On Mon, Jul 11, 2016 at 06:09:03AM +0000, Dmitry Mishin wrote: > If you have used automatic partitioning, than we've created /boot > partition and installed boot loader on every selected disk, which could be > a reason of "boot worked without problems". This is our improvement over > default RHEL behaviour. Need to check why the installer complains still.
In what ways is this an improvement? (I am not saying it is not. I am undecided on this, and am interested in your reasoning.) I test-installed a system from vz-iso-7.0.0-3719.iso last week, and I did notice this detail - and it looked weird to me. By default, with two disks the system has a 1 GB software RAID-1 for /boot, but has no redundancy for the rest, using one disk for host system and the other for /vz (IIRC, the installer said it'd be for "data"). I don't see much use in having /boot survive failure of the disk hosting the system's root fs - this will require recovery with hardware intervention (or at least virtual media or network boot) regardless of whether /boot survives or not. No? So is this improvement solely to workaround issues where the installer and the machine's firmware have different opinion on which disk the system should boot from? I think it would be far more useful to offer an optional installer-supported way to install the whole system on software RAID-1. With only two disks, offer /vz to be on the same two disks as well (same or separate RAID-1 and fs). Alexander _______________________________________________ Users mailing list Users@openvz.org https://lists.openvz.org/mailman/listinfo/users
