On Thu, May 03, 2018 at 10:53:00AM +0200, richard lucassen wrote: > > > There are multiple reports on #ganeti that this update breaks > > > networking in certain circumstances, probably multiple tun/tap > > > device configurations. No more details or a proper bug report yet > > > as I haven't experienced this myself, but mentioning in case it > > > saves anyone else breakage.[...] > > > > I believe I understand this. Creating a tun/tap device using a name > > pattern such as "tun%d" (or empty name) will now fail if the number > > substituted is not 0. There is an upstream fix for this that I failed > > to spot in time. > > There is also an big increase in time before random is initialized: > > [ 182.811840] random: crng init done > > This is a machine on bare metal. On other environments like proxmox I've > seen: > > [ 303.993638] random: crng init done > > Downgrading to the previous kernel resolves the problem (normally a few > seconds). One of the consequences is that openntpd (or a program like > rdate) hangs until the crng is initialized.
I'd think it's a fix for [1], [2] but it does not appear on the list of CVEs fixed. 1. https://security-tracker.debian.org/tracker/CVE-2018-1108 2. https://bugs.chromium.org/p/project-zero/issues/detail?id=1559