On Fri, Jan 24, 2025 at 1:07 PM Christopher Huhn <c.h...@gsi.de> wrote:
> > sda 8:0 0 7.3T 0 disk > > sdb 8:16 0 7.3T 0 disk > > sdc 8:32 0 447.1G 0 disk > > sdd 8:48 0 447.1G 0 disk > as you can see here the large disks are sda (disk1) and sdb (disk2) > while the SSDs are sdc and sdd. > This mapping is not stable through reboots, so you have to make sure > your disks are in the intended order in your disklist. > Ooof. Embarrassing. I'm aware it's unstable, but it had been so stable in my previous tests I stopped paying attention to that. Thanks for the second pair of eyes. Unfortunately though, that doesn't fix my problem. After correcting for that, I'm still seeing the same LVM error. > We use a very basic hook to sort our disks by size: > > > # cat class/31-sortdisks.sh > > disklist=$(grep -H '^[1-9]' /sys/block/[hsv]d[a-z]/size > /sys/block/nvme*n[0-9]/size | sort -n -t : -k 2 | cut -f4 -d / | xargs) > > echo "disklist=\"$disklist\"" >> "$LOGDIR/additional.var" > This looks REALLY helpful! Thanks! I'm making one small modification, using sort's -s modifier... in the (likely futile) hope that the drives might still be in bus order. :)
