Just got new disks, and a question about them

[lebedevanton]

Hi folks, I just got a tray of 16 SATA disks, 828G each.

When I created an Aggregate out of them it used 14 usable disks and 2 for the double parity.

Now I thought I was going to get more usable space, but I only see  8.5Tb usable space.

aggr show_space -h sata
Aggregate 'sata'

    Total space      WAFL reserve    Snap reserve    Usable space       BSR NVLOG           A-SIS
         9932GB           993GB           446GB          8491GB                           0KB             0KB

No volume's contained in the aggregate

Aggregate                       Allocated            Used           Avail
Total space                           0KB             0KB          8491GB
Snap reserve                        446GB             0KB           446GB
WAFL reserve                        993GB           256KB           993GB

I had 14 raw tb's in the 16 disks, and I thought that I would get at least like 10 TBs usable.

Can I get more out of this disk than I already have ?

Thanks.

Anton

[radek_kubka]

Hi Anton,

First of all - what do you mean by 'a tray of 16 disks'? Normally you get either DS14 with 14 drives, or fully populated DS4243 with 24 drives, or half-populated DS4243 with 12 drives.

If we are talking about 16 SATA drives, then your default RAID group size for RAID-DP is 14, so you end up with two RAID groups.

Did you look into this thread re raw vs. usable capacity?

http://communities.netapp.com/message/29008#29008

Regards,
Radek

[reinoud7]

The only thing what you can do is set the aggr snap reserve to 0 and set the aggr snap sched to 0 0 0 and delete the existing aggr snapshots (if they are there).

We do this for all our aggregates, with one exception: you can't do this for a metrocluster.

You will save 446 GB.

[lebedevanton]

Sorry i mistakenly wrote 16 disks, its 14 in that tray.

I  just destoryed it and rebuilding it again, i think I may have misconfigured it from the get go, thats why it wasnt looking right.

Now to save the snap space, is it aggr options 'aggrname' nosnap=on  ?

[eric_barlier]

Hi,

That option will probably do it. As Reynoud mentioned you can do this

snap reserve -A aggr 0

snap sched -A aggr 0 0 0

wait a while and see  if all aggr snapshots have been deleted

snap list -A  aggr

Cheers,

Eric

[lebedevanton]

Thank you, I will do that on this aggr.

I have another question, about unowned disk, when i do disk show -n I get 26 disks that are not owned, does that mean these disk dont do anything, and are not serving any data ?

I can assign them to spare, Ive tried to do it to one, but what if I want to assign that disk to already existing aggregate ?

so question is two fold, those unowned disks, are they working for me, or they are slacking off... and can I add them to an already existing aggregate ?

Thank you all

[eric_barlier]

you got 26 disks that are unowned! bless. if a disk is unowned its NOT in use (so you ve just got yourself another 26 disks), it cant be put in an aggregate. you now need to sit down

and review your aggr. raid layout and come up with a good layout. If you re not too bothered you can assign the disks to which controller

you want with

VIC-NDC-FAS2040-99A*> disk assign

        assign {<disk_name> | all | [-T <storage type>] -n <count> | auto} [-p <pool>] [-o <ownername>] [-s <sysid>]

Once a disk is owned by a controller you can then

1. create aggr.

2 add to existing aggr.

think about trying to avoid mixing disk types on same controller, load balancing when you set up your RAID layout. feel free to

post your current layout and ask for advice.

aggr options

will tell us what we need to know.

Eric