Why does CPU and IO report having high load?

Issue

• A bare metal server running IBM DataStage that is logging blocked task messages. The OS is on local storage, but DataStage itself is stored on a fibre channel drive provided by a CLARiiON CX4-960.
• An example of the output we're getting:

INFO: task osh:25444 blocked for more than 120 seconds.
"echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
osh           D ffffffff801508e3     0 25444  25443                     (NOTLB)
ffff810f065cded8 0000000000000086 0000000007cff966 ffffffff80047f4a
ffff810f065cde48 0000000000000007 ffff81122e535080 ffff8101203b6080
0000181fcf1a6786 000000000005c312 ffff81122e535268 0000000101014120
Call Trace:
[<ffffffff80047f4a>] pagevec_lookup_tag+0x1a/0x21
[<ffffffff80063c6f>] __mutex_lock_slowpath+0x60/0x9b
[<ffffffff80022205>] __up_read+0x19/0x7f
[<ffffffff80063cb9>] .text.lock.mutex+0xf/0x14
[<ffffffff800f3473>] __sync_inodes+0x85/0xaa
[<ffffffff800e44a7>] sync_filesystems+0x13/0x102
[<ffffffff800e1185>] do_sync+0x28/0x5a
[<ffffffff800e11c5>] sys_sync+0xe/0x12
[<ffffffff8006149d>] sysenter_do_call+0x1e/0x76

• There is a high load on CPU and on IO devices. Why is this hapenning ?