Elastic Compute Service:instance FAQ

How do I choose an ECS instance that suits my business?

For detailed guidance, see best practices for instance selection.

How do I check whether an instance can be purchased in a specific region or zone?

You can go to the Instance Types Available for Each Region page to view the instance types available in each region.

What should I do if resources are sold out when I purchase an instance?

How long does it take to activate an ECS instance?

Why is no instance created after I successfully pay for an instance?

If the specified instance type is sold out in the chosen zone, your account will automatically be refunded. Should you not receive a refund within 30 minutes, please submit a ticket.

You can go to the Instance Types Available for Each Region page to view the instance types available in each region.

Why is the memory size different from the instance type definition after I purchase an instance?

The memory size specified for each instance type represents the total memory capacity, which accounts for the system's memory usage, including BIOS reserved memory, along with memory overhead from the kernel and hypervisor. Consequently, the available memory on your instance will be less than the defined size in the instance type. The system memory usage percentage can vary slightly across different instance families due to variations in their underlying technology stacks.

What are enterprise-level instances and entry-level instances?

Enterprise-level instances are part of a range of instance families launched by Alibaba Cloud in September 2016. They offer high performance, consistent computing power, and balanced network capabilities. Designed with exclusive and stable computing, storage, and networking resources, enterprise-level instances are ideal for businesses demanding high operational stability.

Entry-level instances, on the other hand, cater to individuals or small to medium-sized websites. Unlike enterprise-level instances, they prioritize resource sharing and offer variable computing performance. Nevertheless, entry-level instances are more cost-effective.

What are the essential differences between enterprise-level instances and entry-level instances?

Enterprise-level instances employ a CPU-bound scheduling scheme, where each vCPU is tied to a specific CPU hyperthread. This ensures that instances do not vie for CPU resources and consistently deliver the computing performance guaranteed by the service-level agreement (SLA).

In contrast, entry-level instances utilize a CPU-unbound scheduling scheme, with each vCPU being assigned to any available CPU hyperthread. As a result, vCPUs from different instances may compete for CPU resources, leading to variable computing performance during peak traffic periods. While shared instances are covered by an availability SLA, they lack a performance SLA.

Which instances are enterprise-level instances and which are entry-level instances?

Currently, the entry-level instances among the available instance families for sale include e, t6, t5, s6, n4, mn4, xn4, and e4. All others are classified as enterprise-level instances.

In which business scenarios do I need to purchase enterprise-level instances?

For detailed information on business scenarios that are best suited for various enterprise-level instances, see instance families and best practices for instance selection.

How is the network performance of enterprise-level instances?

Network performance for enterprise-level instances depends on their specific configurations. Generally, instances with higher specifications deliver better network performance. For detailed information on network performance across different instance configurations, see instance families.

What are the limits on changing the configurations of enterprise-level instances?

For details on the restrictions associated with modifying enterprise-level instance configurations, please refer to supported instance types for configuration changes.

Can entry-level instances be changed to enterprise-level instances?

For more information, see supported instance types for configuration changes.

Can I directly deploy Redis applications on persistent memory-optimized instances when persistent memory is used as memory?

Utilizing persistent memory-optimized instances for Redis applications can lead to substantial savings in total cost of ownership (TCO) per GiB of memory. However, to optimize performance, it's necessary to adapt your Redis applications to segregate data, placing less frequently accessed data in persistent memory and frequently accessed data in standard memory.

I have a Redis application cluster based on regular memory and want to migrate my business to persistent memory-optimized instances. What should I do and what should I pay attention to?

To ensure a smooth migration, it's crucial to maintain business continuity and data integrity. We suggest initially acquiring a persistent memory-optimized instance and conducting tests with a limited scope of your operations to verify if the instance's fundamental performance and capacity align with your business needs. Once confirmed, you can expand your deployment to additional persistent memory-optimized instances, which can then fully assume the responsibilities of your entire Redis cluster.

Can I directly deploy my parameter server (PS) on persistent memory-optimized instances when persistent memory is used as memory?

In a parameter server (PS) setup, server nodes traditionally store all the training parameters of the clusters in their memory. These parameters can consume a significant amount of memory, leading to high costs. By running a PS on persistent memory-optimized instances and utilizing persistent memory for parameter storage—while keeping hash tables in regular memory—you can substantially lower the total cost of ownership (TCO) for your training clusters.

You can adapt your applications to meet specific business needs or seek technical support from Alibaba Cloud. For technical assistance, or submit a ticket.

How do I configure persistent memory as local disks after purchasing a persistent memory-optimized instance?

You can configure persistent memory to function as local disks using specific tools. For more information, see configure persistent memory.

Which of my applications require higher performance local disks?

Can persistent memory be directly used to deploy Redis or MySQL applications when used as local disks? Is application modification required as with memory usage?

No modifications are necessary. When persistent memory is utilized as local disks, Redis and MySQL applications identify it as standard SSDs.

How does the performance of persistent memory used as local disks compare to local NVMe SSDs or cloud disks?

How reliable is persistent memory?

The reliability of data stored in persistent memory hinges on the dependability of both the persistent memory devices and the physical servers they are connected to. This scenario heightens the risk of encountering single points of failure. To ensure the dependability of your application data, we advise implementing data redundancy at the application layer and storing long-term business data on disks.

Furthermore, data within persistent memory is automatically erased when persistent memory-optimized instances are terminated. Therefore, we recommend backing up your data prior to instance termination. Clearing data from persistent memory takes additional time, resulting in a slightly extended release process for persistent memory-optimized instances compared to other instance types.

Which user groups and business scenarios are u1 instances intended for?

u1 instances are designed for small to medium-sized enterprises and cater to a wide range of general-purpose applications that do not demand high CPU performance. These include web applications, corporate websites, office software, offline data analysis, and databases of small to medium scale.

Which business scenarios are unsuitable for u1 instances?

• Scenarios demanding high computing power and performance for an optimal user experience, such as gaming and high-frequency trading, should utilize c7, g7, or r7 instances.

• Since u1 instances support multiple server platforms, which can lead to significant performance variations, scenarios requiring consistent business performance, like high-availability setups where primary and secondary servers must perform equally, are better served by c7, g7, or r7 instances.

What processor models do u1 instances support?

Currently, u1 instances support the Intel^® Xeon^® Platinum 6149 (Skylake), 8163 (Skylake), 8269CY (Cascade Lake), 8260 (Cascade Lake), 8369HB (Cooper Lake), and 8369HC (Cooper Lake) models.

Do u1 instances support specifying processor models?

No, they cannot.

What are the base frequency and turbo frequency of u1 instance processors?

U1 instances are compatible with server platforms offering varying levels of computing power. The base and turbo frequencies of U1 instance CPUs depend on the specific CPU models of the host platforms. Furthermore, as instances may transition between CPU platforms throughout their lifecycle, their actual base and turbo frequencies can vary as a result.

How to understand the migration of u1 instances to different server platforms during their lifecycle?

U1 instances may be transferred to new hosts throughout their lifecycle for reasons including host upgrades or maintenance activities. As a result, the processors in the new hosts might be different from those in the original ones. This type of host migration can occur with any instance during its lifecycle. For more information on the processor changes before and after migrating U1 instances, see universal instances (U instances).

Which business scenarios are not suitable for e instances?

Scenarios at the enterprise level that demand consistent performance and a guaranteed performance SLA.

What is the CPU model of e instances? Can you specify the processor model?

e instances are powered by Intel^® Xeon^® Platinum scalable processors. It is not possible to specify the processor model.

e instances operate on mainstream, cost-effective CPU platforms and may undergo migrations across different CPU platforms throughout their lifecycle. Technical measures ensure enhanced business compatibility across these platforms.

What are the base frequency and turbo frequency of e instances?

The base frequency stands at 2.5 GHz. Turbo frequency is not guaranteed as e instances may migrate across different CPU platforms throughout their lifecycle.

How to understand the migration of e instances to different server platforms during their lifecycle?

E instances may undergo migration to new hosts throughout their lifecycle, prompted by factors like host upgrades or maintenance activities. Consequently, the processors in the new hosts might vary from the original ones.

There's no need for you to monitor the underlying hosts. Alibaba Cloud employs resource pooling technology and an intelligent scheduling algorithm to dynamically manage resources. This ensures that your applications receive consistent computing power, stability, availability, and elasticity.

What is the essential difference between ECS bare metal instances, traditional cloud hosts (virtual machines), and traditional physical machines?

To understand the distinctions between ECS Bare Metal Instances, traditional virtual machines (cloud hosts), and physical servers, please refer to the ECS Bare Metal Instance specifications.

How is the network performance of ecs bare metal instances?

The network performance of ECS Bare Metal Instances varies according to their specifications. Lower-specification instances deliver reduced network performance, whereas higher-specification instances provide enhanced network capabilities. For detailed network performance metrics across different instance specifications, see instance families.

What disk types are supported by ecs bare metal instances? How many data disks can be mounted?

ECS Bare Metal Instances support ultra disks, standard SSDs, and enhanced SSDs (ESSDs). You can attach up to 16 data disks to a single ECS Bare Metal Instance.

Do ecs bare metal instances support configuration changes and failover?

ECS Bare Metal Instances do not support configuration changes. However, in the event of hardware failure, failover is possible, ensuring data retention on the disks.

How to use SCC RDMA features?

To utilize RDMA features in an SCC instance, choose an Alibaba Cloud Marketplace image tailored for SCC, which includes support for RDMA RoCE drivers and OFED stacks. You can leverage SCC's RDMA capabilities through the IB verbs API or enable RDMA-based communication using MPI.

Which ECS instance types support preemptible instances?

Instance types eligible for pay-as-you-go also support preemptible instances. Should a particular instance type be unavailable due to resource constraints, you may opt for an alternative instance type.

Which regions support preemptible instances?

You can create preemptible instances in any region. Should you encounter resource shortages that prevent creation in a particular region, consider trying an alternative region.

Why is the preemptible instance option not visible when purchasing an ECS instance?

The availability of preemptible instances is contingent on your ECS usage. If you do not see the option for a preemptible instance, it means that your current account does not yet qualify for this feature.

How to bid for preemptible instances?

When creating a preemptible instance, you must set a maximum hourly price within your budget. If your bid exceeds the current market price, the instance will be activated and billed at the market rate. For more information, see creating preemptible instances.

What is the relationship between the maximum bid price and the current market price of preemptible instances?

If your maximum bid price exceeds the current market price, the preemptible instance is successfully launched and charged at the market price. Once launched, the instance enjoys a protection period where it's immune to automatic release due to market price surges, even if they surpass your maximum bid.

Upon the protection period's conclusion, the system evaluates the market price and available inventory for the instance type at five-minute intervals. Should the market price rise above your maximum bid or if there's a shortage of the instance type, your running preemptible instance will be terminated.

Can I view the current market price before purchasing a preemptible instance?

Yes. You can view the market price range and historical pricing for the chosen instance type directly in the console when creating a preemptible instance. After selecting an instance type, the total cost for configuration, encompassing the instance type, storage, and network bandwidth, will be displayed at the bottom of the creation page. The listed price for the instance type reflects its current market value.

Can I view the historical prices of preemptible instance types? How can I view them?

Yes, you can. When creating a preemptible instance through the ECS console, you have the option to view the historical prices for the chosen instance type. Additionally, you can query historical prices using the DescribeSpotPriceHistory operation.

How many preemptible instances can be purchased with one account?

The allocation of preemptible instances is determined by the number of vCPUs rather than the instance count. Upon selecting an instance type for a preemptible instance, you can check your remaining vCPU quota. For more information, see instance limits.

Can preemptible instances be converted to subscription instances?

No, preemptible instances cannot be converted to subscription instances.

Do preemptible instances support changing instance types?

No, they cannot.

How to switch to higher configurations at low cost when preemptible instance configurations are insufficient?

If you need to upgrade your preemptible instance but changing the instance type is not supported, you can create a custom image from the system disk of your current instance. Then, use this custom image to launch a new preemptible instance with higher specifications. For more information, see create a custom image from a snapshot and create an instance from a custom image.

When purchasing preemptible instances, will they be released due to insufficient inventory if automatic bidding is selected?

Choosing automatic bidding caps the bid price at the pay-as-you-go rate, which safeguards against instance release due to spot price changes. Nonetheless, instances might still be released if there's an inventory shortage for the chosen instance type.

When purchasing preemptible instances, if the maximum price is set to the highest possible bid, which is the pay-as-you-go price, is it still possible for the instance to be released?

You can set the maximum bid at the pay-as-you-go rate to protect your instance from being reclaimed due to spot price changes. Nevertheless, your instance might still be released if there's a shortage of resources for that particular instance type.

Can I cancel or reschedule the release time of a preemptible instance after it is set?

Yes, you can cancel or reschedule the automatic release of your preemptible instance at any time before the release occurs.

Why was my preemptible instance released even though I have no overdue payments?

Your preemptible instance is automatically released after its protection period expires if the spot price surpasses your bid or resources become insufficient.

How is the release rate on the sales page calculated? Why is it mostly between 0% and 3%?

The release rate is determined by zone and instance type. It represents the percentage of instances released in a single day, calculated by dividing the number of released instances by the total held that day. Typically, release rates range from 0% to 3%. Alibaba Cloud strives to minimize the release rate, enhancing the stability of preemptible instances for users.

Are notifications sent when preemptible instances are released? How are they sent?

Yes, notifications are sent. If a preemptible instance is to be released due to a spot price change or insufficient resources, it first transitions to the To Be Released state and will be automatically released after approximately 5 minutes.

To subscribe to notifications, you can use CloudMonitor to set up alerts for preemptible instance interruption events. For more information, see set event notifications.

To verify the To Be Released state, you can check the instance metadata or the OperationLocks information provided by the DescribeInstances API call. For more information, see overview of ECS instance metadata and DescribeInstances.

Can data be automatically retained after a preemptible instance is released?

No, data is not retained upon release. However, you can create snapshots to back up your data and environment before releasing the instance. Then, when necessary, you can purchase a new preemptible instance. For more information, see create snapshots.

What to do if a preemptible instance cannot be started after being stopped in economical mode?

If a preemptible instance in economical mode fails to start due to resource shortages, you can preserve your business data by following these steps:

1. Create snapshots of the instance's disks (referred to as instance A). For details, see create snapshots.

   Once the snapshot creation progress reaches 100%, the snapshot is successfully created.

2. Generate a custom image from the snapshot. For more information, see create a custom image from a snapshot.

   If the image status is Available, the custom image has been created.

3. Create an instance, referred to as instance B, using the custom image you've created.

   Create a new instance (referred to as instance B) using the custom image. For details, see create an instance from a custom image.

4. Verify the accuracy of the business data on instance B.

   • If the data is correct, you can release instance A. For more information, see release an instance.

   • If the business data is incorrect, you can detach the disks from instance A (ensure not to select the option to release the disks along with the instance upon detachment), attach them to instance B, and then terminate instance A. For more information, see how to detach a data disk and how to terminate an instance.

After the protection period of a preemptible instance ends, if the spot price > bid price, the instance enters the to be released state. During the 5-minute notification period before the instance is released, if the spot price < bid price, will the instance be retained?

Yes, the preemptible instance will be released. Receiving the notification that the instance is to be released sets an irreversible status, ensuring the instance's scheduled release.

How to retain the public IP address of a preemptible instance as much as possible?

When a preemptible instance is released, its public IP address is reclaimed. To retain a specific public IP address, consider using the Elastic IP Address (EIP) feature. For more information, see Elastic IP Address.

You can convert an existing public IP address to an EIP. For detailed instructions, see Convert a static public IP address to an Elastic IP Address and Convert a static public IP address of a classic network ECS to an Elastic IP Address.

How to view the bid price of a preemptible instance?

To see the bid price for a preemptible instance, check the Billing Information section on the instance details page.

How is a preemptible instance billed if it is released in less than 1 hour?

If you release a preemptible instance within one hour, billing is calculated based on the actual usage time in seconds. The spot price at the instance's successful creation determines the hourly rate, with the per-second rate being the hourly price divided by 3,600. Charges are rounded to two decimal places, and any accrued costs under USD 0.01 are not billed.

What resources are included in the price discount for preemptible instances?

The discount for preemptible instances applies exclusively to the instance types. Other resources, including system disks, data disks, and network bandwidth, are priced the same as pay-as-you-go instances and do not receive a discount.

Why do spot prices for the same instance type vary significantly across different zones within the same region at the same time?

Price disparities arise due to differences in inventory levels. Each zone is associated with distinct data centers. Within a single region, data centers across various zones can differ in construction conditions, capacity, sales, and deployment policies. Consequently, the availability of the same instance type can fluctuate between zones. Spot prices for preemptible instances are subject to supply and demand dynamics, resulting in variable pricing.

Can the spot price of preemptible instances exceed the pay-as-you-go price?

No, the spot price for a preemptible instance may reach, but will not surpass, the pay-as-you-go rate. Rest assured that you will not be charged more for a preemptible instance than you would for an equivalent pay-as-you-go instance.

Will the fees I pay for a preemptible instance change with market price fluctuations within the first hour of operation?

No. The hourly rate for a preemptible instance is fixed at the start of each hour and applies for that entire hour.

Will charges continue after a preemptible instance is stopped?

Why is the billing price from 8:30 to 9:00 USD 1.5 per hour when I bid USD 2 per hour at 8:00 and the spot price became USD 2 per hour at 8:30?

Between 8:30 and 9:00, the instance was within the creation protection period and was billed at the transaction price of USD 1.5 per hour. Once the protection period ended, billing switched to the current spot price.

Do preemptible instances continue to be billed when they enter the to be released state?

Once a preemptible instance enters the 'to be released' state, it will be automatically released within approximately 5 minutes. The instance remains operational and incurs charges during this time. Billing ceases upon the instance's release.

Which is more cost-effective: preemptible instances without a protection period or with a protection period?

For optimal cost savings, preemptible instances without a protection period are the more economical choice, providing roughly a 10% additional discount compared to those with a protection period.

Is the release rate of preemptible instances without a protection period higher than those with a protection period?

Are preemptible instances without a protection period prioritized for release?

No, they do not.

Will all newly created preemptible instances be without a protection period after creating one without a protection period?

Are preemptible instances without a protection period still notified 5 minutes before release?

Yes, you will still receive a notification 5 minutes prior to release, providing the same processing time as preemptible instances that have a protection period.

Resource quantity: do preemptible instances without a protection period have fewer resources than those with a protection period?

No, preemptible instances are allocated the same amount of resources regardless of whether they have a protection period.

Does it support switching between instances with and without a protection period?

No, it is not. Preemptible instances come with a default protection period of one hour upon creation. The protection period for these instances can only be set at the time of creation and cannot be modified afterwards.

Can I upgrade the instance type and configurations of a subscription ECS instance?

Yes, you can. For more information, see how to upgrade instance types for subscription instances.

Can I upgrade the instance type and configurations of a pay-as-you-go ECS instance?

Yes, but you must first stop the instance. For more information, see how to change the instance type of a pay-as-you-go instance or use the ModifyInstanceSpec API operation to upgrade the configurations.

How long does it take to upgrade an ECS instance?

• You can upgrade the instance type of a subscription ECS instance without interruption, and the process typically completes in about 15 minutes.

• To upgrade the instance type of a pay-as-you-go ECS instance, you must first stop the instance. This upgrade also takes approximately 15 minutes.

• Bandwidth upgrades for an ECS instance can be performed without downtime, and the process is usually finished within 5 minutes.

How are the fees calculated for upgrading ECS instances?

Upon upgrading the instance type or other configurations of an ECS instance, the applicable fees and their calculation method will be displayed on the upgrade page. For detailed fee information, you can visit the or Billing and Cost page.

Does upgrading ECS instances affect the configurations of my cloud-based business?

Upgrading pay-as-you-go instances requires you to stop the instances beforehand. For subscription instances, a restart is necessary for the new configurations to activate. Upgrading temporarily disrupts services running on the instances. It is advisable to perform upgrades during off-peak hours to minimize impact. Post-upgrade, the instance will continue to operate smoothly without any need for reconfiguration.

How to upgrade ECS resources?

Why does upgrading instance configurations have no effect?

To apply the upgraded configurations to your instance, you must restart it either through the console or by invoking an API operation.

Can upgrade orders be directly canceled to revert to the original configurations?

Once an upgrade order is effective, the configurations are permanently upgraded and cannot be reversed. To restore the instance to its original configurations, you can perform a downgrade. Please note that downgrading will incur charges.

Which of my applications require higher-performance local disks?

Problem description: An ECS instance remains in the Starting state for an extended period after launch and then shuts down automatically. Upon logging into the system, you discover that the Aliyun Assist Service has been deleted or disabled.

How to upload files using FTP tools on macOS?

Method 1: Using macOS's Built-in Terminal

You can upload files using the macOS Terminal or iTerm2 (Download iTerm2). Ensure you select the correct path for file uploads.

1. Connect to the FTP service.

2. Navigate to the appropriate directory. For Windows systems, no change is needed. For Linux systems, switch to htdocs.

3. Execute the put command to start the upload.

Method 2: Using a Third-Party FTP Tool

1. Download Yummy FTP.

2. Install Yummy FTP on your macOS.

3. Input the server IP address, username, and password for the FTP service. Set the protocol to Standard (FTP) and the port to 21, unless you're using a different port. Do not select an SSH key.

4. Click Connect.

5. In the right window, select the directory for the upload (Windows users need not select a directory; Linux users should choose the htdocs directory). Then, in the left window, select the files you want to upload and click upload.

   If you receive a prompt during Yummy FTP's installation stating that your security preferences only permit applications from the Mac App Store and recognized developers, follow these steps to configure your computer:

   1. Open System Preferences > Security & Privacy.

   2. Click the security lock in the lower-left corner and enter the administrator password to make changes.

   3. In the Allow Apps Downloaded From section, choose Anywhere.

   After configuring the settings, you can proceed with the normal software installation.

If the issue remains unresolved, you can submit a ticket.

How to file a domain name after purchasing an ECS instance?

Your purchased ECS instance must comply with the ICP filing requirements. For more information, see the ICP filing server preparation and check.

Why does the ECS instance fail to load the kernel at startup?

Problem description: Upon selecting an option from the GRUB menu during system startup, there is no response. After mounting the LiveCD image and accessing the system on the ECS instance, the file system permissions appear normal, and the message logs do not indicate any exceptions.

Cause: The system has been compromised by ransomware.

Solution: Ensure you back up critical data and then reinitialize the system.

How to change the logon password inside an instance?

For detailed instructions, see how to change the logon password within an instance.

Purchased an ECS instance and want to add a sound card and video card, but found it cannot be added. What is the reason?

Alibaba Cloud ECS instances are designed as standard servers and are not equipped with multimedia capabilities such as sound or video cards. Consequently, these components cannot be added to the system.

Can the remaining usage time of an ECS instance be transferred to another ECS instance?

No, transferring the remaining usage time of an ECS instance to another is not supported. To balance flexibility and cost, consider selecting a pay-as-you-go instance for purchase and combine it with reserved instances. For more information, see the reserved instance overview.

Does the cloud server provide a database by default?

Can databases be set up on cloud servers?

Yes, you can install database software and tailor the environment to your requirements on ECS, as there are no imposed restrictions. Alternatively, you can opt for Alibaba Cloud's RDS database services.

Does elastic compute service support Oracle databases?

Yes, it does. Before installing an Oracle database, we recommend conducting a performance stress test on the ECS instance to ensure it meets the database's read and write requirements.

Are public and private IP addresses independent, and can you specify or add IP addresses?

In a classic network, public and private IP addresses operate independently. Private IP addresses facilitate communication within ECS instances and between ECS instances, OSS buckets, or RDS instances. An ECS instance with a public bandwidth of 0 Mbit/s will not be assigned a public IP address. Under normal conditions, IP addresses remain constant, and it is not possible to specify, choose, or acquire additional IP addresses.

In a Virtual Private Cloud (VPC), public IP addresses are linked to private IP addresses through NAT. You have the option to designate secondary private IP addresses or expand your IP address allocation using the multi-IP feature.

Can cloud servers perform load balancing?

Cloud servers running on both Linux and Windows support load balancing. It's essential to maintain consistent configurations across web servers and website code. While it's possible to implement load balancing with a single ECS instance, for optimal performance, using two or more ECS instances is recommended.

Can I change the region after purchasing an ECS instance?

No, you cannot change the region after purchasing an ECS instance. However, you can use the CloudOps Orchestration Service public template ACS-ECS-CloneInstancesAcrossRegion to replicate an ECS instance in another region. The replicated instance will have identical disk data, but the IP address may change.

Can you adjust the partition size of purchased disks?

To ensure system security and stability, repartitioning system disks on Windows or Linux instances is not possible. Using third-party tools to repartition system disks forcefully may introduce risks like system crashes or data loss.

While data disks can be repartitioned, it is a risky process that can result in data loss and is generally not recommended.

How to replace the static public IP address of an instance with another EIP?

If you chose the static public bandwidth option when purchasing an ECS instance and have acquired a new EIP, follow these steps:

• Method 1:

  1. Switch the static public bandwidth to a pay-as-you-go model. For more information, see Public Bandwidth Billing.

  2. Change the static public IP address to an EIP. For more information, see Converting a VPC ECS Instance's Static Public IP to an EIP.

  3. Detach the converted EIP. For more information, see Detaching an EIP.

  4. Attach the new EIP. For more information, see Attaching an EIP to an ECS Instance.

• Method 2:

  In the OOS console, utilize the ACS-ECS-ConvertsPublicIPToNewEIPByInstanceId maintenance template to switch an instance's public IP to a new EIP. For the China (Hangzhou) region, visit ACS-ECS-ConvertsPublicIPToNewEIPByInstanceId. You can access the link and adjust the region via the top menu bar in the OOS console.

How to view subscription instances in all regions within the same account?

When can you force stop an instance? What are the consequences?

If an instance fails to stop through the proper shutdown procedure, you can force a stop. This action is akin to a power outage and may lead to the loss of any unsaved data within the instance's operating system.

Why does the reboot fail?

Why was an ECS instance in a scaling group automatically released even though release protection was enabled when using the auto scaling feature?

Even if release protection is enabled for an ECS instance from the instance list in the ECS console or through the ModifyInstanceAttribute operation after Auto Scaling has created it, this does not stop Auto Scaling from releasing the instance automatically.

To prevent automatic release, you can set the ECS instance to a protected state via the Auto Scaling console. For more information, see how to change an instance to the protected state.

Can I directly deploy Redis or MySQL applications when persistent memory is used as local disks? Do I need to modify applications as I do when persistent memory is used as memory?

For a comprehensive guide to packet forwarding testing, consult network performance testing methods. For specialized packet forwarding testing methods for g7, c7, and r7 instances, refer to network performance testing with Pktgen.

How does the performance of persistent memory compare with that of local NVMe SSDs and cloud disks when persistent memory is used as local disks?

For detailed guidance on copying ECS instances across regions using custom images, see copying ECS instances across regions with custom images.

Additionally, the Server Migration Center (SMC) facilitates full, incremental, batch, and VPC-based migrations. Data can also be transferred between ECS instances within the SMC console. For more information, see cross-account and same-account ECS instance migration.

How to check Linux instance disks and perform partition formatting?

You can use the df -h command to check disk space and usage, and the fdisk -l command to view disk information. For guidance on partitioning and formatting Linux instance disks, see initializing a data disk (Linux).

How to upload files to a Linux instance?

You can utilize the FTP service to transfer files to a Linux instance.

How to change the owner and group of directories and files on a Linux instance?

Incorrect file or directory permissions on a web server can lead to a 403 error when trying to access the website. It's essential to determine the identity under which the server process runs before making any adjustments to files and directories.

You can use the ps and grep commands to find out the identity that the process of the files and directories is running under.

To check the owner and group of files and directories, use the ls -l command.

To modify permissions, the chown command is used. For instance, the command chown -R www.www /alidata/www/phpwind/ changes the owner and group of all files and directories within the /alidata/www/phpwind directory to the www user.

What are the limits for transferring and changing the public IP address of an ECS instance?

If the problem persists, submit a ticket.

Can ECS instances access the Amazon website?

Yes, ECS instances can access the Amazon website, provided they have a proper Internet connection.

Why can't I access websites outside China after logging on to an ECS instance?

To access websites hosted outside the region of your ECS instance, ensure that both the instance has a proper Internet connection and the websites adhere to local laws and regulations.

What is the AliVulfix process in ECS instances?

AliVulfix is a program from Security Center designed to scan ECS instances for security vulnerabilities.

Can elastic compute service instances defend against network attacks?

Alibaba Cloud ECS instances come with Security Center, which provides defense against DDoS attacks. In the event of network attacks, CloudMonitor automatically identifies and mitigates abnormal traffic. For particularly large-scale attacks, Alibaba Cloud employs blackhole filtering to safeguard your ECS instances. Additionally, for enhanced security, you can install protective security software and disable infrequently used ports.

What security services does Alibaba Cloud offer?

Alibaba Cloud Security Center leverages advanced data analytics to provide a comprehensive suite of security solutions. This suite includes essential services such as security vulnerability scanning, website malware detection, host intrusion prevention, and robust anti-DDoS measures, offering complete protection for your digital assets.

For further information on security services, refer to the security products page.

How to detect if a server is mining?

To determine whether a server is running mining programs, refer to the characteristics of mining programs and use the methods below:

• Monitor the server's CPU for normal activity.

  Note

  If you notice the server's CPU utilization spikes to 80% or higher, accompanied by an unknown process continuously sending packets, it's likely that a mining program is active on your server.

• Visit the Security Center console security alert page to review any ongoing attack alerts.

For further details, see best practices for defending against mining programs and handling mining programs.

How to handle mining programs or apply for unlocking?

Addressing Mining Programs

You can address mining programs or request to unlock affected servers through the Security Center, or manage mining worms via the Cloud Firewall console. For more information, see best practices for defending against mining programs and best practices for handling mining programs.

Requesting Server Unlock

If your server is locked due to a mining virus infection or attack, you can submit an unlock request. For the unlock portal, visit the penalty list.

How to activate Windows instances in a VPC network?

For details on activating Windows instances in a VPC network using a KMS domain name, refer to activation methods for Windows instances in a VPC network.

How reliable is persistent memory?

You can view the expiration date of subscription instances on the Instance List page of the ECS console, under the Billing Method column.