CGI Script Repository & Execution Environment

Manage, monitor, and execute legacy CGI scripts for the archive preservation pipeline. All scripts run in isolated chroot environments.

Scripts
Execution Console
Logs
Documentation
Script Path Status Last Run Action
fetch-archive /cgi-bin/archive/fetch.cgi Active 2 min ago
render-static /cgi-bin/render/static.cgi Queued 14 min ago
geo-cities-import /cgi-bin/import/geocities.cgi Error 1 hr ago
link-graph /cgi-bin/analysis/links.cgi Active 5 min ago
backup-snapshot /cgi-bin/maintenance/snap.cgi Idle 1 day ago
cgi-exec@archive-node-01
[root@cgi-env]# systemctl status cgi-runner
● cgi-runner.service - 1990 Web Archive CGI Execution Daemon
Active: active (running) since Mon 2024-01-15 08:30:00 UTC
Tasks: 14 Memory: 248M CGroup: /system.slice/cgi-runner.service
[root@cgi-env]# queue --list --pending
[2] render-static.cgi | ETA: 00:04:12
[3] link-graph.cgi | ETA: 00:01:45
[root@cgi-env]# execute --script fetch-archive --mode deep
  • 10:42:18INFOfetch-archive: Successfully retrieved 412 pages from Wayback mirror
  • 10:40:05WARNrender-static: CSS parser fallback triggered for Netscape 4.0 layout
  • 10:38:51ERRgeo-cities-import: Timeout after 120s on /geocities/Athens/2481/
  • 10:35:22INFOlink-graph: Completed crawl of 1.2k nodes, 4.8k edges
  • 10:30:00INFOSystem: Daily integrity check passed. 0 corrupted archives detected.
  • 09:55:14WARNbackup-snapshot: Disk usage at 78%. Throttling writes.
  • 09:12:08INFOfetch-archive: MIME type detection updated for .hqx and .sit files
  • 08:45:33ERRrender-static: Failed to parse frameset in legacy document

CGI Path Conventions

All executable scripts must reside under /cgi-bin/ or its subdirectories. Relative paths in scripts are resolved against the chroot environment at /var/cgi-root/.

Environment Variables

  • ARCHIVE_BASE → /mnt/archive-1990
  • RENDER_MODE → exact|modern|hybrid
  • LOG_LEVEL → debug|info|warn|err
  • TIMEOUT → 120 (seconds)
\n

Execution Limits

Scripts are subject to ulimit -t 300 (CPU seconds) and ulimit -v 512000 (virtual memory MB). Long-running tasks must use the --daemonize flag and register with the queue manager.

Security Notes

All CGI processes run under the www-cgi user with read-only access to archive volumes. Network access is restricted to approved archival mirrors and DNS resolvers. Shell execution is disabled by default.

"}