Hello,
I have two separate systems and with ext4 I cannot get speeds greater than
~350MiB/s when using ext4 as the filesystem on top of a raid5 or raid0.
It appears to be a bug with ext4 (or its just that ext4 is slower for this
test)?
Each system runs 2.6.33 x86_64.
Can someone please confirm?
Here is ext4:
# dd if=/dev/zero of=bigfile bs=1M count=10240
10240+0 records in
10240+0 records out
10737418240 bytes (11 GB) copied, 29.8556 s, 360 MB/s
The result is the same regardless of the RAID type (RAID-5 or RAID-0)
Note, this is not a bandwidth problem:
# dd if=/dev/zero of=/dev/md0 bs=1M count=10240
10240+0 records in
10240+0 records out
10737418240 bytes (11 GB) copied, 17.6871 s, 607 MB/s
With XFS:
p63:~# mkfs.xfs -f /dev/md0
p63:~# mount /dev/md0 /r1
p63:~# cd /r1
10240+0 records in
10240+0 records out
10737418240 bytes (11 GB) copied, 17.6078 s, 610 MB/s
NOTE: With a HW raid controller (OR using XFS), I can get > 500 MiB/s,
this problem only occurs with SW raid (Linux/mdadm).
Example (3ware 9650SE-16PML RAID-6, 15 drives (using EXT4)
$ dd if=/dev/zero of=bigfile bs=1M count=10240
10240+0 records in
10240+0 records out
10737418240 bytes (11 GB) copied, 21.1729 s, 507 MB/s
Justin.
On Sat, 27 Feb 2010 08:47:48 -0500 (EST)
Justin Piszcz <[email protected]> wrote:
> Hello,
>
> I have two separate systems and with ext4 I cannot get speeds greater than
> ~350MiB/s when using ext4 as the filesystem on top of a raid5 or raid0.
> It appears to be a bug with ext4 (or its just that ext4 is slower for this
> test)?
>
> Each system runs 2.6.33 x86_64.
Could be related to the recent implementation of IO barriers in md.
Can you try mounting your filesystem with
-o barrier=0
and see how that changes the result.
NeilBrown
>
> Can someone please confirm?
>
> Here is ext4:
>
> # dd if=/dev/zero of=bigfile bs=1M count=10240
> 10240+0 records in
> 10240+0 records out
> 10737418240 bytes (11 GB) copied, 29.8556 s, 360 MB/s
>
> The result is the same regardless of the RAID type (RAID-5 or RAID-0)
>
> Note, this is not a bandwidth problem:
>
> # dd if=/dev/zero of=/dev/md0 bs=1M count=10240
> 10240+0 records in
> 10240+0 records out
> 10737418240 bytes (11 GB) copied, 17.6871 s, 607 MB/s
>
> With XFS:
>
> p63:~# mkfs.xfs -f /dev/md0
> p63:~# mount /dev/md0 /r1
> p63:~# cd /r1
> 10240+0 records in
> 10240+0 records out
> 10737418240 bytes (11 GB) copied, 17.6078 s, 610 MB/s
>
> NOTE: With a HW raid controller (OR using XFS), I can get > 500 MiB/s,
> this problem only occurs with SW raid (Linux/mdadm).
>
> Example (3ware 9650SE-16PML RAID-6, 15 drives (using EXT4)
> $ dd if=/dev/zero of=bigfile bs=1M count=10240
> 10240+0 records in
> 10240+0 records out
> 10737418240 bytes (11 GB) copied, 21.1729 s, 507 MB/s
>
> Justin.
> --
> To unsubscribe from this list: send the line "unsubscribe linux-raid" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
On Sun, 28 Feb 2010, Neil Brown wrote:
> On Sat, 27 Feb 2010 08:47:48 -0500 (EST)
> Justin Piszcz <[email protected]> wrote:
>
>> Hello,
>>
>> I have two separate systems and with ext4 I cannot get speeds greater than
>> ~350MiB/s when using ext4 as the filesystem on top of a raid5 or raid0.
>> It appears to be a bug with ext4 (or its just that ext4 is slower for this
>> test)?
>>
>> Each system runs 2.6.33 x86_64.
>
> Could be related to the recent implementation of IO barriers in md.
> Can you try mounting your filesystem with
> -o barrier=0
>
> and see how that changes the result.
>
> NeilBrown
Hi Neil,
Thanks for the suggestion, it has been used here:
http://lkml.org/lkml/2010/2/27/66
Looks like an EXT4 issue as XFS does ~600MiB/s..?
Its strange though, on a single hard disk, I get approximately the same
speed for XFS and EXT4, but when it comes to scaling across multiple disks,
in RAID-0 or RAID-5 (tested), there is a performance problem as it hits a
performance problem at ~350MiB/s. I tried multiple chunk sizes but nothing
seemed to made a difference (whether 64KiB or 1024KiB), XFS performs at
500-600MiB/s no matter what and EXT4 does not exceed ~350MiB/s.
Is there anyone on any of the lists that gets > 350MiB/s on a mdadm/sw raid
with EXT4?
A single raw disk, no partitions:
p63:~# dd if=/dev/zero of=/dev/sdm bs=1M count=10240
10240+0 records in
10240+0 records out
10737418240 bytes (11 GB) copied, 92.4249 s, 116 MB/s
A single raw disk formatted with XFS, no partitions:
p63:~# mkfs.xfs /dev/sdm > /dev/null 2>&1
p63:~# mount /dev/sdm -o nobarrier,noatime /r1
p63:/r1# dd if=/dev/zero of=file bs=1M count=10240
10240+0 records in
10240+0 records out
10737418240 bytes (11 GB) copied, 85.7782 s, 125 MB/s
A single raw disk formatted with EXT4, no partitions:
p63:~# mkfs.ext4 /dev/sdm > /dev/null 2>&1
p63:~# mount /dev/sdm -o nobarrier,noatime /r1
p63:/r1# dd if=/dev/zero of=file bs=1M count=10240
10737418240 bytes (11 GB) copied, 85.1501 s, 126 MB/s
p63:/r1#
EXT2 vs. EXT3 vs. EXT4
http://lkml.org/lkml/2010/2/27/77
XFS tests: (550-600MiB/s)
http://linux.derkeiler.com/Mailing-Lists/Kernel/2010-02/msg10572.html
Justin.
Justin Piszcz wrote:
>
>
> On Sun, 28 Feb 2010, Neil Brown wrote:
>
>> On Sat, 27 Feb 2010 08:47:48 -0500 (EST)
>> Justin Piszcz <[email protected]> wrote:
>>
>>> Hello,
>>>
>>> I have two separate systems and with ext4 I cannot get speeds
>>> greater than
>>> ~350MiB/s when using ext4 as the filesystem on top of a raid5 or raid0.
>>> It appears to be a bug with ext4 (or its just that ext4 is slower
>>> for this
>>> test)?
>>>
>>> Each system runs 2.6.33 x86_64.
>>
>> Could be related to the recent implementation of IO barriers in md.
>> Can you try mounting your filesystem with
>> -o barrier=0
>>
>> and see how that changes the result.
>>
>> NeilBrown
>
> Hi Neil,
>
> Thanks for the suggestion, it has been used here:
> http://lkml.org/lkml/2010/2/27/66
>
> Looks like an EXT4 issue as XFS does ~600MiB/s..?
>
> Its strange though, on a single hard disk, I get approximately the
> same speed for XFS and EXT4, but when it comes to scaling across
> multiple disks,
> in RAID-0 or RAID-5 (tested), there is a performance problem as it hits a
> performance problem at ~350MiB/s. I tried multiple chunk sizes but
> nothing
> seemed to made a difference (whether 64KiB or 1024KiB), XFS performs at
> 500-600MiB/s no matter what and EXT4 does not exceed ~350MiB/s.
>
> Is there anyone on any of the lists that gets > 350MiB/s on a mdadm/sw
> raid
> with EXT4?
>
> A single raw disk, no partitions:
> p63:~# dd if=/dev/zero of=/dev/sdm bs=1M count=10240
> 10240+0 records in
> 10240+0 records out
> 10737418240 bytes (11 GB) copied, 92.4249 s, 116 MB/s
I hate to say it, but I don't think this measures anything useful. When
I was doing similar things I got great variabilty in my results until I
learned about the fdatasync option so you measure the actual speed to
the destination and not the disk cache. After that my results were far
slower and reproducible.
--
Bill Davidsen <[email protected]>
"We can't solve today's problems by using the same thinking we
used in creating them." - Einstein
On Sat, 27 Feb 2010, Bill Davidsen wrote:
> Justin Piszcz wrote:
>>
>>
>> On Sun, 28 Feb 2010, Neil Brown wrote:
>>
>>> On Sat, 27 Feb 2010 08:47:48 -0500 (EST)
>>> Justin Piszcz <[email protected]> wrote:
>>>
>>>> Hello,
>>>>
>>>> I have two separate systems and with ext4 I cannot get speeds greater
>>>> than
>>>> ~350MiB/s when using ext4 as the filesystem on top of a raid5 or raid0.
>>>> It appears to be a bug with ext4 (or its just that ext4 is slower for
>>>> this
>>>> test)?
>>>>
>>>> Each system runs 2.6.33 x86_64.
>>>
>>> Could be related to the recent implementation of IO barriers in md.
>>> Can you try mounting your filesystem with
>>> -o barrier=0
>>>
>>> and see how that changes the result.
>>>
>>> NeilBrown
>>
>> Hi Neil,
>>
>> Thanks for the suggestion, it has been used here:
>> http://lkml.org/lkml/2010/2/27/66
>>
>> Looks like an EXT4 issue as XFS does ~600MiB/s..?
>>
>> Its strange though, on a single hard disk, I get approximately the same
>> speed for XFS and EXT4, but when it comes to scaling across multiple disks,
>> in RAID-0 or RAID-5 (tested), there is a performance problem as it hits a
>> performance problem at ~350MiB/s. I tried multiple chunk sizes but nothing
>> seemed to made a difference (whether 64KiB or 1024KiB), XFS performs at
>> 500-600MiB/s no matter what and EXT4 does not exceed ~350MiB/s.
>>
>> Is there anyone on any of the lists that gets > 350MiB/s on a mdadm/sw raid
>> with EXT4?
>>
>> A single raw disk, no partitions:
>> p63:~# dd if=/dev/zero of=/dev/sdm bs=1M count=10240
>> 10240+0 records in
>> 10240+0 records out
>> 10737418240 bytes (11 GB) copied, 92.4249 s, 116 MB/s
>
> I hate to say it, but I don't think this measures anything useful. When I was
> doing similar things I got great variabilty in my results until I learned
> about the fdatasync option so you measure the actual speed to the destination
> and not the disk cache. After that my results were far slower and
> reproducible.
fdatasync:
http://lkml.indiana.edu/hypermail/linux/kernel/1002.3/01507.html
Justin Piszcz wrote:
>
>
> On Sat, 27 Feb 2010, Bill Davidsen wrote:
>
>> Justin Piszcz wrote:
>>>
>>>
>>> On Sun, 28 Feb 2010, Neil Brown wrote:
>>>
>>>> On Sat, 27 Feb 2010 08:47:48 -0500 (EST)
>>>> Justin Piszcz <[email protected]> wrote:
>>>>
>>>>> Hello,
>>>>>
>>>>> I have two separate systems and with ext4 I cannot get speeds
>>>>> greater than
>>>>> ~350MiB/s when using ext4 as the filesystem on top of a raid5 or
>>>>> raid0.
>>>>> It appears to be a bug with ext4 (or its just that ext4 is slower
>>>>> for this
>>>>> test)?
>>>>>
>>>>> Each system runs 2.6.33 x86_64.
>>>>
>>>> Could be related to the recent implementation of IO barriers in md.
>>>> Can you try mounting your filesystem with
>>>> -o barrier=0
>>>>
>>>> and see how that changes the result.
>>>>
>>>> NeilBrown
>>>
>>> Hi Neil,
>>>
>>> Thanks for the suggestion, it has been used here:
>>> http://lkml.org/lkml/2010/2/27/66
>>>
>>> Looks like an EXT4 issue as XFS does ~600MiB/s..?
>>>
>>> Its strange though, on a single hard disk, I get approximately the
>>> same speed for XFS and EXT4, but when it comes to scaling across
>>> multiple disks,
>>> in RAID-0 or RAID-5 (tested), there is a performance problem as it
>>> hits a
>>> performance problem at ~350MiB/s. I tried multiple chunk sizes but
>>> nothing
>>> seemed to made a difference (whether 64KiB or 1024KiB), XFS performs at
>>> 500-600MiB/s no matter what and EXT4 does not exceed ~350MiB/s.
>>>
>>> Is there anyone on any of the lists that gets > 350MiB/s on a
>>> mdadm/sw raid
>>> with EXT4?
>>>
>>> A single raw disk, no partitions:
>>> p63:~# dd if=/dev/zero of=/dev/sdm bs=1M count=10240
>>> 10240+0 records in
>>> 10240+0 records out
>>> 10737418240 bytes (11 GB) copied, 92.4249 s, 116 MB/s
>>
>> I hate to say it, but I don't think this measures anything useful.
>> When I was doing similar things I got great variabilty in my results
>> until I learned about the fdatasync option so you measure the actual
>> speed to the destination and not the disk cache. After that my
>> results were far slower and reproducible.
>
> fdatasync:
> http://lkml.indiana.edu/hypermail/linux/kernel/1002.3/01507.html
>
I wasn't expecting a huge change in value, your data size is large. But
thanks, the total time without sync can be off by at least seconds,
making it hard to duplicate results. You missed nothing this time.
Did you use any of the options with ext4? I found about 15-20% with
options, but I didn't take good enough notes to quote now. :-(
That doesn't mean there wasn't more, I tested on FC9, ext4 was
experimental then.
--
Bill Davidsen <[email protected]>
"We can't solve today's problems by using the same thinking we
used in creating them." - Einstein
On Sun, Feb 28, 2010 at 4:45 AM, Justin Piszcz <[email protected]> wrote:
>
>
> On Sat, 27 Feb 2010, Bill Davidsen wrote:
>
>> Justin Piszcz wrote:
>>>
>>>
>>> On Sun, 28 Feb 2010, Neil Brown wrote:
>>>
>>>> On Sat, 27 Feb 2010 08:47:48 -0500 (EST)
>>>> Justin Piszcz <[email protected]> wrote:
>>>>
>>>>> Hello,
>>>>>
>>>>> I have two separate systems and with ext4 I cannot get speeds greater
>>>>> than
>>>>> ~350MiB/s when using ext4 as the filesystem on top of a raid5 or raid0.
>>>>> It appears to be a bug with ext4 (or its just that ext4 is slower for
>>>>> this
>>>>> test)?
>>>>>
>>>>> Each system runs 2.6.33 x86_64.
>>>>
>>>> Could be related to the recent implementation of IO barriers in md.
>>>> Can you try mounting your filesystem with
>>>> ?-o barrier=0
>>>>
>>>> and see how that changes the result.
>>>>
>>>> NeilBrown
>>>
>>> Hi Neil,
>>>
>>> Thanks for the suggestion, it has been used here:
>>> http://lkml.org/lkml/2010/2/27/66
>>>
>>> Looks like an EXT4 issue as XFS does ~600MiB/s..?
>>>
>>> Its strange though, on a single hard disk, I get approximately the same
>>> speed for XFS and EXT4, but when it comes to scaling across multiple disks,
>>> in RAID-0 or RAID-5 (tested), there is a performance problem as it hits a
>>> performance problem at ~350MiB/s. ?I tried multiple chunk sizes but
>>> nothing
>>> seemed to made a difference (whether 64KiB or 1024KiB), XFS performs at
>>> 500-600MiB/s no matter what and EXT4 does not exceed ~350MiB/s.
>>>
>>> Is there anyone on any of the lists that gets > 350MiB/s on a mdadm/sw
>>> raid
>>> with EXT4?
>>>
>>> A single raw disk, no partitions:
>>> p63:~# dd if=/dev/zero of=/dev/sdm bs=1M count=10240
>>> 10240+0 records in
>>> 10240+0 records out
>>> 10737418240 bytes (11 GB) copied, 92.4249 s, 116 MB/s
>>
>> I hate to say it, but I don't think this measures anything useful. When I
>> was doing similar things I got great variabilty in my results until I
>> learned about the fdatasync option so you measure the actual speed to the
>> destination and not the disk cache. After that my results were far slower
>> and reproducible.
>
> fdatasync:
> http://lkml.indiana.edu/hypermail/linux/kernel/1002.3/01507.html
How did you format the ext3 and ext4 filesystems?
Did you use mkfs.ext[34] -E stride and stripe-width accordingly?
AFAIK even older versions of mkfs.xfs will probe for this info but
older mkfs.ext[34] won't (though new versions of mkfs.ext[34] will,
using the Linux "topology" info).
On Sun, 28 Feb 2010, Bill Davidsen wrote:
> Justin Piszcz wrote:
>>
[ .. ]
>> fdatasync:
>> http://lkml.indiana.edu/hypermail/linux/kernel/1002.3/01507.html
>>
> I wasn't expecting a huge change in value, your data size is large. But
> thanks, the total time without sync can be off by at least seconds, making it
> hard to duplicate results. You missed nothing this time.
>
> Did you use any of the options with ext4? I found about 15-20% with options,
> but I didn't take good enough notes to quote now. :-(
> That doesn't mean there wasn't more, I tested on FC9, ext4 was experimental
> then.
Yes:
I tried nearly every option in the ext4 readme:
more:
p63:~# tune2fs -o journal_data_writeback /dev/md0
tune2fs 1.41.10 (10-Feb-2009)
p63:~# mount /dev/md0 /r1 -o noatime,barrier=0,data=writeback,nobh,commit=100,nouser_xattr
p63:~#
p63:~# cd /r1
p63:/r1# dd if=/dev/zero of=file bs=1M count=10240
10240+0 records in
10240+0 records out
10737418240 bytes (11 GB) copied, 35.7193 s, 301 MB/s
p63:/r1#
p63:~# mount /dev/md0 /r1 -o noatime,barrier=0,data=writeback,nobh,commit=100,nouser_xattr,nodelalloc
p63:~#
p63:/r1# dd if=/dev/zero of=file bs=1M count=10240
10240+0 records in
10240+0 records out
10737418240 bytes (11 GB) copied, 30.5846 s, 351 MB/s
p63:/r1#
p63:~# mount /dev/md0 /r1 -o noatime,barrier=0,data=writeback,nobh,commit=100,nouser_xattr,nodelalloc,max_batch_time=0
p63:~#
p63:/r1# dd if=/dev/zero of=file bs=1M count=10240
10240+0 records in
10240+0 records out
10737418240 bytes (11 GB) copied, 30.8501 s, 348 MB/s
p63:/r1#
p63:~# mount /dev/md0 /r1 -o noatime,barrier=0,data=writeback,nobh,commit=100,nouser_xattr,nodelalloc,min_batch_time=10000
p63:~#
p63:/r1# dd if=/dev/zero of=file bs=1M count=10240
10240+0 records in
10240+0 records out
10737418240 bytes (11 GB) copied, 31.0127 s, 346 MB/s
p63:/r1#
p63:~# mount /dev/md0 /r1 -o noatime,barrier=0,data=writeback,nobh,commit=100,nouser_xattr,nodelalloc,journal_ioprio=0
p63:~#
p63:/r1# dd if=/dev/zero of=file bs=1M count=10240
10240+0 records in
10240+0 records out
10737418240 bytes (11 GB) copied, 31.1559 s, 345 MB/s
p63:/r1# cd
p63:~# mount /dev/md0 /r1 -o noatime,barrier=0,data=writeback,nobh,commit=100,nouser_xattr,nodelalloc,journal_ioprio=7
p63:~#
p63:/r1# dd if=/dev/zero of=file bs=1M count=10240
10240+0 records in
10240+0 records out
10737418240 bytes (11 GB) copied, 31.4713 s, 341 MB/s
p63:/r1#
p63:~# mount /dev/md0 /r1 -o noatime,barrier=0,data=writeback,nobh,commit=100,nouser_xattr,nodelalloc,journal_async_commit
p63:~#
p63:/r1# dd if=/dev/zero of=file bs=1M count=10240
10240+0 records in
10240+0 records out
10737418240 bytes (11 GB) copied, 30.7633 s, 349 MB/s
p63:/r1#
p63:~# mount /dev/md0 /r1 -o noatime,barrier=0,data=writeback,nobh,commit=100,nouser_xattr,nodelalloc,journal_async_commit,oldalloc
p63:~#
p63:/r1# dd if=/dev/zero of=file bs=1M count=10240
10240+0 records in
10240+0 records out
10737418240 bytes (11 GB) copied, 30.7607 s, 349 MB/s
p63:/r1#
p63:~# mount /dev/md0 /r1 -o noatime,barrier=0,data=writeback,nobh,commit=100,nouser_xattr,nodelalloc,journal_async_commit,stripe=1024
p63:~#
Justin.
On Sun, 28 Feb 2010, Mike Snitzer wrote:
> On Sun, Feb 28, 2010 at 4:45 AM, Justin Piszcz <[email protected]> wrote:
[ .. ]
>
> How did you format the ext3 and ext4 filesystems?
>
> Did you use mkfs.ext[34] -E stride and stripe-width accordingly?
> AFAIK even older versions of mkfs.xfs will probe for this info but
> older mkfs.ext[34] won't (though new versions of mkfs.ext[34] will,
> using the Linux "topology" info).
Yes and it did not make any difference:
http://lkml.org/lkml/2010/2/27/77
Incase anyone else wants to try too, you can calculate by hand, or if you
are in a hurry, I found this useful:
http://busybox.net/~aldot/mkfs_stride.html
I believe there is something fundamentally wrong with ext4 when performing
large sequential I/O when writing, esp. after Ted's comments.
Justin.
Justin Piszcz wrote:
>
>
> On Sun, 28 Feb 2010, Mike Snitzer wrote:
>
>> On Sun, Feb 28, 2010 at 4:45 AM, Justin Piszcz
>> <[email protected]> wrote:
>
> [ .. ]
>
>>
>> How did you format the ext3 and ext4 filesystems?
>>
>> Did you use mkfs.ext[34] -E stride and stripe-width accordingly?
>> AFAIK even older versions of mkfs.xfs will probe for this info but
>> older mkfs.ext[34] won't (though new versions of mkfs.ext[34] will,
>> using the Linux "topology" info).
>
> Yes and it did not make any difference:
> http://lkml.org/lkml/2010/2/27/77
>
> Incase anyone else wants to try too, you can calculate by hand, or if you
> are in a hurry, I found this useful:
> http://busybox.net/~aldot/mkfs_stride.html
>
> I believe there is something fundamentally wrong with ext4 when
> performing large sequential I/O when writing, esp. after Ted's comments.
>
> Justin.
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-raid" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
>
I'm going to have to do some testing now, I just tested ext4 against the
raw speed of the device (single device test) and they were quite close
to identical. I'm going to order one more drive to bring my test setup
up to five devices, and do some testing on how it behaves.
More later.
--
Bill Davidsen <[email protected]>
"We can't solve today's problems by using the same thinking we
used in creating them." - Einstein
On Sun, 28 Feb 2010, Bill Davidsen wrote:
> Justin Piszcz wrote:
>>
>>
>> On Sun, 28 Feb 2010, Mike Snitzer wrote:
>>
>>> On Sun, Feb 28, 2010 at 4:45 AM, Justin Piszcz <[email protected]>
>>> wrote:
>>
>> [ .. ]
>>
>>>
>>> How did you format the ext3 and ext4 filesystems?
>>>
>>> Did you use mkfs.ext[34] -E stride and stripe-width accordingly?
>>> AFAIK even older versions of mkfs.xfs will probe for this info but
>>> older mkfs.ext[34] won't (though new versions of mkfs.ext[34] will,
>>> using the Linux "topology" info).
>>
>> Yes and it did not make any difference:
>> http://lkml.org/lkml/2010/2/27/77
>>
>> Incase anyone else wants to try too, you can calculate by hand, or if you
>> are in a hurry, I found this useful:
>> http://busybox.net/~aldot/mkfs_stride.html
>>
>> I believe there is something fundamentally wrong with ext4 when performing
>> large sequential I/O when writing, esp. after Ted's comments.
>>
>> Justin.
>>
>> --
>> To unsubscribe from this list: send the line "unsubscribe linux-raid" in
>> the body of a message to [email protected]
>> More majordomo info at http://vger.kernel.org/majordomo-info.html
>>
> I'm going to have to do some testing now, I just tested ext4 against the raw
> speed of the device (single device test) and they were quite close to
> identical. I'm going to order one more drive to bring my test setup up to
> five devices, and do some testing on how it behaves.
>
> More later.
Thanks, let me know how it goes, I see the same thing, on a single hard
drive, there is little difference between EXT4 and XFS:
http://permalink.gmane.org/gmane.linux.kernel/955357
However, when multiple disks are involved, it is a different story.
Justin.