Series
A series about operating the infrastructure underneath everything else – the migrations, the failures, and the checks that turned out to be pointed at the wrong thing.
The special vdev of a 113 TB pool sat on two qcow2 files – both on the same consumer SSD, with 11% of its endurance left. Moving it meant operating on a live mirror with no room to be wrong. And it surfaced two checks that had been pointed at the wrong thing the whole time.
Our pre-push gate took almost six minutes. The container read the repository through a bind mount, and that alone accounted for 324 of the 357 seconds.
Codex reported success in write mode although no file appeared. The fault was in the Linux sandbox, but the lasting lesson is about our gate.
From home, every mail port ran into a timeout while HTTPS on the same host answered normally. The cause was a second [DEFAULT] ignoreip in jail.d that replaces FAIL2BAN_IGNOREIP from docker-mailserver.
