Options
- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Mute
- Printer Friendly Page
FAS8060, 4*24*900GB SAS disk shelves - how to configure with with maximum efficiency?
2017-04-13
03:02 AM
4,101 Views
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hello folks,
Wellk we got our new replacement Netapp system. It is 8060, two controllers in one box. Unfortunatelly, number of drives came less than I expected - 4 shelves, 900GB 10K SAS drives.
My goal is to minimize number of wasted disk. With 7-mode we've been running for last 5 years without a problem, we have just one aggregate per filer, which has root volume and data volumes as well. There are 3 shelves attached per filer (6 total), coming with 72 15K 450GB drives. I had 2 spare drives for the whole system, and configured 28 disk per RG to maximize the space/capacity. As I said, it worked fine.
Now, with the new system and even with 9.1, do I understand correctly that I have no choice to allocate 3 disk for root aggregate per node and dedicate one spare drive for the root aggregate AND one more - for data aggregates for each node? In other words, from 48 disks available per node, 3 goes to the root aggregate with one spare, one more spare to data aggregate, 4 disks dor DP (two RAID groups for data aggregate) and as a result, only 48-9=39 data disks?
Is above calculation correct?
Is there a way to minimize wasted disks?
Thanks,
P.S. Why can't Netapp have the option to have shared root aggregate for the cluster? Why it has to be per node? Or even - why I has to be on disks - either SAS or flash? Why it can't be build on montherboard/NVRAM/Flash and be completely independed from data shelves/drives?
Solved! See The Solution
1 ACCEPTED SOLUTION
irakli_natsvlishvili has accepted the solution
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Unfortunately, no.
root-data or root-data-data partitioning of SSD's are completely orthogonal to the 4-way partitioning of SSD's for Flash Pools.
4 REPLIES 4
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Almost.
48 HDDs.
Root aggr: 1D + 2P = 3
Data aggr: 2 x (20D + 2P) = 44
Spare: 1
Root-data partitioning can improve the efficiency, but it's not available in your configuration.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Aha, I don't need to dedicate 1 spare for the root aggregate only.
Question - if there is one box with flash drives, is it possbile to create a minimum allowable partition size for root and dedicate the rest of partition(s)/drive(s) for flash pool for caching SAS drives?
As an example - there is one box filled with 24 600GB (as an example dize) flash disks for the whole system in addition to 2x24 SAS disks per node. 11 of flash disks are allocated for one node, 11 - for another node.
Using that 11 drives, is it possible to create a small partition for root (whatever minimum is allowed) and the remainder on each flash disk be used for flash pool to cache data aggregate created on SAS drives?
irakli_natsvlishvili has accepted the solution
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Unfortunately, no.
root-data or root-data-data partitioning of SSD's are completely orthogonal to the 4-way partitioning of SSD's for Flash Pools.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Good question. This may be possible, although probably highly unusual. I would speak with your NetApp contact to ensure such configuration is supported.
