Three dominant themes fromthe discussion
| Theme | Key takeaway | Representative quote |
|---|---|---|
| 1. Ballast/reserve files – keeping a small, non‑essential file(s) on each system to buy time when disks run low. | Allows the OS to keep writing lock files or other temporary data while admins clean up or add capacity. | flanfly: “A neat trick I was told is to always have ballast files on your systems. Just a few GiB of zeros that you can delete in cases like this.” |
| 2. Alarm fatigue & inadequate threshold alerts – simple %‑based warnings often miss fast‑filling disks and can be ignored or cause false confidence. | Alerts must be reliable, low‑noise, and preferably predictive rather than just “> 50 % used”. | evil‑olive: “Surely a 50% warning alarm on disk usage covers this without manual intervention?” … “a ‘warning alarm’ is a terrible concept, in general. it’s a perfect way to lead to alert fatigue.” |
3. Proactive monitoring with lightweight tools & quotas – using tools like gdu, ncdu, or system‑level quotas/reservations to detect and remediate space pressure early. |
Early warning lets you act before a full disk brings services down. | Neil44: “I also discovered gdu recently. It's really good. It saves me running du -h --max-depth=1 | sort -h a million times trying to find where the space has gone while you're stressing about production being down.” |
These three themes capture the community’s consensus on how to avoid being caught off‑guard by disk‑space exhaustion.