Bug#567468: please set hostname via sysctl early on

2010-01-29 Thread martin f krafft
Package: initramfs-tools Version: 0.93.4 Severity: wishlist mdadm in incremental mode needs the hostname to be able to work without a configuration file, which would make RAID-boot much more robust. Instead of mdadm touching /proc/sys/kernel/{host,domain}name, it feels like initramfs should do

Bug#567468: please set hostname via sysctl early on

2010-01-29 Thread Marco d'Itri
On Jan 29, martin f krafft madd...@debian.org wrote: mdadm in incremental mode needs the hostname to be able to work without a configuration file, which would make RAID-boot much more robust. Instead of mdadm touching /proc/sys/kernel/{host,domain}name, it feels like initramfs should do

Bug#567468: please set hostname via sysctl early on

2010-01-29 Thread martin f krafft
also sprach Marco d'Itri m...@linux.it [2010.01.30.0233 +1300]: mdadm in incremental mode needs the hostname to be able to work without a configuration file, which would make RAID-boot much more robust. Instead of mdadm touching /proc/sys/kernel/{host,domain}name, it feels like initramfs

Bug#567468: please set hostname via sysctl early on

2010-01-29 Thread Marco d'Itri
On Jan 29, martin f krafft madd...@debian.org wrote: What happens if the hostname is changed but the initramfs is not rebuilt? How likely is that to happen? How likely is it to happen that the hostname is changed, an array built (thus carrying the new hostname), and the initramfs not

Bug#567468: please set hostname via sysctl early on

2010-01-29 Thread martin f krafft
also sprach Marco d'Itri m...@linux.it [2010.01.30.0839 +1300]: What happens if the hostname is changed but the initramfs is not rebuilt? How likely is that to happen? How likely is it to happen that the hostname is changed, an array built (thus carrying the new hostname), and the