What data for this category looks like
For FastGPT v4.17.0 deployment scenarios, declaration materials and case review data for mineral mining and processing primarily originate from geological exploration reports, mining license application documents, mineral processing process records, safety supervision department inspection case files, and monthly production reports. Data updates fall into two categories: new mining project declaration materials are submitted in one-time batches, while daily case files are updated weekly or monthly. Documents include structured fields such as mining license number, mining area size, and annual ore processing capacity, plus unstructured text such as safety assessment reports and process descriptions. Attachments mostly use formats including PDF and CAD drawings. Fields must match standard units used in the mining industry.
What constraints do these characteristics impose during deployment and upgrade
Multi-format, large-volume attachments require configuring object storage services with large capacity support during deployment to avoid upload failures. Complex associated queries for structured fields require separate relational database instances to prevent resource contention with other services. Batch-imported declaration materials and frequently updated daily case files require reserving incremental synchronization configuration entries during upgrades. Mixed storage of long-text reports and CAD drawings requires adjusting the maximum length limit of database indexes to prevent index failure.
How to configure settings
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
STORAGE_S3_ENDPOINT | Adapt to local object storage or cloud service S3-compatible endpoints, set single-bucket storage limit to 1000 GB | Mining declaration materials include large-volume CAD drawings and long-text reports, requiring large-capacity, highly reliable storage support |
PG_URL | Deploy a separate PostgreSQL instance, set maximum connection pool size to 20–30 | Structured approval fields require complex associated queries, separate instances prevent resource contention with other services |
DB_MAX_LINK | 800–1200 characters | Structured fields of a single declaration material have a large combined length, requiring adaptation to long-text index requirements |
REDIS_URL | Configure persistent storage, set expiration time to 7200 seconds | Contextual association for case review requires temporary caching of session data to prevent session loss |
FILE_TOKEN_KEY | Use a randomly generated 32-bit string as the key | Secure verification of upload and download permissions for declaration material attachments is required to prevent unauthorized access |
Three common mistakes
- Phenomenon: A "connection refused" error occurs during docker-compose deployment, and the service fails to start. Cause: Connection address and port for
MONGODB_URIorPG_URLare not configured correctly, preventing the service from initializing database connections. - Phenomenon: The knowledge base cannot support continuous questioning, and the second question produces irrelevant answers. Cause: Persistence parameters for
REDIS_URLare not configured correctly, leading to expiration or loss of session context cache, preventing association with historical questions. - Phenomenon: Locally modified code cannot be synchronized to the online FastGPT instance, and no changes appear on the page. Cause: The frontend domain name for
FE_DOMAINis not configured correctly, and the built image push and restart process is not executed, leading to the online service not loading the latest code.
How to confirm correct configuration
- Log in to the FastGPT backend system settings page, check storage configuration items, and confirm that the address of
STORAGE_S3_ENDPOINTmatches the currently deployed object storage service. - Upload a CAD drawing or long-text report commonly used for mining declarations, verify that upload, storage, and download processes work normally with no error prompts.
- Initiate two consecutive case review-related questions, confirm that the second question can associate context from the first question, with expected output.
- Check service runtime logs, confirm that connections for
PG_URLandMONGODB_URIare normal, with no timeout or connection refused error messages.
The values provided are common starting points, and should be evaluated against the reader's own samples.
Question material comes from public community discussions. Configuration values are common starting points and should be measured against your own samples. Verified on 2026-09-28.