SYSTEM MESSAGE: WELCOME TO THE MANTiS DIRECTORY LISTING. IF IT IS YOUR FIRST VISIT, CHECK OUT THE
FAQ BEFORE POSTING.
MEMBERS: 6 |
THREADS: 20 |
POSTS: 70 · NEWEST:
BOOT_SECTOR
OCTOBER UPTIME REPORT
POSTED BY
ADMIN ON 11-01-2026 07:42 PM
6 REPLIES, 68 VIEWS
Closing the books on October.
Total uptime: 99.84%. The 21 minutes of downtime were the backup migration on the 18th — old host to new host, WAL file transferred cleanly, no post rollback.
Total posts this month: 34. Up from 11 in September. Three new members: kbdwarrior, boot_sector, and vector_eye (who joined in late September but started posting in October).
Plan for November: switch WAL checkpointing from PASSIVE to TRUNCATE on a 03:00 cron. Should keep the WAL file from bloating past 4 MB as post volume grows.
Numbers and the migration script are below. Questions welcome.
// POSTS
Confirmed — 99.84% for October. The 21 minutes of downtime were the backup migration on the 18th. Old host to new host, WAL file transferred cleanly, no post rollback. Backup script ran twice on the 19th by accident, no harm done.
POSTED BY
SYSOP ON
11-01-2026 09:14 PM
Solid numbers for SQLite on a single host. Hats off. Most people assume you need Postgres for this kind of uptime.
POSTED BY
PLAYER1 ON
11-02-2026 08:50 AM
What is the WAL checkpoint interval right now? That is usually where I would tune next — PASSIVE lets the WAL grow without bound under write pressure.
PASSIVE right now. Considering TRUNCATE on a cron at 03:00 local. The WAL is sitting at about 1.8 MB after a week, which is fine, but I want a hard ceiling before it becomes a problem. Thoughts?
POSTED BY
ADMIN ON
11-02-2026 06:09 PM
TRUNCATE at 03:00 is the safe bet. Just make sure your backup job runs AFTER the checkpoint, or you will snapshot a half-written WAL. The order matters: checkpoint first, backup second, verify checksums third.
Three weeks in on the TRUNCATE schedule. Clean. No WAL bloat. Backup job runs at 03:05, finishes at 03:06, checksum verifies on the 03:10 cron. Has not missed a night yet.