Reallocate or not?

[dunno1958]

ONTAP 7.3.2

Hello

I've just added 10 additonal disks into an exisitng aggregate. The aggregate contains NFS datastores which in turn contain live VMDK files. The NetApp volumes containing the NFS datstores are deduplicated.

Do I reallocate or not? If yes, which command do I use?

Thanks

[ekashpureff]

You may want to. How many data disks did you have before the addition ?

What is the resulting RAID layout ?

You cna run 'reallocate measure' on a volume or two on the aggregate - look for a 'hot spots' index number.

See the man page for reallocate: http://now.netapp.com/NOW/knowledge/docs/ontap/rel735/html/ontap/cmdref/man1/na_reallocate.1.htm

I hope this response has been helpful to you.

At your service,

Eugene E. Kashpureff
ekashp@kashpureff.org
Senior Systems Architect / NetApp Certified Instructor
http://www.linkedin.com/in/eugenekashpureff

(P.S. I appreciate points for helpful or correct answers.)

[aborzenkov]

Could you explain what hotpspot (in reallocate output) actually means and how to intepret it? I do not remember any paper describing it.

[scottgelb]

We often have customers run a perfstat, or we run statit separately (perfstat contains statit) and statit will show disk utilization percentage.  If mostly writes, you will see the new disks have heavier utilization but see how much.... if 50% or less it usually isn't an issue.

measure_layout is also a good method like Eugene mentioned, but statit may give what you need much quicker, or you can run both.  For stati,  "priv set advanced ; statit -b"  wait a while for a representative sample of the counters then "priv set advanced ; statit -enr" and you will see util% on all disks.  Check the disks in your aggregate and see the range of % between disks.  For example if you see 10 disks at 95% and the rest at 20% it might make sense to reallocate....but if you see those 10 disks at 40% and the other disks at 20% it wouldn't.

[dunno1958]

Thanks for the replies gents.

Disks in aggergate: 40

RAID groups: 2 x 20 disks (18 data + 2 parity)

Scott, good point about the statit command. I've checked util% and its running at 5% for the original disks and 1% for the new disks I've added.

I've seen the reallocate man pages, however I raised the question because of some other posts that I've seen. One guy says don't reallocate on a volume which contains live virtual machines. Is there any truth in that comment?

Just for clarity, the measure_layout increments range between 1 (optimised) to 10 (very unpotimised)? I get '2' for all volumes in the aggregate.

Thanks

[radek_kubka]

Have you seen this thread already?

http://communities.netapp.com/message/20969#20969

One guy says don't reallocate on a volume which contains live virtual machines. Is there any truth in that comment?

To me it does make sense to be cautious - reallocate process will consume some resources, so scheduling it off peak is recommended.

Regards,

Radek

[scottgelb]

Very good.  At 5% it wouldn't be worth changing anything.  Put more VMs on that aggregate

[cameron_gocke]

I realize this thread is older, but just FYI:  The NetApp recommendation is to never run reallocate on deduplicated volumes.  The deduplication process itself eliminates the benefits of reallocation and vice versa.

[aborzenkov]

Do you have reference to (document with) this recommendation?

[cameron_gocke]

TR-3505 page 21 under the heading 4.4 deduplication and read reallcoation:

"...

A read reallocation scan does not rearrange blocks on disk that are shared between files by deduplication

on deduplicated volumes. Since read reallocation does not predictably improve the file layout and the

sequential read performance when used on deduplicated volumes, performing read reallocation on

deduplicated volumes is not supported. Instead, for files to benefit from read reallocation, they should be

stored on volumes that are not enabled for deduplication"

[stokesjc]

TR-3929 page 9

4.4 Deduplication and Compression Starting in Data ONTAP 8.1 deduplicated data can be reallocated using physical reallocation or read_realloc space_optimized. Although data may be shared by multiple files when deduplicated, reallocate uses an intelligent algorithm to only reallocate the data the first time a shared block is encountered. Prior versions of Data ONTAP do not support reallocation of deduplicated data and will skip any deduplicated data encountered. Compressed data will not be reallocated by reallocate or read reallocate, and it is not recommended to run reallocate on compressed volumes.

.... so the dedup rules have changed