2 Replies Latest reply: Jul 4, 2013 12:54 AM by AdvUni-MD RSS

FAS3140 - OnTap 8.1r2 - DS4243 Disk Shelves

GEOFFREY.DEAN
Currently Being Moderated

I wonder if anyone can offer some advice:

 

I have a FAS 3140 that is running OnTap 8.1r2 7-mode and have recently purchased 2 * DS4243 disk shelves full of 3TB disks.

 

What is the best configuration for these aggregates in terms of performance and size etc.  The filer is used primarily for CIF shares and snap mirror destinations.

 

Do we go with 1 large aggregate using 44 disks in 4RG of 11 leaving 4 spare.  Use 38 disks in 2RG 19 leaving 10 spare?

 

A recommendation has been made to create two aggregates, Aggr1 uses 22 disks RG size of 11 with Aggr2 using 24 disks RG size of 12, is this a good happy medium?

 

Thanks

  • Re: FAS3140 - OnTap 8.1r2 - DS4243 Disk Shelves
    martin.fisher
    Currently Being Moderated

    Hi - to some extent it depends on what the system (filer) is being used for. If you wanted performance you would probably allocate the max number of disks per RG to the aggr to scale the I/O, in accordance with the NetApp Best practise docs etc. As you are using 3TB Disks and is mostly used for Snapmirror destination and CIFS, I would have run with the Happy Medium option, creating a couple of aggr's so all the data is not allocated to one aggr and trying to maximise the no of disks used. Check the RG sizes and you should have a spare disk or two.

    HTH

  • Re: FAS3140 - OnTap 8.1r2 - DS4243 Disk Shelves
    AdvUni-MD
    Currently Being Moderated

    The FAS3140 has an aggregate limit of 75TiB in 8.1.x

    The maximum (optimal) RAID config, according to the "Storage Subsystem Technical FAQ", is 2 RAID groups with 17 drives each.

    That would leave you with 1 aggregate with 34 disks in two RAID groups and 1 aggregate with 12 disks in 1 RAID group (leaving 2 spares)

     

    That's probably what I would go with, except if you want (almost) equal space for each of the two aggregates, in which case your Idea (22 Disks for aggr1 and 24 disks for aggr2) looks good, although you'll loss a few TiB in total because of the two additional Parity/DParity drives for the fourth RAID group

     

    -Michael

More Like This

  • Retrieving data ...