Skip to content

linux-service-setup.md

Latest commit

 

History

History
87 lines (70 loc) · 3.11 KB

linux-service-setup.md

File metadata and controls

87 lines (70 loc) · 3.11 KB
Raw

Linux Service Setup

This document is legacy planning material for a possible future Linux service. It is not the supported production scheduling mode for PROD-001. Supported production scheduling is cron or manual one-cycle execution of the packaged Python command documented in the production operator runbook.

This document is a non-installing template for operator planning. It does not install a service, enable a service, start a service, create a user, read configuration, load credentials, connect to PostgreSQL, run SQL, or download sources. It must not be used to imply daemon, system service, or installer support for production operation. The .NET runtime has a directly runnable scheduled-worker entrypoint baseline, but its run command remains a not-yet-implemented placeholder and does not make .NET production-ready. Implementation-specific service files should be added only after a future task explicitly scopes and reviews service support for the selected implementation.

Service Responsibilities

A future Linux service setup would need to define:

  • Working directory
  • Environment variables
  • Database configuration source
  • Raw archive path
  • Service user
  • Restart policy
  • Logging destination
  • Basic service management commands

Conceptual systemd Unit

The example below is conceptual and not a supported production unit. The exact command would depend on a future reviewed service implementation. Until that implementation exists, keep ExecStart as an operator-owned placeholder and use cron or manual one-cycle execution for supported production scheduling. Do not point production systemd units at the PROD-003 .NET run-once placeholder for real ingestion.

[Unit]
Description=CarbonOps-Parser background ingestion service
After=network.target

[Service]
WorkingDirectory=/opt/carbonops-parser
EnvironmentFile=/etc/carbonops-parser/runtime.env
ExecStart=/opt/carbonops-parser/<approved-runtime-entrypoint>
Restart=on-failure
User=carbonops
Group=carbonops
KillSignal=SIGTERM
TimeoutStopSec=300

[Install]
WantedBy=multi-user.target

The environment file path above is illustrative. It must be created and managed outside the repository and must not be committed with environment-specific values.

Management Commands

Typical service management commands after a future reviewed unit is installed:

sudo systemctl daemon-reload
sudo systemctl start carbonops-parser
sudo systemctl status carbonops-parser
sudo journalctl -u carbonops-parser -f
sudo systemctl stop carbonops-parser

Notes

Linux service documentation must explain how configuration is provided, where raw files are archived, how logs are reviewed, and how graceful stop is handled. Do not add automatic enablement, destructive cleanup, branch or worktree cleanup, schema deletion, or ad hoc database mutation to service management steps.

For the supported production install, configure, validate, run, stop, diagnose, and rollback flow, see Production Packaging And Operator Runbook. For project-level Python/.NET parity requirements, see Production Parity Contract.