Details
-
Bug
-
Resolution: Fixed
-
Minor
-
None
-
None
-
3
-
9223372036854775807
Description
This issue was created by maloo for Andreas Dilger <adilger@whamcloud.com>
This issue relates to the following test suite run: https://testing.whamcloud.com/test_sets/9822f4c9-679e-4798-aa60-f9c1af2f6c38
lustre-initialization failed with the following error in the test logs when running an interop test from master (2.13.52) with b2_12 (2.12.4):
2020-03-22T03:03:50 Starting mds1: -olocalrecov /dev/mapper/mds1_flakey /mnt/lustre-mds1 2020-03-22T03:03:50 CMD: trevis-46vm4 mkdir -p /mnt/lustre-mds1; mount -t lustre -olocalrecov /dev/mapper/mds1_flakey /mnt/lustre-mds1 [ 91.607933] LDISKFS-fs (dm-3): Unrecognized mount option "localrecov" or missing value 2020-03-22T03:03:50 trevis-46vm4: mount.lustre: mount /dev/mapper/mds1_flakey at /mnt/lustre-mds1 failed: Invalid argument
and the following in the MDS logs:
[ 91.607933] LDISKFS-fs (dm-3): Unrecognized mount option "localrecov" or missing value
This is due to the landing of patch https://review.whamcloud.com/36025 "LU-12722 target: disable recovery for local clients", which adds the "-olocalrecov" mount option for the MDTs, even though this option is not supported by older Lustre versions.
Unfortunately, adding this in local.sh is probably too early to be checking the MDS version, so it makes sense to move this option to init_lustre_params() or similar. A better long-term solution would be to have the servers provide their own config for the local version, rather than having it generated by the client (which may be running a different release), but that is likely a lot more effort.
VVVVVVV DO NOT REMOVE LINES BELOW, Added by Maloo for auto-association VVVVVVV
lustre-initialization lustre-initialization - 'lustre-initialization timed out'