshots of the original system disk will be preserved. However, automatic snapshots will be deleted if the "Delete Automatic Snapshots While Releasing Disk" feature is activated.

Note

Note that after replacing a system disk, the disk ID will change, making it impossible to use the original system disk's snapshots to roll back the new system disk.

What must I be aware of before I replace a system disk?

Before replacing a system disk, it is recommended to create snapshots for backup purposes. Ensure that the new system disk has at least 1 GiB of free space. Without adequate free space, the instance may fail to start properly after the system disk replacement.

For details on system disk pricing, visit the or the ECS pricing page.

How do I resize a system disk?

You can resize the system disk either through the ECS console or by invoking the ResizeDisk operation.

Can I shrink a disk after I resize it by extending its capacity?

No, once you have increased a disk's capacity, it is not possible to reduce its size. After the new, larger capacity is in effect, the disk cannot be downsized. It is advisable to carefully plan your storage needs in advance. To reduce a disk's size, you can utilize the Alibaba Cloud Server Migration Center (SMC). For more information, see how to shrink a disk.

What EBS devices can be resized when they are used as system disks? Do regional limits apply to this operation?

System disks such as ESSD AutoPL, ultra disks, standard SSDs, and ESSDs can be resized. This operation is available in all regions.

Can the system disks of both subscription and pay-as-you-go ECS instances be resized?

Yes, you can resize the system disks for both subscription and pay-as-you-go ECS instances.

What is the storage capacity range of a system disk?

The storage capacity of a system disk depends on the operating system used. For detailed guidelines on disk resizing, see the referenced document.

Can I specify a new size for the system disk when I replace the system disk?

You cannot resize the system disk until the start of a new billing cycle following the downgrade of a subscription instance's configurations upon renewal.

How do I create a disk from a snapshot of a data disk to resize the data disk without data loss?

If you cannot resize a data disk due to a disk error without risking data loss, consider purchasing a pay-as-you-go disk to temporarily store the data from the original disk before reformatting it. Follow these steps:

Create a snapshot of the current data disk. For more information, see create a snapshot.
Visit the disk purchase page, select the same region and zone as your ECS instance, and click Create Disk From Snapshot. Choose the snapshot you created earlier.
Log on to the ECS console and attach the newly purchased data disk to your ECS instance.
Log on to the ECS instance and execute the mount command to attach the new disk. For guidance on attaching a disk created from a snapshot, see create a data disk from a snapshot.
Verify that the files on the new data disk match those on the original disk.
Once you've confirmed data consistency, use the fdisk command to remove the original partition table. Then, partition and format the original data disk with the fdisk and mkfs.ext3 commands, ensuring the new disk's free space matches its new capacity. For details, see resize the partition and file system of a data disk in legacy Linux.
Use the cp -R command to copy all data from the new data disk to the original one.

You can use the --preserve=all parameter to maintain all file attributes.
Detach the new data disk by running the umount command.
Finally, detach and release the new data disk from the ECS instance through the ECS console.

What do I do if the error message "Bad magic number in super-block while trying to open /dev/vdb1" appears when I resize a disk of a Linux instance?

Problem description: When attempting to resize and format the disk using the e2fsck -f /dev/vdb command, the error message Bad magic number in super-block while trying to open /dev/vdb1 is displayed.
Cause: This error typically indicates an invalid start sector for the disk partition. For instance, using the fdisk -lu command may show a start sector of 63 before resizing, which changes to 2048 after re-creating the partition.
Solution:
1. Check the partition's start sector by running the fdisk -lu /dev/vdb command and note the value.
2. Create a new partition and adjust the start sector with the fdisk /dev/vdb command. Use the r command to return to the main menu, then save and exit with wq.
3. Execute the e2fsck -f /dev/vdb and resize2fs /dev/vdb commands to resize the disk.
4. Reattach the disk by running the mount command.

What do I do if I have resized a disk offline but do not want to restart the instance?

If you've resized a disk offline and prefer not to restart the instance, you can perform an online disk resize while the instance is running to apply the changes. Note that online resizing incurs charges. Adjust the disk size according to your business needs.

For instance, if you've increased an instance's disk size to 60 GiB through the ECS console without restarting, the change won't be effective. To implement the resizing, you can adjust the disk size online via the ECS console, perhaps to 61 GiB. Following the online resize, the operating system will reflect the increased disk capacity of 61 GiB.

Can I partition a data disk for data storage?

Yes, you can partition a data disk into multiple sections for storage. It is advisable to utilize the system's partitioning tool for this purpose.

For a disk with multiple partitions, are snapshots created for the entire disk or only for a specific partition?

Snapshots capture the entire disk, not just individual partitions.

What must I be aware of before I re-partition a disk?

To safeguard your data, it's advisable to create a snapshot of the disk prior to re-partitioning. Should an error occur, this allows you to revert the disk to its previous state. For more information, see creating a snapshot and reverting a disk using a snapshot.

What is the relationship between data writing and partitioning and formatting?

To utilize a new disk or disk partition, it must first be initialized and have a data structure established on it. Formatting's primary purpose is to create file systems. As such, creating a file system on a disk involves writing the file system's data onto the disk. The volume of data written to disks during the formatting process depends on the chosen file system:

For a Windows instance, you can format a data disk using one of two approaches:
- Quick formatting: This method quickly allocates file systems to partitions and rewrites the directory table, using less space than full formatting.
- Full formatting: This comprehensive method allocates file systems to partitions, rewrites the directory table, and checks for bad sectors, marking them accordingly. It also fills empty data blocks on the disk, effectively writing data to the entire disk. Consequently, the size of the initial full snapshot is roughly equivalent to the disk's capacity.
In a Linux instance, the size of the initial snapshot is influenced by the file system format if the disk remains unwritten post-formatting.

I rolled back a data disk by using a snapshot after I re-partitioned the disk. How many partitions are available in the disk?

Rolling back a data disk to a snapshot reverts the disk to its state at the time the snapshot was captured. If the disk had not been re-partitioned at that time, the disk will contain only one partition.

An error message similar to the following one is returned when I want to roll back a disk: "A disk can be rolled back only when the instance to which the disk is attached has been stopped and the disk has no snapshots being created. If the operating system of the current ECS instance has been replaced, the snapshot taken before the operating system is replaced cannot be used to roll back the new system disk." What do I do?

Problem Description: Attempting to roll back a disk using a snapshot may result in the error message: "A disk can only be rolled back when the instance to which it is attached has been stopped, and there are no snapshots being created. If the operating system of the current instance has been replaced, the snapshot taken before the replacement cannot be used to roll back the new system disk."
Cause: This issue could stem from an invalid disk attribute or disk state.
Solution: Address this issue by checking the instance state or snapshot state.
- Ensure that the instance associated with the disk is in a stopped state.
  Ensure the instance to which the disk is attached is stopped. Disks can only be rolled back when their associated instance is in the Stopped state. To verify the instance's state, log on to the ECS console and visit the Instances page.
- Verify if the snapshot-associated instance has had its system disk replaced.
  Verify if the system disk associated with the snapshot has been replaced. When the operating system of an instance is replaced, a new system disk is created using the new image, resulting in a different system disk ID. Snapshots from the original system disk cannot be used to roll back the new one. Instead, create a custom image from the snapshot and use it to replace the instance's system disk, restoring it to the state captured in the snapshot. For more information, see create a custom image from a snapshot and replace the operating system (non-public image).
- Verify that there are no snapshots currently being created for the disk you want to roll back.
  Check if there is an ongoing snapshot creation for the disk you want to roll back. Alibaba Cloud prohibits disk rollbacks during snapshot creation to maintain data consistency. On the Instance Details page, click the Snapshots tab and ensure that the Progress is not at 100% and the Status shows as Incomplete.
  
  If you need to cancel the snapshot creation to proceed with the rollback, select the snapshot and click Delete Snapshot to stop the snapshot task.

How do I migrate data from the system disk of a Linux instance?

Suppose you have a Linux instance without additional data disks, and over time, the system disk's usage nears 100%, no longer sufficing for your business requirements. In this scenario, you can add a data disk to the instance and use the mv command to transfer data from the system disk to the new data disk.

How do I copy data across instances?

To copy data across instances, depending on the operating system, you can use one of the following methods:

Copy data between Linux instances:
- Utilize the lrzsz tool:
  
  Log on to the instance and install lrzsz. You can then use the rz command to upload data and the sz command to download data.
  
  You can use the sz command to download data to your computer, and then employ the rz command to upload that data to a different instance.
- Employ the FTP service:
  
  For SFTP tool usage, it's recommended to log on to instances using the root account for file uploads and downloads.
- Execute the wget command:
  
  Compress the files or folders, place them in a web directory, and create a download URL. Then, use the wget command on another instance to download the files or folders.
- Utilize the secure copy (SCP) utility:
  
  Log on to the instance and use the SSH credentials of the source instance to execute SCP commands for copying data. For additional information on SCP commands in Linux, refer to the relevant documentation.
Copy data between a Linux instance and a Windows instance:

It is recommended to download files from the Linux instance to your computer using the SFTP tool and then upload them to the Windows instance using the FTP service.
Copy data between Windows instances:
- Use the FTP service:
- You can use TradeManager to transfer data between two instances.