On the 29.06.2017, NetApp ONTAP OS version 9.2 as GA (General Availibility) has published its storage. The release Canidate (RC) was available for some time, but many of you shy at the label “Release Canidate” also to engage them. NetApp as a reminder, versions are available also on his RC full production support.
But now to the innovations of ONTAP 9.2:
- The name: We spoke always of clustered DataONTAP or 7 mode. Today, one speaks only of ONTAP at NetApp. In my opinion, to cut the last braids to the old 7 Mode now also through the naming convention mode.
- Aggregate inline Deduplication: Who to do had you been with NetApp will already know that de-duplication technology. First only as a downstream process, then with the large-scale spread of SSDs also as inline de-Duplikation and of course also for non SSD units. The deduplication happened on volume level. That means I have many VMs of same type in a volume I can reach high de-duplication rates. I had many small volumes and have distributed my VMs on it, the de-duplication rate was not exactly great. The Unit brings to my understanding inline de Duplikation now remedy, by both on volume level, as deduplicate at the aggregate level. In my opinion it is can screw further the yield this efficiency mechanisms upwards. I rely on marketing statements there but not for now, until I saw it myself 😉
- Fabric pools: Advance times sum up: ‘Hot crap’ :). A Storage Tiering in the cloud means simply fabric pools.
In your “hot” data on a SSD Tier resources are provided and your cold data, so that used little by little are automatically in the cloud. Target Amazon S3 is supported as a cloud and NetApps StorageGRID in the first stage of the FabricPool. NetApp has expressed ambitions but also to expose the FabricPool for Microsoft azure. From my point of view, the advantage of FabricPool clearly, that I must provide not much hardware in my data center is located. It is also difficult to estimate how many hot and cold data I actually have to ever enter a valid sizing. The scope for my cold data is too small, so I’m often in the situation to create new Diskshelfs with slow disks, or I put cold data areas on “expensive” SSD. FabricPool might have a very exciting solution. But for the time being only for all Flash FAS systems. Also burning, I would be interested what makes NetApp, if the connection to the Internet is lost. Is the entire Aggregate offline? That would be bad. At the moment I suspect that they can buffer a temporary Internet failure mechanisms and increased wait times with a series of sophisticated caching. - QoS minimum: so far you had in your onTap systems a quality of service policy with which you could set the maximum number of IOPS or MB / s. You could prevent, for example, that a workload “eats away all resources of your storage” and no “air” was more for the other workloads and they were thus impair performance.
OnTap 9.2 I can now also a minimal value for IOPS or MB / s set. For example, I can guarantee its performance the volume my ERP system running this. It is self-evident I hope, that I can guarantee no more IOPS than the overall system can produce. So 50,000 IOPS guarantee at a FAS who can afford only 10,000 IOPS total is not in there. - There are ADPv2 for FAS systems: in this post I already explained how ADP. New now is that ADPv2 now also for the FAS 9000, FAS 8200 and the FAS support the 80xx series. It is no longer necessary SSDs used to have, but the ADPv2 we on SAS and NL SAS panels work. So more net capacity 🙂
I hope I could give you a little insight into the innovations of ONTAP 9.2 with these TOP five. The complete list of all new features can be found here. Disclaimer: This post is based on insights that I gained through research on NetApps Web sites. Subject to misinterpretation or misunderstanding.