[HPDD-discuss] Using multiple NIDs for the same failnode

Hi. I'm trying to perform a rolling upgrade of a Lustre 2.1.4 filesystem to 2.4.2. The each server has multiple NIDs with failover. Here's tunefs.lustre output from an OST: checking for existing Lustre data: found CONFIGS/mountdata Reading CONFIGS/mountdata Read previous values: Target: pfs-OST0001 Index: 1 Lustre FS: pfs Mount type: ldiskfs Flags: 0x2 (OST ) Persistent mount opts: errors=remount-ro,extents,mballoc Parameters: ost.quota_type=ug mgsnode=192.168.115.8@o2ib,172.23.46.8@tcp failover.node=192.168.115.9@o2ib,172.23.46.9@tcp Permanent disk data: Target: pfs-OST0001 Index: 1 Lustre FS: pfs Mount type: ldiskfs Flags: 0x2 (OST ) Persistent mount opts: errors=remount-ro,extents,mballoc Parameters: ost.quota_type=ug mgsnode=192.168.115.8@o2ib,172.23.46.8@tcp failover.node=192.168.115.9@o2ib,172.23.46.9@tcp When I attempt to failover the OST to a 2.4.x server, I am seeing some type of parsing error: Mar 3 18:42:57 ki-oss02 kernel: LDISKFS-fs (dm-1): Unrecognized mount option "172.23.46.8@tcp" or missing value Mar 3 18:42:57 ki-oss02 kernel: LustreError: 789:0:(osd_handler.c:5410:osd_mount()) pfs-OST0001-osd: can't mount /dev/mapper/360080e5000245170000004aa4ebc173c: -22 This looks a lot like the bug described in LU-4460. Is there a fix/workaround for this in a current maintenance release? ------- Yemi