What the data for this category looks like
Semiconductor wafer fabrication supplier and process quality audit data typically comes from MES systems, equipment FDC modules, defect detection and metrology equipment, wafer map files, process recipe libraries, and experimental datasets. Data update frequencies include per wafer batch generation, real-time synchronization during equipment operation, and daily batch aggregation. Daily processed documents include process flow cards, batch run records, equipment recipe parameters, SPC statistical charts, exception handling sheets, defect maps, and yield reports. Fields include batch identifiers, process parameter values, defect position coordinates, and equipment runtime durations. Units include nanometer-scale units such as nanometers, and commonly used industrial units like hours.
What constraints these characteristics impose during deployment and upgrade
Data for this category is multi-source, heterogeneous, and has highly varied update rhythms. Deployment requires configuration adjustments to support integration with multiple systems and real-time synchronization. Nanometer-scale process parameters and large-volume wafer map files require storage configurations with sufficient bandwidth and storage space. Batch data volume per wafer batch increases with process complexity, so database connection pool parameters must be adjusted to avoid connection overflow. During upgrades, compatibility logic for historical batch run records and defect maps must be retained to avoid interrupting ongoing wafer production processes. Additionally, limited rework windows require that deployment and upgrade operations be completed during non-production windows, so offline upgrade plans must be configured.
How to configure settings
The following configurations apply to FastGPT v4.17.0.
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
DB_MAX_LINK | 80–120 connections | Adapts to concurrent access demands for batch data per wafer batch and real-time equipment data. This range covers synchronization scenarios for most wafer fabrication facilities. |
REDIS_URL | redis://default:custom password@redis host:6379/1 | Caches real-time equipment FDC data and audit session information. Uses database 1 to isolate business data. |
STORAGE_S3_ENDPOINT | Access endpoint corresponding to the enterprise object storage | Stores large-volume wafer maps, defect maps, and process documents. Supports upload and read operations for multi-source files. |
| FE_DOMAIN | wafer-quality-audit.enterprise domain | Configures a secondary domain name to distinguish the service, and resolves secondary domain name access issues for Docker deployments. |
PG_URL | postgresql://username:password@database host:5432/fastgpt_wafer?sslmode=require | Stores structured yield reports and exception handling sheets. Ensures transmission security for sensitive process data. |
Three common misconfigurations
- Symptom: After Docker deployment, the service cannot be accessed via the configured secondary domain name. Logs show 404 errors or domain name resolution failures. Cause: The
FE_DOMAINparameter was not configured correctly, or the secondary domain name was not pointed to the public IP of the deployment server in the domain name resolution service provider's console. - Symptom: Containers fail to start after upgrading to a new version, with prompts about missing dependency packages or incompatible configuration items. Cause: Original
REDIS_URLandPG_URLconfiguration information was not retained, or running batch data synchronization tasks were not stopped during the upgrade, resulting in overwriting of configuration files. - Symptom: Text chunks split from the knowledge base contain overly long nanometer-scale process parameters, or multiple merged segments still do not meet reasonable length requirements. Cause: Chunking rules were not adjusted for the field characteristics of wafer manufacturing documents. Generic document chunking configurations were used directly, without matching the actual length characteristics of documents in this category.
How to verify configurations are correct
- Access the configured secondary domain name, check if the service login page loads normally, to verify that the
FE_DOMAINconfiguration takes effect. - Connect to the configured
PG_URLandREDIS_URL, run basic read and write tests, to confirm that database and cache service connections are working properly. - Upload a wafer process flow card and a defect map, check if the files can be stored and read normally, to verify that the
STORAGE_S3_ENDPOINTconfiguration takes effect. - Check the database connection pool status, confirm that the current number of connections does not exceed the value set for
DB_MAX_LINK, to verify that the concurrent access configuration is reasonable.
The values provided are common starting points and should be measured against the reader's own samples.