The following table lists the software required to use Fujitsu Enterprise Postgres.
No. | Product name | Version | Fujitsu Enterprise Postgres product name | Remarks | |
|---|---|---|---|---|---|
AE | SE | ||||
1 | PRIMECLUSTER (*1) | 4.7A00 or later (*1) | Y | Y | Mandatory when performing failover operation integrated with PRIMECLUSTER. |
2 | Oracle Call Interface (OCI) libraries | 19c or later | Y | Y | Required when using oracle_fdw. |
3 | Amazon Web Services Command Line Interface (AWS CLI) | 2.x or later | Y | Y | Required for monitoring database operations using Amazon CloudWatch. Required for Autoscale of Replica Servers through Amazon EC2 Auto Scaling Integration. (*2) |
4 | amazon-cloudwatch-agent | v1.300064 or later | Y | N | Required for Autoscale of Replica Servers through Amazon EC2 Auto Scaling Integration. |
5 | amazon-ssm-agent | 3.3 or later | Y | N | Required for Autoscale of Replica Servers through Amazon EC2 Auto Scaling Integration. |
6 | MCP Toolbox for Databases | v0.27.0 or later | Y | Y | Required when using MCP server feature for business data utilization. |
Y: Can be used
N: Cannot be used
*1: The applicable products are shown below:
PRIMECLUSTER Enterprise Edition
PRIMECLUSTER HA Server
*2: This feature can only be used in AE.
If you want to use the semantic text search and automatic vectorization (including standalone use of model management within the database), you need the following Python packages:
Install to a location other than the default Python package installation destination (purelib) and set the environment variable PYTHONPATH.
Example of verifying the default Python package installation destination (purelib)
# python -c "import sysconfig;print(sysconfig.get_path('purelib'))"
/usr/local/lib/python3.11/site-packagesExample of installing Python packages using the pip command
pip install <package-name>==<version> -t <python package-install-destination>
Setting PYTHONPATH
export PYTHONPATH=<python package-install-destination>:$PYTHONPATH
Package name | Remarks |
|---|---|
anthropic | Install 0.44.0. |
backoff | Install 2.2.1. |
boto3 | Install 1.35 or higher and 2.0 or lower. |
click | Install 8.0 or higher and 9.0 or lower. |
cohere | Install 5.13.11. |
datadog_lambda | Install 6.9 or higher and 7.0 or lower. |
datasets | Install 3.1.0. |
google-cloud-aiplatform | Install 1.78.0 or higher and 2.0 or lower. |
langchain-openai | Install 0.1 or higher and 1.0 or lower. |
langchain-text-splitters | Install 0.3.5 or higher and 1.0 or lower. |
litellm | Install 1.58.2. |
ollama | Required when using Ollama as the embedded provider. |
openai | Required when using OpenAI as the embedded provider. |
pgvector | Python library for handling the pgvector data type, separate from the pgvector Fujitsu Enterprise Postgres extension. |
pydantic | Install 2.0 or higher and 3.0 or lower. |
python-dotenv | Install 1.0 or higher and 2.0 or lower. |
pytimeparse | Install 1.1 or higher and 2.0 or lower. |
structlog | Install 24.0 or higher and 25.0 or lower. |
tiktoken | Install 0.7 or higher and 1.0 or lower. |
tritonclient[grpc] | Install 2.20.0 or higher. When performing vectorization using embedded models imported into the database, use the same version as the Triton Inference Server you set up as the inference server. |
typing_extensions | Install 4.0 or higher and 5.0 or lower. |
voyageai | Required when using VoyageAI as the embedded provider. |
The following table lists client that can be connected to the Fujitsu Enterprise Postgres server feature.
OS | Product name |
|---|---|
Windows Linux | Fujitsu Enterprise Postgres Client 18 or later |
The following table lists server assistant that can be connected to the Fujitsu Enterprise Postgres server feature.
OS | Product name |
|---|---|
Linux Windows | Fujitsu Enterprise Postgres Server Assistant 18 or later |