On Sat, Jun 26, 2021 at 06:49:17AM +0200, Martin Husemann wrote: > Also any reason to use 9.1 instead of 9.2 or 9.2_STABLE? > (Not that I think it would make a difference for azalia)
Practical reason: I start to update the node I'm doing my main programing/developing work on and I then, after having verified that things are rolling and with some delay---specially if the node is a remote production server that it is not possible to update easily and for safety only when I have physical access to it in case of problem (this time: there was)---I put other nodes in sync to not have to cross-compile between NetBSD versions. When I updated the developing node, NetBSD was at 9.1. Since, for what I know (not much), virtualization always(?) present a defined common pseudo-hardware interface, I imagine that there is no virtualization that will allow to test a kernel in a VM, with access to an image of the real hardware present, so that one can verify that a tentative kernel will run on the actual hardware before switching kernels? I have still to verify that an UEFI bootloader will allow to implement by scripting a "boot once", so that if a new kernel (on a remote host) crashes, it reboots with a kernel that is known to work. It is probably possible to implement this with the existence of persistent storage of UEFI variables. -- Thierry Laronde <tlaronde +AT+ polynum +dot+ com> http://www.kergis.com/ http://kertex.kergis.com/ http://www.sbfa.fr/ Key fingerprint = 0FF7 E906 FBAF FE95 FD89 250D 52B1 AE95 6006 F40C