Shared ACFS filesystem was full on a 2-node Oracle Cluster and the nodes became unresponsive.

Issue

• Shared ACFS filesystem was full on a 2-node Oracle Cluster and the nodes became unresponsive. Following messages were seen on the console just before the hang:

kernel: INFO: task df:4286 blocked for more than 120 seconds.
kernel: "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
kernel: df            D ffffffff80151248     0  4286   4282                     (NOTLB)
kernel:  ffff81037411bb38 0000000000000082 0000000000000000 ffff810880003c10
kernel:  0000000000000292 0000000000000005 ffff810581fa7080 ffff81011df650c0
kernel:  001767e59c11013a 0000000000085412 ffff810581fa7268 0000000a7f388000
kernel: Call Trace:
kernel:  [<ffffffff88663713>] :oracleoks:KsWaitEvent+0x113/0x171
kernel:  [<ffffffff8008e16d>] default_wake_function+0x0/0xe
kernel:  [<ffffffff88264d5c>] :dm_mod:dm_any_congested+0x38/0x3f
kernel:  [<ffffffff886651e8>] :oracleoks:KsCondWaitTimeout+0x59/0x1c8
kernel:  [<ffffffff886fb74a>] :oracleacfs:OfsAcquireDLMLock_int+0x80d/0xfab
kernel:  [<ffffffff886f67de>] :oracleacfs:ofs_dlm_blkrtn+0x0/0x126
kernel:  [<ffffffff886fbf44>] :oracleacfs:OfsOpAcquireDLMLock_int+0x5c/0x445
kernel:  [<ffffffff80019b4b>] __follow_mount+0x26/0x7f
kernel:  [<ffffffff887194c4>] :oracleacfs:OfsGenStatFs+0x1d/0xd3
kernel:  [<ffffffff8875343d>] :oracleacfs:ofs_statfs+0xba/0x14d
kernel:  [<ffffffff8000a64f>] __link_path_walk+0xe14/0xf42
kernel:  [<ffffffff800e0ac2>] vfs_statfs+0x63/0x7f