2017-03-31 02:58 AM
I have an envioremnt in which I have a SAS aggregate with flash pool (aggr1) and a SSD aggregate (aggr2), serving different purposes. Per node, of course.
When creating my SVM (one per node, again), I 'm not sure about the best practices about which aggregate choose as the root aggregate for the root volume. The "where" question. When creating SVMs which will serve data volumes residing in the SAS aggrs, I choose those aggr for the root volume. Easy. But what about when creating SVMs which will serve data volumes residing in the SSD aggrs? I see two options:
- Esthetic: I place the root volume in the same SSD aggr in which i will place the data volume.
- Economize: I place the root volume in a SAS aggr, so I dont waste SSD resources. Yay, i know it's just 1Gb.
I have been reading some articles and best-pratices KBs, with lots of how/what/why/when usefull tips (including LS mirros), but can not find anything about "where". I tell myself that, as my data volumes namespace are under the root volume namespace, I should use the SSD aggr in this case, but I want to be sure.
Thanks to everybody in advance! I hope this is not a very stupid question.
Solved! See The Solution
2017-04-10 11:34 PM
Your svm root volume can be in any of the data aggr you have. (as long as its not in node's root aggr, which have cfo failover policy)
I'm lil curious by this.... "When creating my SVM (one per node, again),"
There is no such requirements. The best paractice is to keep SAN proptocol in one SVM and NAS for another SVM
But usually customers have verious requirement to virtulize their storage.. so they can create as many as they want.
keep in mind managibilty will be an issue.
There is not much data under SVM root volume.. its mostly for namespace.
Its used to be 20M volume and netapp increase it to 1G with Ontap 8.3
Most of the information in that kept in memory.. unless there is a change.. system use the info from memory.
So it dosen't matter you store them in SSD or SAS drive.
Hope this help..