Platform For AI:DSW FAQ

Q: DSW billing and idle charges

• DSW supports subscription and pay-as-you-go billing methods. For billing details, see DSW Billing.

• With pay-as-you-go, you are charged while your instance is running. Billing continues as long as the instance is in a 'running' state, even if you are not executing code or have the WebIDE open.

Q: How do I view my DSW bill?

Pay-as-you-go users can view bill details on the Expenses and Costs page. View bill details.

Q: Charges after stopping an instance

Charges that continue after you stop an instance typically occur for two main reasons:

• Billing delay: For pay-as-you-go DSW instances, there is a delay in bill generation and deductions. A charge you receive may be for usage that occurred before the instance was stopped, not for usage after it was stopped.

• System disk expansion: If you expanded the system disk when you created the DSW instance, you continue to incur storage fees for the disk, even when the instance is in a 'stopped' state. An expanded system disk cannot be downsized. To stop all storage charges, you must delete the instance. Before you delete the instance, make sure to back up all necessary data. Data cannot be recovered after an instance is deleted.

  To check if the system disk was expanded, click the instance name to go to the instance details page. Check if the system disk capacity exceeds the free quota. Instances in a public resource group receive 100 GiB of free capacity.

  

Q: Stop all billing

• The only way to completely stop all billing for a DSW instance is to delete the instance. Before you delete the instance, make sure to back up all necessary data. Data cannot be recovered after the instance is deleted.

  

• Switch between different workspaces and regions to ensure all instances are deleted.

  

Q: Pay-as-you-go per-minute billing

Pay-as-you-go instances are billed by the minute. The formula is: Bill Amount = (Hourly Rate / 60) × Actual Service Duration (in minutes).

Q: "Failed to pull image" error for private repositories

To create a DSW instance using an image URL from a private repository, such as a private Alibaba Cloud Container Registry (ACR) repository, you must enter the username and password for the repository.

Q: Image creation error: "insufficient capacity of ephemeral storage"

Cause: This error occurs when the available space on the system disk is insufficient for the image's write layer.

Solution: In the DSW Terminal, run df -h to check disk usage. Ensure that the space used by the overlay does not exceed the available space on /dev/vda4. If it does, resolve this by setting a Custom Excluded Path when creating the image.

Q: How to use Docker images in DSW?

• To start a DSW instance from a Docker image, push the image to Alibaba Cloud Container Registry (ACR), add it as a custom image in your PAI workspace, and then select it when creating the instance.

  • To push a Docker image to Container Registry (ACR), see Use a Personal Edition instance to push and pull images.

  • To add a PAI custom image, see custom images.

• To package the current DSW environment as an image for starting other instances or deploying models, see Create a DSW Instance Image.

• Install and use Docker in the DSW cloud IDE: Only DSW instances created from a Lingjun Resource Group or a Version 1.0 General Resource Group support the nested container feature. Manage Nested Containers with DockerBoard and Use Docker in DSW.

Q: Why does DSW image creation fail or time out?

• Image size limit exceeded: When you create a DSW image, a single image layer cannot exceed 10 GiB. Otherwise, the build fails. Reduce the image size to prevent this error.

• Region mismatch: The DSW instance and the Container Registry (ACR) instance must be in the same region. Otherwise, the system cannot find the corresponding image repository during image creation.

• Insufficient system disk space: During image creation, if the available system disk space is less than the data size required for the image layer, an "insufficient capacity of ephemeral storage" error occurs.

• Network issues: The Personal Edition of ACR pushes images over the public network. Network jitter or long transfer times can cause large image pushes to fail or time out. For faster and more stable performance, bind an Enterprise Edition of ACR instance to the same VPC as the DSW instance to push images over the internal network.

Q: "Create Image" button is disabled or image repository not found

1. Incorrect instance status: You can use the "Create Image" feature only on running DSW instances. The button is disabled if the instance is stopped or in any other state.

2. Prerequisites not met or configuration error:

   • You must first create a Container Registry (ACR) instance in the same region as the DSW instance, and then create a namespace and an image repository within it.

   • Ensure that the DSW instance and the ACR instance are in the same region.

Q: Image creation error: "Push container failed..."

When creating an image, ensure that the data in a single layer does not exceed 10 GiB; otherwise, the build will fail. For DSW instances that use public resource groups, you can set a Custom Exclusion Path to exclude files or directories from the final image. Alternatively, you can mount a storage path, such as an OSS path, to store and access your data.

System disk size and usage

By default, files and data in a DSW instance are stored on the system disk, which comes with a free quota.

• Check the free quota

  DSW instances in a public resource group include a 100 GiB free quota. General-purpose compute resources include a free system disk only with specific instance types. Lingjun Intelligent Computing Resources do not provide free cloud disks. Check the free quota in the System Disk section of the instance configuration page:

  1. Click the instance name on the Instance List page.

  2. In the upper-right corner, click Change Settings and scroll down to the System Disk section.

  

• Check system disk usage

  Click the DSW instance name to view the system disk usage in the Environment Information section.

  

• What to do when the system disk is full

  If system disk usage exceeds the free quota, you can expand the system disk or mount a dataset.

System disk scale-in

No. A DSW system disk cannot be scaled in after an expansion. If a DSW instance's system disk is larger than you need, you can back up important information by using a dataset, OSS, NAS, or CPFS. After the backup, you can delete the DSW instance to stop billing. Then, create a new DSW instance with an appropriate system disk size.

Q: Mounting and using a custom file system

You can mount an OSS, NAS, or CPFS file system when you create a new instance. You can then access the mounted directory through the DSW Terminal to view and use your files.

Q: NAS mount error: "MountTarget ... is not in VPC"

• Cause: This error occurs because you configured a mount point when creating the NAS dataset.

• Solution: Leave the mount point field empty when you create the dataset.

Q: Mount error: "wrong fs type, bad option, bad superblock"

• Symptom

• Solution

  Before you run the mount command, install the nfs-utils package.

  yum install nfs-utils

Q: "Input/output error" when accessing a mounted OSS dataset

This issue occurs because the role lacks the required OSS access permissions (AliyunPAIDLCAccessingOSSRole). For specific authorization steps, see PAI service account authorization.

Q: Reducing OOM risk when mounting OSS with JindoFuse

You can resolve this issue using one of two methods:

• Method 1: Use JindoFuse 6.8.1, which has optimized memory usage.

  {
      "fs.jindo.fuse.pod.image.tag":"6.8.1"
  }

  

• Method 2: Use ossfs.

  When you submit the task, specify the following:

  {
      "mountType": "ossfs"
  }

  

  To reduce metadata cache usage when listing directory contents and mitigate OOM issues, you can disable the readdirplus optimization with the following configuration:

  {
      "mountType": "ossfs",
      "fs.ossfs.args": "-oreaddirplus=false"
  }

Q: Mounted OSS not visible in the JupyterLab file browser

This is because the DSW file browser shows the instance's working directory by default, which is typically /mnt/workspace. The mount path you specified for OSS, for example, /mnt/data, is not under the default working directory, so the file browser does not display it in the file list.

Solutions:

• Access through code: Your files are already successfully mounted. In your code, you must use the full mount path to access them, such as open('/mnt/data/my_file.csv').

• Change the mount point: To easily view the files in the UI, you can set the mount path to a subdirectory of the working directory during configuration, such as /mnt/workspace/my_oss_data. After the mount is complete, you can see your OSS files in the my_oss_data folder in the file browser.

• Access through the Terminal: You can use the cd /mnt/data command in the DSW Terminal to change to the mounted directory, and then use commands like ls to view and manage files.

Q: OSS mount errors: "Transport endpoint not connected" or "Input/output error"

This error indicates that the DSW instance has lost its mount connection to OSS. Possible causes and troubleshooting steps are as follows:

1. RAM role permission issue: Check if the RAM role configured for your DSW instance has permissions to access OSS (for example, AliyunPAIDLCAccessingOSSRole). Insufficient permissions are a common cause of this connection failure.

2. Insufficient resources for the mount service: During high-intensity random read/write operations or operations on a large number of small files, the ossfs or JindoFuse process responsible for mounting may crash due to an Out of Memory (OOM) error. In the advanced mount configuration, you can disable the metadata cache or increase the memory allocation. JindoFuse.

3. Restore the connection:

   • For mounts configured at startup, the easiest way to restore the connection is to restart the DSW instance. The system automatically remounts the path.

   • You can also use the PAI SDK to execute a dynamic mount command to remount the path without restarting the instance.

Q: Supported mount types in DSW

DSW can mount cloud storage services like OSS, NAS, and CPFS, either by creating datasets or by directly mounting paths.

• Alibaba Cloud Drive is not supported: DSW does not currently support directly mounting a personal Alibaba Cloud Drive. We recommend storing data that needs to be processed in OSS.

• Mounting MaxCompute tables is not supported: MaxCompute (formerly ODPS) tables cannot be directly mounted as a file system in DSW. Use the SDKs or APIs provided by PAI to read and write this data. Read and write MaxCompute tables using PyODPS.

Q: Data persistence and migration for DSW instances

The system disk of a DSW instance is ephemeral storage. For instances in a public resource group, data is cleared if the instance is stopped for more than 15 days. For instances in a dedicated resource group, data on the system disk is also cleared when the instance is stopped or deleted.

To persist your code and data and migrate them between instances, use externally mounted storage.

• Persistence solution: Save all your important data, code, and models under mounted OSS or NAS paths. This ensures your assets remain safe in your own OSS or NAS even if the DSW instance is deleted.

• Migration solution: When you need to migrate data from one DSW instance to another, simply mount the same OSS or NAS path that contains the data to the new instance. This is the most convenient way to migrate data.

Q: Working directory files not appearing in mounted OSS

The default OSS mount path is /mnt/data, which is separate from the default DSW working directory at /mnt/workspace. Therefore, files in your working directory are not automatically visible in the mounted OSS bucket. To make them visible, copy them to the mount path at /mnt/data. After copying, the files will appear in your OSS bucket.

cp -r /mnt/workspace/. /mnt/data/

Q: Reading data from OSS in DSW

Use a Python SDK or API to read data from OSS. Read and write data in OSS.

Q: Uploading or downloading a folder

DSW does not support uploading or downloading folders directly. As a workaround, compress the folder into a single file. The DSW Terminal includes standard Linux tools, such as tar, gzip, and unzip, to compress and extract files. The following example demonstrates how to use tar.

1. Run tar --version to check if tar is installed. If not, install it using the following commands.

   # Installation command for Debian-based systems (for example, Ubuntu)
   sudo apt install tar
   
   # Installation command for Red Hat-based systems (for example, CentOS or Fedora)
   sudo yum install tar

2. Compressing and extracting a folder.

   # Compress a folder. Replace /path/to/directory with the path to your folder.
   tar -cvf archive_name.tar /path/to/directory
   
   # Extract a folder.
   tar -xvf archive_name.tar

Q: Transferring data between DSW instances

You can use one of the following methods:

• Mount a dataset, OSS, NAS, or CPFS: Mount the same dataset or OSS path to both DSW instances. Both instances can then access any data stored in the shared path.

• Upload and download files: Download the data from the source DSW instance and then upload it to the target DSW instance.

Q: Download fails or is unresponsive

This issue is typically caused by network congestion or browser issues. Try the following steps:

1. Wait a moment. Large files take longer to respond and download.

2. Switch to a different browser or use your browser's incognito mode and try again.

3. If you are downloading large files (over 200 MB) or have an unstable network connection, we recommend mounting OSS to download them.

Q: File Transfer Station is out of space

The File Transfer Station has a total capacity of 10 GB. To free up space, go to its management page and clear the transfer station files. If the page does not update automatically, refresh your browser.

Q: Redirection to File Transfer Station during upload

This is the expected behavior. To ensure upload stability and speed, all files larger than 10 MB are automatically routed through the File Transfer Station. They are then saved to your instance after the transfer is complete.

Q: Uploading large files or datasets to DSW

The system disk provides only limited temporary storage. Instead of uploading large files directly, upload data to OSS and mount it to your DSW instance. Mount a dataset, OSS, NAS, or CPFS.

Q: ProxyClient SSH connection disconnects

When you connect to a DSW instance over SSH, your session may disconnect after a period of inactivity. You may see the following error:

To permanently fix this issue, we recommend using the more stable Remote connection: Direct SSH connection method.

Q: Cannot open local folders in VSCode

This issue is typically caused by the VSCode client. We recommend uploading the files from your local folder to your DSW instance. For instructions, see Upload and download files.

Q: Direct SSH connection setup failsFailed to update private zone items: Failed to add zone

This error occurs when internal DNS resolution is disabled. To enable it, see Enable internal DNS resolution.

Q: How do I use third-party libraries in DSW?

DSW supports installing third-party libraries. Manage third-party libraries.

Q: Are packages installed with pip and my code lost after stopping a DSW instance?

If you use a cloud disk as the system disk, your data will not be lost. All data on the instance's disk, including the environments in the /mnt/workspace and /root directories, is preserved. The next time you start the instance, all your environments and files are still available. Data is permanently cleared only when you delete the instance.

Q: Why does a third-party library I installed not work?

If you install a third-party library by using the pip command but cannot find the library when you run the import command, first try restarting the service or Kernel. If the error persists, check which environment you are using. By default, DSW installs third-party libraries to the Python 3 environment. To install a library in a different environment, you must first switch to that environment. The following example shows how to do this.

Install in the Python 2 environment.
source activate python2
pip install --user xxx
Install in the TensorFlow 2.0 environment.
source activate tf2
pip install --user xxx

Replace xxx with the name of the third-party library you want to install.

Q: How to fix pip install failures in DSW from dependency or version errors?

This issue is typically caused by an incompatible environment. To troubleshoot this issue, follow these steps in order:

1. Recommended solution: Change the image. Stop the current instance, create a new DSW instance, and select a different official image. For example, if the PyTorch 2.1 image does not work, try the PyTorch 2.3 image. You can also try an image from the modelscope series, which typically offers better compatibility.

2. Install a specific version. Check the official documentation for the library to find a version that supports your current DSW environment, including the Python and CUDA versions. Then, run pip install package_name==x.y.z.

3. Change the download source. Try using a different mirror, such as the Tsinghua University mirror: pip install -i https://pypi.tuna.tsinghua.edu.cn/simple <yourLibraryName>.

Q: Why is a library installed in the Terminal not found in a Jupyter Notebook?

This issue can occur if the Terminal and Jupyter Notebook use different Python environments. You can run the which python command to check which Python environment is active. Alternatively, you can install the required library directly in the Notebook, for example:

Q: The CUDA driver version is too low. Should I upgrade the NVIDIA driver in DSW?

Do not upgrade the driver version. DSW instances have pre-installed, locked drivers and CUDA. You cannot modify them manually — doing so can irreparably damage the instance. Instead, stop the current instance, create a new one, and select an official image with a higher CUDA version.

For example, the official image modelscope:1.9.4-pytorch2.0.1tensorflow2.13.0-gpu-py38-cu118-ubuntu20.04 has cu118 in its name, which indicates CUDA version 11.8.

Q: Can I use Docker in DSW to deploy my application?

To use Docker with Lingjun resources, submit a ticket to be added to the allowlist. DSW instances that do not use Lingjun resources do not currently support running Docker inside the instance container.

Q: How can I extract files if the unzip or 7z command is missing?

You can use the apt-get command to install the tools.

• To install unzip: in the Terminal, run apt-get update && apt-get install -y unzip. Then, use unzip your_file.zip.

• To install p7zip (for 7z files): in the Terminal, run apt-get update && apt-get install -y p7zip-full. Then, use 7z x your_file.7z.

Q: Why does the installation of a third-party library get stuck or time out?

If the installation of a third-party library is stuck, times out, or is extremely slow, this is typically a network issue. Follow these steps to troubleshoot the issue:

Step 1: Check network connectivity

In the Terminal, run the ping www.aliyun.com command to test whether the instance can access the internet. If the network is unreachable, proceed to Step 2.

Step 2: Check the gateway configuration

On the instance configuration page, check the type of the Internet Access Gateway:

• Public Gateway: By default, DSW uses a Public Gateway to access the internet. You can confirm the gateway type on the DSW instance configuration page. The Public Gateway has limited bandwidth, which may be insufficient for downloading large files. In this case, you can switch to a Dedicated Gateway.

• Dedicated Gateway: Provides higher network speeds. You must create an Internet NAT Gateway, associate an EIP, and configure an SNAT entry in the VPC. Otherwise, you cannot access the internet. Improve public network access speed through a dedicated gateway.

Step 3: Change the pip download source

DSW uses an Alibaba Cloud mirror source by default, but issues can occur during peak hours or due to network fluctuations. We recommend switching to a different mirror source in China:

# Use the Tsinghua University mirror (recommended)
pip install -i https://pypi.tuna.tsinghua.edu.cn/simple --trusted-host pypi.tuna.tsinghua.edu.cn <yourLibraryName>

# Use the USTC mirror
pip install -i https://pypi.mirrors.ustc.edu.cn/simple --trusted-host pypi.mirrors.ustc.edu.cn <yourLibraryName>

# Use the Douban mirror
pip install -i https://pypi.doubanio.com/simple --trusted-host pypi.doubanio.com <yourLibraryName>

To permanently change the default pip source, see View or change the pip source.

Step 4: Use offline installation

If the network is unreachable or very unstable, you can perform an offline installation:

1. On your local computer (with good network connectivity), download the .whl library file:

   # Run on your local computer
   pip download <yourLibraryName> -d ./packages

2. Upload the downloaded .whl file to your DSW instance. For instructions, see Upload and download files.

3. In DSW, perform the offline installation:

   pip install /path/to/your-package.whl

Q: How do I get root permissions in the DSW WebIDE?

Most official DSW images run as root by default. If the command prompt shows root@..., you already have root permissions. You can safely ignore the "not recommended to run as root" warning during pip installation. If your image does not use root, switch to an image that supports root access.

Q: How do I start X server in DSW?

DSW does not support starting X server.

Deploy a DSW model

• Use the EAS model deployment service

  After modeling, use PAI-EAS to deploy the model as an online service. Deploy a model as an online service.

• Download the model for local deployment

  You can right-click a model generated by DSW to download it to your local machine.

Q: Why am I prompted to log in again after inactivity?

For security, your DSW login session expires after three hours. You must log in again, but this does not interrupt running tasks. To run long tasks, use thenohup command in the DSW Terminal.

Q: Do tasks continue to run after I close my browser or PC?

Yes. A DSW instance runs in the cloud, independent of your local device. However, instances with an idle auto-shutdown policy (enabled by default for free trials) may stop automatically if CPU and GPU utilization remains below a threshold.

Q: Why can't DSW start Docker?

Because DSW runs in a container, DSW does not support installing Docker. The corresponding CUDA version is pre-installed on the underlying virtual machine and cannot be changed. You can use nvidia-smi to view the corresponding CUDA version.

Q: Why is tab auto-completion unavailable in the terminal?

Some images have usage restrictions that disable these features. To enable them, type bash in the Terminal and press Enter.

Q: What if the instance specifications are insufficient?

You can update the DSW instance specifications by following these steps:

1. In the DSW instance list, click the instance name to open the instance details page.

2. On the Instance Settings tab, click Change Settings.

3. In the Change Instance Settings panel, update the instance specifications.

   Note

   Updating the specifications of a running instance will cause it to restart immediately. Ensure that you have saved your work before proceeding.

Q: How do I free up memory?

If high memory usage is affecting your work, you can resolve the issue in one of two ways.

• If the Terminal is unresponsive due to high memory usage, click Stop Instance in the upper-right corner. Alternatively, you can return to the DSW console and click the Stop button for that instance. Wait for the instance to stop before you restart it.

• If the Terminal is still responsive, you can run thetop command to view process memory usage. The %MEM column shows the percentage of memory used, and PID is the process ID.

  To terminate a memory-intensive process, run the following command:

  kill PID

  Replace PID with the process ID of the process you want to terminate. After you run the command, memory usage decreases.

Q: RuntimeError: CUDA error: too many resources requested for launch

Cause: This error occurs when a CUDA kernel requests more resources than are available, which is typically caused by GPU hardware limitations.

Solution: Try restarting the instance and running the program again. If the problem persists, select a GPU instance with higher specifications.

Q: Can I create swap space if the instance runs out of memory?

No. DSW runs in a container and does not support creating or managing swap space.

This is for the following reasons:

• Permission restrictions: The kernel permissions of the container are restricted, which prevents it from mounting a swap file. You cannot bypass the resource policies of the host even if you obtain root permissions in the container.

• Platform policy: The platform schedules and restricts resources to ensure the stability and security of the multitenancy environment.

Suggestion: If memory is insufficient, optimize your code or upgrade your instance specifications.