Issue

• Cluster node hung at shutdown during GFS2 umount, reporting:

Unmounting GFS2 filesystems

• Upon reboot (init 6), the system is hanging when it comes to unmounting file systems.  A screenshot of the console shows the following backtrace:

Turning off swap:                                           [  OK  ]
Turning off quotas:                                         [  OK  ]
Unmounting pipe file systems:                               [  OK  ]
Unmounting file systems:    INFO: task pdflush:362 blocked for more than 120 seconds
"echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
pdflush        D 00000034  2040   362     35            363  361 (L-TLB)
        c478ee40 00000046 f3e32a4e 00000034 00000ae0 eda8342c f92bfa46 0000000
        f7f75aa0 f3e373e4 00000034 00004996 00000000 f7f75bac c4512800 ef33474
        f925b21c c45131a0 ed65c8f0 00000000 eda8342c 00000018 ed65c8d0 fffffff
Call Trace:
 [<f92bfa46>] __put_lkb+0x28/0xd5 [dlm]
 [<f925b21c>] gdlm_bast+0x0/0x78 [lock_dlm]
 [<f931d10e>] just_schedule+0x5/0x8 [gfs2]
 [<c061c065>] __wait_on_bit+0x33/0x58
 [<f931d109>] just_schedule+0x0/0x8 [gfs2]
 [<f931d109>] just_schedule+0x0/0x8 [gfs2]
 [<c061c0ec>] out_of_line_wait_on_bit+0x62/0x6a
 [<0c43642c>] wake_bit_function+0x0/0x3c
 [<f931d102>] gfs2_glock_wait+0x27/0x2e [gfs2]
 [<f932c676>] gfs2_write_inode+0x5c/0x15a [gfs2]
 [<-----66f>] gfs2_write_inode+0x55/0x15a [gfs2]