SQLite WAL & Safe Backups
Write-ahead logging pragmas, busy timeouts, checkpoint starvation limits, and atomic online snapshots without halting readers.
Open SQLite checklist →Concise production defaults, verification commands, and subtle edge-case notes distilled from everyday Linux host maintenance and embedded database tuning.
Write-ahead logging pragmas, busy timeouts, checkpoint starvation limits, and atomic online snapshots without halting readers.
Open SQLite checklist →Immutable build asset headers, conditional revalidation with ETag and 304 responses, and avoiding mixed Vary keys.
Open HTTP caching guide →Persistent timers with jitter, MemoryMax boundaries, and read-only filesystem protections audited via systemd-analyze security.
Open systemd notes →Common low-overhead commands used before reaching for heavier profiling harnesses:
| Question | Command | Signal to Inspect |
|---|---|---|
| Which process holds open deleted files? | lsof +L1 /var | Large deleted logs still consuming unlinked inodes |
| Are timers drifting or missed after reboot? | systemctl list-timers --all | LAST vs NEXT intervals and Persistent=true wakeups |
| Which unit consumed recent disk writes? | systemd-cgtop --order=io | Sustained throughput per cgroup slice |
| Verify SQLite integrity offline | sqlite3 db.sqlite "PRAGMA quick_check;" | Returns single line ok across all B-tree pages |
Prefer declarative operating system primitives over bespoke watchdog scripts. Every configuration line documented in these three topic notes relies strictly on built-in kernel, SQLite, RFC 9111 HTTP semantics, and standard service manager features.