I deploy a test file system on 2 nodes, with lustre-2.10.4, then I will try to update lustre from 2.10.4 to 2.12.13
Mgs(management Service ) and mdt(metadata service) are installed on one node, this node can either use ldiskfs or zfs as underlying partition to make lustre file system. In both test nodes, I already created one zpool(ost-001) on each node.
umdist11#mkfs.lustre --backfstype=zfs --fsname=t3test --mgs --mdt ost-001/mdtmgs umdist11#zfs list NAME USED AVAIL REFER MOUNTPOINT ost-001 16.4M 220T 256K none ost-001/mdtmgs 13.7M 220T 13.7M none
#tunefs.lustre ost-001/mdtmgs
#mkdir /lustre/mgt # mount -t lustre ost-001/mdtmgs /lustre/mgt/ #lctl dl
umdist12#mkfs.lustre --backfstype=zfs --fsname=t3test --mgsnode=10.10.2.120@tcp --ost --index=0 ost-001/ost0001 umdist12#zfs list umdist12#zfs get all -s local umdist12#tunefs.lustre ost-001/ost0001 umdist12#mount -t lustre ost-001/ost0001 /mnt/ost-001/ #lctl dl
bl-2-1#mount -t lustre 10.10.2.120@tcp:/t3test /lustre/t3test/ [root@bl-2-1 ~]# cd /lustre/t3test/ [root@bl-2-1 t3test]# dd if=/dev/zero of=file2 bs=10M count=1000 [root@umdist12 tools]# df -h /mnt/ost-001 Filesystem Size Used Avail Use% Mounted on ost-001/ost0001 219T 9.8G 219T 1% /mnt/ost-001 [root@bl-2-1 t3test]# lfs df -h /luste/t3test UUID bytes Used Available Use% Mounted on t3test-MDT0000_UUID 218.7T 13.6M 218.7T 0% /lustre/t3test[MDT:0] t3test-OST0000_UUID 218.7T 9.8G 218.7T 0% /lustre/t3test[OST:0] filesystem_summary: 218.7T 9.8G 218.7T 0% /lustre/t3test-- WenjingWu - 07 Feb 2020