Hi, On 05-03-2025 22:06, Mathias Gibbens wrote:
For the past day-ish or so I've been seeing arm64 CI runs fail[1] on various hosts due to disk space exhaustion. This seems to be a new issue, and limited just to the arm64 CI hosts. Has some config recently changed for the CI hosts?
It's due to a mix of things. I have configured /tmp on tmpfs nowadays on arm64, but apparently my service file manipulations haven't ensured yet that the necessary steps are *always* taken in case of a reboot. The reboot happened because of a unattended-upgrades which we installed everywhere several months ago.
So yes, recent changes, but not so recent. The cause is a bug. I think (well, hope) I fixed it today.
Paul
Attachment:
OpenPGP_signature.asc
Description: OpenPGP digital signature