You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I've been building a Void-ish system using a read-only squashfs root filesystem. Unfortunately, the script /etc/core-services/03-filesystems.sh, at line 67, drops to an emergency shell if root cannot be mounted read-write.
This makes sense for the majority of desktop and server use cases, but for some embedded devices and some desktops/servers, having a read-only rootfs makes sense.
Proposal: have core-services/03-filesystems.sh check if /proc/cmdline contains ro, or readonly, or ro=true, or some such string, to support a wider range of use cases.
My temporary solution is to just remove the || emergency_shell from line 67, but this is undesirable because I wish to stay as close to upstream as possible
The text was updated successfully, but these errors were encountered:
ro is a default kernel cmdline flag for majority of setups though, it mounts the root read-only and then it's the target system's job to remount it as rw.
Greetings!
I've been building a Void-ish system using a read-only squashfs root filesystem. Unfortunately, the script /etc/core-services/03-filesystems.sh, at line 67, drops to an emergency shell if root cannot be mounted read-write.
This makes sense for the majority of desktop and server use cases, but for some embedded devices and some desktops/servers, having a read-only rootfs makes sense.
Proposal: have core-services/03-filesystems.sh check if /proc/cmdline contains
ro
, orreadonly
, orro=true
, or some such string, to support a wider range of use cases.My temporary solution is to just remove the
|| emergency_shell
from line 67, but this is undesirable because I wish to stay as close to upstream as possibleThe text was updated successfully, but these errors were encountered: