Hi Misc,
For number of years I had a very simple scheme to backup my OpenBSD infrastructure servers running critical network services for our small university lab. Namely, I would put a low profile usb flash drive and use /altroot facility in the daily(8) scripts to backup root partition to it as described in FAQ https://www.openbsd.org/faq/faq14.html#altroot I started doing that many years ago, before sysupgrade was available. It worked like a charm. Once sysupgrade became available I noticed that it would get confused by an extra disk in the server. My "solution" was to remove usb drive before running sysupgrade and that worked OK until Covid 19 when the physical access to my servers became more challenging. I had a quick look at the sysupgrade.sh script http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/usr.sbin/sysupgrade/sysupgrade.sh?rev=1.40&content-type=text/x-cvsweb-markup and I have to admit that it is not clear to me how the target disk for the installation is picked. I completely understand that sysupgrade is designed not to be configurable in order to be foolproof. I am curios how the other people adapted their use of /altroot facility in the era of sysupgrade. Obviously people who use full disk encryption or RAID 1 are facing similar "issues". For me personally those are of secondary importance as I always have the access to my laptop where the disk is fully encrypted and I always have the access to my desktop where softraid RAID 1 mirror is used for the OS. Most Kind Regards, Predrag Punosevac
