Details
-
Bug
-
Resolution: Fixed
-
Critical
-
None
-
Lustre 1.8.6
-
None
-
Virtualized environment involving 4 OSS's, 1 MDT, and 4 clients
-
2
-
6497
Description
running fdisk -l from the MDT shows sdb as the correctr size of our lustre config and shows no partition table. Running df from ANY of the clients results in a hang.
From the MDT, we've captured the last few lines of the dmesg output: Lustre: MDS lustre-MDT0000: lustre-OST0010_UUID now active, resetting orphans Lustre: 3334:0:(quota_master.c:1718:mds_quota_recovery()) Only 0/32 OSTs are active, abort quota recovery Lustre: lustre-OST0012-osc: Connection restored to service lustre-OST0012 using nid 192.168.1.31@tcp. Lustre: MDS lustre-MDT0000: lustre-OST0012_UUID now active, resetting orphans Lustre: 3334:0:(quota_master.c:1718:mds_quota_recovery()) Only 0/32 OSTs are active, abort quota recovery Lustre: 3334:0:(quota_master.c:1718:mds_quota_recovery()) Skipped 4 previous similar messages Lustre: lustre-OST0014-osc: Connection restored to service lustre-OST0014 using nid 192.168.1.31@tcp. Lustre: Skipped 4 previous similar messages Lustre: MDS lustre-MDT0000: lustre-OST0014_UUID now active, resetting orphans Lustre: Skipped 4 previous similar messages