If a RAID0 mdadm array is created with a mix of SATA and NVMe drives, is the performance of the array limited to SATA speed? Also, when I destroyed and recreated my existing array, I noticed that I created it using ext4 partitions on the disks as the raid devices instead of the disks themselves. Would this have had any performance impact on the array?
After recreating my array with the devices themselves instead of partitions, it fails to mount after rebooting, but if I run
sudo mdadm --create --verbose /dev/md0 --level=0 --raid-devices=5 /dev/nvme2n1 /dev/nvme4n1 /dev/nvme1n1 /dev/nvme3n1 /dev/nvme6n1
and confirm creating the array, I’m able to mount it again and all of the data is still there. However, even if I update the array layout in /etc/mdadm/mdadm.conf and run sudo update-initramfs -u -k all, on the next reboot, sudo mdadm --assemble --scan doesn’t see a RAID superblock on any of the devices.
Any ideas on how to fix this?
Try adding --homehost=any to your create command.
What’s the lsblk -o+FSTYPE output? And mdadm --detail /dev/md0? Also after reboot.
2 Likes
Thanks for the reply. I figured it out and hadn’t gotten around to posting the solution. Though I had wiped the partitions on the disks prior to creating the array, I failed to clear the existing GPT partition table on each disk. After running sudo sgdisk -z /dev/nvme... on each drive and re-creating the array, it was able to persist after reboot.
1 Like