2015-03-10 07:48 AM
We have a HA 2 controller Filer setup havng (NetApp Release 8.1.4P7 7-Mode). Recently we wish to setup VMWare in our environment in such a way that the ESXi Datastores are
stored onto Netapp Filer disks via NFS protocol.
For this we have purchased 12 new disks. I have few questions regarding this and appreciate any feedback/suggestions in this regard.
1. We already have 2 aggregates (first aggregate having 500 GB SATA disks) and (second having 1 TB BSAS disks). The newly purchased
disks are 2 TB SATA disks. As per the Netapp Manual , they say , it is better not to mix 2 different types of disks in one aggregate.
also preferably not to mix 2 different sizes disks in one aggregate. If I assume this 2 points correctly then the obvious
choice is to create a new aggregate and add 12 disks into it.
2. Then use this aggregate to create the datastores for varous ESXi VMs. and then we are good to go? atleast I assume.
I would like to get your views on this. Moreover if you think there are better ways in which the Netapp Filer can be setup for VMWare then
please let me know.
These are the commands I would follow while adding the disks.
After inserting all the 12 disks in filer
1. disk show -n (To list all unnassigned disks)
2. disk assign all (in our setup disk.auto_assign is off ..so manually assigning all the disks. This will also put all the disks in spare?? )
3. aggr create newaggr -n -d diskname1 diskname2 (likewise add all the 11 disks in one go. 1 disk I would like to keep in spare)
This should create newaggr with 11 disks with raid_dp group. That means 1 for dparity + 1 for parity and rest 8 for data. Correct?
4. In the end, confirm this by "aggr status -s" and "aggr status -r"
5. Thats it ...I hope so. Or am I missing something?
I would say thanx in advance for taking your time to read/comment the post.
Any questions please let me know.
2015-03-10 08:08 AM
- with the -n switch in 'aggr create' it wouldn't do anything, just show a 'what if'
other than that, it looks fine. Once done, create an NFS volume and present it to ESX (either from VSC or manually)
also make sure you have the correct NFS.MaxQueueDepth value in ESX to avoid this issue: