Large Projects Crashes Import From Spreadsheet

by ADMIN 47 views

Introduction

In the realm of data ingestion, large projects often pose significant challenges. One such issue is the crash of import from spreadsheet, resulting in incomplete or partially imported projects. This article delves into the details of this critical problem, exploring its causes, affected projects, and potential solutions.

Describe the Bug

While uploading a large spreadsheet (i.e., with > 10K biomaterials), ingestion at one point stops, leading to a 104 error and/or a partially imported project (missing any entity or linking depending on when it stopped). This issue is particularly concerning, as it can result in the loss of valuable data and hinder the progress of ongoing projects.

To Reproduce

To replicate this behavior, follow these steps:

Step 1: Go to Ingest

Navigate to the ingest page at https://contribute.data.humancellatlas.org/.

Step 2: Click on "Register a Project"

Click on the "Register a project" button to initiate the project registration process.

Step 3: Use the Provided Spreadsheet

Use the example spreadsheet provided at https://docs.google.com/spreadsheets/d/1h1qVfzqUMws1vXDS5q_V3Tf8i76VLc3H/edit?usp=drive_link&ouid=101176945353039944995&rtpof=true&sd=true.

Step 4: Wait for Ingest to Crash

Wait for the ingestion process to crash, resulting in an error message and/or an incomplete submission.

Step 5: Observe the Error Message

Observe the error message, which typically indicates a connection reset by peer (104 error).

Expected Behaviour

The expected behavior is for all entities and linkings to be complete and uninterrupted, ensuring a seamless and successful project import.

Affected Projects

The following projects are affected by this issue:

HeartReconstructionPostHF

dcp1514250681821643

Fetal/ Maternal Interface

Environment

The issue is observed in the production environment.

Browser

The issue is observed in Chrome.

Potential Solutions

To address this critical issue, the following potential solutions can be explored:

1. Optimize Ingestion Process

Optimize the ingestion process to handle large spreadsheets more efficiently, reducing the likelihood of crashes and errors.

2. Implement Load Balancing

Implement load balancing to distribute the ingestion workload across multiple servers, ensuring that no single server becomes overwhelmed and crashes.

3. Increase Server Resources

Increase the server resources (e.g., CPU, memory, and storage) to handle large projects and prevent crashes.

4. Develop a Backup System

Develop a backup system to ensure that data is not lost in case of a crash or error during ingestion.

5. Provide Real-time Feedback

Provide real-time feedback to users during the ingestion process, allowing them to monitor the progress and identify potential issues early on.

Introduction

In our previous article, we discussed the critical issue of large projects crashing import from spreadsheet, resulting in incomplete or partially imported projects. In this article, we will address some of the frequently asked questions (FAQs) related to this issue, providing clarity and insights to help users better understand the problem and its potential solutions.

Q&A

Q1: What is the cause of the crash during large project import?

A1: The crash during large project import is primarily caused by the ingestion process becoming overwhelmed by the sheer volume of data, leading to a connection reset by peer (104 error).

Q2: What are the affected projects, and how can I identify them?

A2: The affected projects are HeartReconstructionPostHF, dcp1514250681821643, and Fetal/ Maternal Interface. You can identify them by checking the project IDs and ingest links provided in our previous article.

Q3: What is the expected behavior during large project import?

A3: The expected behavior is for all entities and linkings to be complete and uninterrupted, ensuring a seamless and successful project import.

Q4: How can I prevent the crash during large project import?

A4: To prevent the crash during large project import, you can try the following:

  • Optimize the ingestion process to handle large spreadsheets more efficiently.
  • Implement load balancing to distribute the ingestion workload across multiple servers.
  • Increase server resources (e.g., CPU, memory, and storage) to handle large projects.
  • Develop a backup system to ensure that data is not lost in case of a crash or error during ingestion.

Q5: What are the potential solutions to address this issue?

A5: The potential solutions to address this issue include:

  • Optimizing the ingestion process.
  • Implementing load balancing.
  • Increasing server resources.
  • Developing a backup system.
  • Providing real-time feedback to users during the ingestion process.

Q6: How can I report issues related to large project import?

A6: To report issues related to large project import, please contact our support team at support@humancellatlas.org. We will do our best to assist you and provide a resolution to the issue.

Q7: What is the estimated timeline for resolving this issue?

A7: We are working diligently to resolve this issue as soon as possible. However, the estimated timeline for resolution is dependent on various factors, including the complexity of the issue and the availability of resources.

Q8: How can I stay updated on the progress of resolving this issue?

A8: To stay updated on the progress of resolving this issue, please follow our social media channels (e.g., Twitter, Facebook, and LinkedIn) or sign up for our newsletter to receive regular updates.

Conclusion

In conclusion, the issue of large projects crashing import from spreadsheet is a critical problem that requires immediate attention. By understanding the causes, affected projects, and potential solutions, users can take proactive steps to prevent the crash and ensure a seamless and successful data ingestion experience. We will continue to work tirelessly to resolve this issue and provide regular updates to our users.