On May 12, 2010, at 12:43 PM, Trond Myklebust wrote:
On Wed, 2010-05-12 at 11:17 -0400, Andy Adamson wrote:
I've seen the following intermittently while running the
Connecatathon
special 'write/read 30MB file' test. It's always been with the pNFS
client, either running with the pNFS file module installed, or
running
without it installed (NFSv4.1 only). But given the intermittent
behavior, that could just be because I run a lot more tests on the
pNFS kernel.
I've seen this with 2.6.34-rc3 and up.
Any ideas?
-->Andy
May 11 19:35:41 fedora-64 kernel: INFO: task bigfile:2174 blocked for
more than
120 seconds.
May 11 19:35:41 fedora-64 kernel: "echo 0 > /proc/sys/kernel/
hung_task_timeout_s
ecs" disables this message.
May 11 19:35:41 fedora-64 kernel: bigfile D ffff88005b10d380
0 2174
2137 0x00000084
May 11 19:35:41 fedora-64 kernel: ffff88005ac83c28 0000000000000082
000000000000
0000 0010000000000000
May 11 19:35:41 fedora-64 kernel: ffff88005ac83fd8 ffff8800520d0000
000000000001
3780 ffff88005ac83fd8
May 11 19:35:41 fedora-64 kernel: 0000000000013780 0000000000013780
000000000001
3780 0000000000013780
May 11 19:35:41 fedora-64 kernel: Call Trace:
May 11 19:35:41 fedora-64 kernel: [<ffffffff810aae7a>] ? sync_page
+0x0/0x4a
May 11 19:35:41 fedora-64 kernel: [<ffffffff813b1b26>] io_schedule
+0x3d/0x52
May 11 19:35:41 fedora-64 kernel: [<ffffffff810aaec0>] sync_page
+0x46/0x4a
May 11 19:35:41 fedora-64 kernel: [<ffffffff813b1f61>]
__wait_on_bit_lock+0x46/0
x8f
May 11 19:35:41 fedora-64 kernel: [<ffffffff8103cee6>] ?
release_console_sem+0x1
bd/0x1ee
May 11 19:35:41 fedora-64 kernel: [<ffffffff810aae20>] __lock_page
+0x66/0x6d
May 11 19:35:41 fedora-64 kernel: [<ffffffff81055ad5>] ?
wake_bit_function+0x0/0
x33
May 11 19:35:41 fedora-64 kernel: [<ffffffff810aae63>] lock_page
+0x3c/
0x41
May 11 19:35:41 fedora-64 kernel: [<ffffffff810ab533>] filemap_fault
+0xc3/0x30b
May 11 19:35:41 fedora-64 kernel: [<ffffffff810c3202>] __do_fault
+0x54/0x416
May 11 19:35:41 fedora-64 kernel: [<ffffffff810c39fa>]
handle_mm_fault
+0x436/0x82b
May 11 19:35:41 fedora-64 kernel: [<ffffffff810c7f49>] ?
do_mmap_pgoff
+0x290/0x2f3
May 11 19:35:41 fedora-64 kernel: [<ffffffff813b64e8>] do_page_fault
+0x25e/0x273
May 11 19:35:41 fedora-64 kernel: [<ffffffff813b3bb5>] page_fault
+0x25/0x30
The above is just an ordinary page fault. Something else is apparently
holding a lock on the page and so you are waiting for it to be
released.
Is there a page lock leak somewhere in the pNFS client perhaps?
I'll look. Thanks
-->Andy
Cheers
Trond
--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
