On 18 March 2012 10:51, Pascal Terjan <[email protected]> wrote: > Well unionfs seemed to get different oops in each new version. Anyway > I don't think it should be done like this (having specific code for > unionfs everywhere). > I want to improve current chroot mode (like using a cgroup to easily > kill all processes or set limits) and have a common api to allow other > modes, like building in kvm for example, or using lvm/btrfs rw > snapshots (which should be better supported than unionfs).
For the record, I saw the need for cgroup this weekend. iurt got killed when log file becames too big or if free space is below 99% but the forked commands still run... See http://svnweb.mageia.org/soft?view=revision&revision=3803
