On 7/19/2011 7:20 PM, Dave Chinner wrote:
> On Tue, Jul 19, 2011 at 05:37:25PM -0500, Stan Hoeppner wrote:
>> On 7/19/2011 3:37 AM, Emmanuel Florac wrote:
>>> Le Mon, 18 Jul 2011 14:58:55 -0500 vous écriviez:
>>>> card: MegaRAID SAS 9260-16i
>>>> disks: 14x Barracuda® XT ST33000651AS 3TB (2 hot spares).
>>>> ~ 30TB
>>> This card doesn't activate the write cache without a BBU present. Be
>>> sure you have a BBU or the performance will always be unbearably awful.
>> In addition to all the other recommendations, once the BBU is installed,
>> disable the individual drive caches (if this isn't done automatically),
>> and set the controller cache mode to 'write back'. The write through
>> and direct I/O cache modes will deliver horrible RAID6 write performance.
>> And, BTW, RAID6 is a horrible choice for a parallel, small file, high
>> random I/O workload such as you've described. RAID10 would be much more
>> suitable. Actually, any striped RAID is less than optimal for such a
>> small file workload. The default stripe size for the LSI RAID
>> controllers, IIRC, is 64KB. With 14 spindles of stripe width you end up
>> with 64*14 = 896KB.
> All good up to here.
And then my lack of understanding of XFS internals begins to show. :(
>> XFS will try to pack as many of these 50-150K files
>> into a single extent, but you're talking 6 to 18 files per extent,
> I think you've got your terminology wrong. An extent can only belong
> to a single inode, but an inode can contain many extents, as can a
> stripe width. We do not pack data from multiple files into a single
Yes, I think I meant stripe unit, the 896KB.
> For new files on a su/sw aware filesystem, however, XFS will *not*
> pack multiple files into the same stripe unit. It will try to align
> the first extent of the file to sunit, or if you have the swalloc
> mount option set and the allocation is for more than a swidth of
> space it will align to swidth rather than sunit.
Interesting. Didn't realize this.
> So if you have a small file workload, specifying sunit/swidth can
> actually -decrease- performance because it allocates the file
> extents sparsely. IOWs, stripe alignment is important for bandwidth
> intensive applications because it allows full stripe writes to occur
> much more frequently, but can be harmful to small file performance
> as the aligned allocation pattern can prevent full stripe writes
> from occurring.....
I don't recall reading this before Dave. Thank you for this tidbit.
How much performance decrease are we looking at here? An mkfs.xfs of an
mdraid striped array will by default create sunit/swidth values right?
And thus this lower performance w/small files.
>> this is wholly dependent on the parallel write pattern, and in which of
>> the allocation groups XFS decides to write each file.
> That's pretty much irrelevant for small files as a single allocation
> is done for each file during writeback.
I believe I was already thinking of the concatenated array at this point
and accidentally dropped those thoughts into the striped array discussion.
>> XFS isn't going
>> to be 100% efficient in this case. Thus, you will end up with many
>> partial stripe width writes, eliminating much of the performance
>> advantage of striping.
> Yes, that's the ultimate problem, but not for the reasons you
> suggested. ;)
Thanks for saving me Dave. :) I had the big picture right but FUBAR'd
some of the details. Maybe there's a job in politics waiting for me. ;)
>> These are large 7200 rpm SATA drives which have poor seek performance to
>> begin with, unlike the 'small' 300GB 15k SAS drives. You're robbing
>> that poor seek performance further by:
>> 1. Using double parity striped RAID
>> 2. Writing thousands of small files in parallel
> The writing in parallel is only an issue if it is direct or
> synchronous IO. If it's using normal buffered writes, then writeback
> is mostly single threaded and delayed allocation should be preventing
> fragmentation completely. That still doesn't guarantee that
> writeback avoids RAID RMW cycles (see above about allocation
The RMW was mainly what I was concerned with here.
>> This workload is very similar to the case of a mail server using the
>> maildir storage format.
> There's not enough detail in the workload description to make that
Good point. I should have said "at first glance... seems similar".
>> If you read the list archives you'll see
>> recommendations for an optimal storage stack setup for this workload.
>> It goes something like this:
>> 1. Create a linear array of hardware RAID1 mirror sets.
>> Do this all in the controller if it can do it.
>> If not, use Linux RAID (mdadm) to create a '--linear' array of the
>> multiple (7 in your case, apparently) hardware RAID1 mirror sets
>> 2. Now let XFS handle the write parallelism. Format the resulting
>> 7 spindle Linux RAID device with, for example:
>> mkfs.xfs -d agcount=14 /dev/md0
>> By using this configuration you eliminate the excessive head seeking
>> associated with the partial stripe write problems of RAID6, restoring
>> performance efficiency to the array. Using 14 allocation groups allows
>> XFS to write write, at minimum, 14 such files in parallel.
> That's not correct. 14 AG means that if the files are laid out
> across all AGs then there can be 14 -allocations- in parallel at
> once. If Io does not require allocation, then they don't serialise
> at all on the AGs. IOWs, If allocation takes 1ms of work in an AG,
> then you could have 1,000 allocations per second per AG. With 14
> AGs, that gives allocation capability of up to 14,000/s
So are you saying that we have no guarantee, nor high probability, that
the small files in this case will be spread out across all AGs, thus
making more efficient use of each disk's performance in the concatenated
array, vs a striped array? Or, are you merely pointing out a detail I
have incorrect, which I've yet to fully understand?
> And given that not all writes require allocation and allocation is
> usually only a small percentage of the total IO time. You can have
> many, many more write IOs in flight than you can do allocations in
> an AG....
Ahh, I think I see your point. For the maildir case, more of the IO is
likely due to things like updating message flags, etc, than actually
writing new mail files into the directory. Such operations don't
require allocation. With the workload mentioned by the OP, it's
possible that all of the small file writes may indeed require
allocation, unlike the maildir workload. But if this is the case,
wouldn't the concatenated array still yield better overall performance
than RAID6, or any other striped array?
If I misunderstood your last point, or any points, please guide me to
the light Dave.