Hi, I''m deliberately filling up an OST with dd. Not surprisingly I see some ENOSPC errors on the client: LustreError: 2325:0:(client.c:576:ptlrpc_check_status()) @@@ type =PTL_RPC_MSG_ERR, err == -28 req@f71a4200 x9097/t0 o4->oss2_ost8_UUID@oss2_UUID:28 lens 328/288 ref 2 fl Rpc:R/0/0 rc 0/-28 What puzzels me is that they keep coming on the client long after the dd process has stopped: LustreError: 2325:0:(client.c:576:ptlrpc_check_status()) @@@ type =PTL_RPC_MSG_ERR, err == -28 req@f71a4200 x68032/t0 o4->oss2_ost8_UUID@oss2_UUID:28 lens 328/288 ref 2 fl Rpc:R/0/0 rc 0/-28 LustreError: 2325:0:(client.c:576:ptlrpc_check_status()) Skipped 29683 previous similar messages Meanwhile on the OSS, I see a "billion" lines of these and the just keep coming - also after the dd process has finished. There is no other activity on the filesystem. ... ldiskfs_map_inode_page: error reading block 2219264 ldiskfs_map_inode_page: error reading block 2217728 ldiskfs_map_inode_page: error reading block 2219776 ... Any ideas of whats going on ? I''m using 1.4.8 on vanilla 2.6.12. Compiled with --enable-ldiskfs /Jakob