Details
-
Bug
-
Resolution: Fixed
-
Minor
-
Upstream
-
None
-
3
-
9223372036854775807
Description
if a remote host is down for a while, then local MDS may get stuck as well
Call Trace: dump_stack+0x85/0xc0 ptlrpc_queue_wait+0x27e/0x290 [ptlrpc] osp_remote_sync+0x122/0x1a0 [osp] osp_md_xattr_list+0x4d4/0x7f0 [osp] mdd_iterate_xattrs+0xac/0xa30 [mdd] ? mdo_declare_xattr_set+0x330/0x330 [mdd] ? down_write_nested+0x73/0xb0 mdd_migrate_object+0x18c1/0x2b80 [mdd] mdd_migrate+0x385/0x1740 [mdd] ? mdt_object_lock_internal+0x1d3/0x3f0 [mdt] mdt_reint_migrate+0x1bf6/0x3260 [mdt] ? lustre_msg_add_version+0x1c/0x90 [ptlrpc] mdt_reint_rec+0x117/0x240 [mdt] mdt_reint_internal+0x9e1/0xbd0 [mdt] mdt_reint+0x57/0x100 [mdt] tgt_request_handle+0xbe0/0x1970 [ptlrpc]