What you might miss about NetApp from Aug-Nov 2019, including Insight in Las Vegas? Part 3

NetApp & Rubrik

NetApp & Rubrik announced collaboration. First StorageGRID can be a target for Rubrik archives. And second Rubrik now supports NetApp SnapDiff API. SnapDiff API is a technology in ONTAP which compares two snaps and gives a list of files changed so Rubrik can copy only changed files. While Rubrik is not the first in working with NetApp SnapDiff APIs, others like Catalogic, Commvault, IBM (TSM) and Veritas (NetBackup) can work with it as well, but Rubrik is the first one with backing up data to a public cloud. Will be available in Rubrik Cloud Data Management (CDM) v5.2 in 2020.

NetApp & Veeam

Veeam Availability Orchestrator v3 (VAO) provide a new level of NetApp integration for DP:

  • FULL recovery orchestration for NetApp ONTAP Snapshots
  • Automated testing and reports that have become essential to your DR strategies
  • TR-4777: Veeam & StorageGRID

Continue to read

All announcements from Aug-Nov 2019

Am I missing something?

Please let me know in the comments below!

If you spotted an error, please let me know personally 😉

Disclaimer

Opinions & observations are my own, and not official NetApp information. This post contains future looking statements and may contain errors. If you have spotted an error, please let me know.

What you might miss about NetApp from Aug-Nov 2019, including Insight in Las Vegas? Part 2

MAX Data 1.5

  • Support for ONTAP 9.6 GA and later releases
  • Support for FAS storage systems or ONTAP Select systems running ONTAP 9.7 besides AFF storage systems
  • Resizing application memory allocation
  • Support for Red Hat Enterprise Linux 7.7
  • Support for local snapshots on server-only systems
  • Significant performance improvements with more I/o, less latency: 5.4M I/o 4KB READ @ 12.5usec latency

Previously in 1.4

With version 1.4 you can use MAX Data without AFF. Tiering now works between PMEM and your SSD installed in the server.

Some info leaks that HCI will support MAX Data at some point.

Considering new compute node H615C with the Cascade Lake CPUs, which is by the way, required for Optane memory, so it looks like NetApp putting all together to make it happen.

Continue to read

Announcements from Aug-Nov 2019

Am I missing something?

Please let me know in the comments below!

If you spotted an error, please let me know personally 😉

Disclaimer

Opinions & observations are my own, and not official NetApp information. This post contains future looking statements and may contain errors. If you have spotted an error, please let me know.

What you might miss about NetApp from Aug-Nov 2019, including Insight in Las Vegas? Part 1

E-Series

Performance

End-to-End NVMe with EF600 – More I/o (x2 times more than EF570), less latency:

NVMe in EF600

  • 100Gb NVMe/RoCE
  • 100Gb NVMe/InfiniBand
  • 32Gb NVMe/FC

E-Series Performance Analyzer

An automated installation and deployment of Grafana, NetApp E-Series Web Services, and supporting software for performance monitoring of NetApp E-Series Storage Systems. NetApp intend this project to allow you to quickly and simply deploy an instance of our performance analyzer for monitoring your E-Series storage systems. We incorporate various open source components and tools in order to do so. While they primarily intend it to serve as a reference implementation for using Grafana to visualize the performance of your E-Series systems, I also can be customizable and extensible based on your individual needs.

https://github.com/NetApp/eseries-perf-analyzer

New TR docs about EF & DB

Continue to read

announcements from Aug-Nov 2019

Am I missing something?

Please let me know in the comments below!

If you spotted an error, please let me know personally 😉

Disclaimer

Opinions & observations are my own, and not official NetApp information. This post contains future looking statements and may contain errors. If you have spotted an error, please let me know.

Why GCP Anthos on NetApp HCI is a big deal?

Google Cloud & NetApp announced a new validated design with GKE running on NetApp HCI on-premise.

Read what you might miss from NetApp announcements during Aug-Nov 2019 compressed into a single article.

Kubernetes was originally designed by Google, Google is one of the main contributors to Docker, and obviously the most advanced, mature & stable on the market. If you tried GKE in GCP & other competitive solutions, you know what I’m talking about.

Containers on-premises are difficult when you want to make Enterprise solution for new containerized applications on-premises for number of reasons: Installation, configuration, management, updates of your core infrastructure components, persistent & performant & predictable storage performance, DevOps do not want to deal with infrastructure they want just consume it. These are the key problems to solve and NetApp aims to do it.

NVA-1141: NetApp HCI with Anthos. NVA Design/Deployment

Bullet points why Google Anthos on NetApp HCI is an important announcement:

  • Hybrid cloud. NetApp according to its Data fabric vision, continue to bring hybrid cloud experience to its users in the flash. Now with Anthos on HCI your on-prem data center becomes just another cloud zone. Software updates for GKE & Anthos are on Google’s shoulders, you just consume it. Not just NetApp HCI maintenance like software & firmware updates can be bought as a service, but space as well. You can pay as you go & consume infrastructure as a service: OPEX instead of CAPEX by request with NetApp Keystone
  • NetApp Kubernetes Services (NKS) In addition to NetApp NKS which allows for the deployment & management of on-premises & in the cloud kubernetes clusters, Anthos provides the ability to deploy clusters on-prem and fully integrated with Google Cloud, including the ability to manage from the GKE console. NKS bundled with Istio, Helm & many other components for your microservices which puts DevOps to the next level. Cloud infrastructure on-premises reached your data center
  • Storage automation. NetApp Trident is literally the most advanced storage driver for containers at the market so far which brings automation, API and persistent storage to containerization world. NetApp Trident with NKS & Anthos totally make sense. Speaking about Automation, NetApp Ansible playbooks are also the most advanced on the market at the moment with 106 published & supported modules, and SolidFire itself is known as fully API-driven storage, so you can work with it solely through RESTful API
  • Simple, predictive and performant enterprise storage with QoS whether on-prem or in the cloud: use Trident and Ansible with NetApp HCI on-prem or CVO or CVS in AWS, Azure or GCP, moreover replicate your data to the cloud for DR or Test/Dev
  • NetApp HCI vs other HCI solutions. One of the most notable HCI competitor is Nutanix so I want to use it as an example. Nutanix’s storage architecture with local disk drives certainly interesting but not unique and obviously have some architectural disadvantages, scalability was one issue to name. Local disk drives are blessing & great news for tiny solutions and not so good of idea when you need to scale it up, cheapness of a small solution with commodity HW & local drives might turn into curse at scale. That’s why Nutanix eventually developed dedicated storage nodes connected over the network to overcome the issue while stepping to the very competitive lend of network storage systems. Because dedicated storage nodes connected over the network is not something new & unique for Nutanix, there are plenty of capable & scalable network storage systems out there. Therefore, most exciting part of Nutanix is their ecosystem & simplicity not the storage architecture though. Now thanks to Anthos, NetApp HCI get in to a unique position with scalability, ecosystem, simplicity, hybrid cloud & functionality for microservices where some other great competitors like Nutanix not reached yet, and that gives NetApp a momentum in the HCI market
  • Performance. Don’t forget about NetApp’s Max Data software which already working with VMware & SolidFire, it will take NetApp only one last step to bring DCPMM like Intel Optane to NetApp HCI. Note NetApp just announced on Insight 2019 a compute node with Intel Cascade Lake CPUs which required for Optane. Max Data is not available on NetApp HCI yet, but we can clearly see that NetApp putting everything together to make it happen. Persistent memory in form of a file system for a Linux host server with tiering for cold blocks to “slow” SSD storage can put NetApp on top of all the competitors in terms of performance

HCI Performance

Speaking about which, take a look on these two performance tests:

  1. IOmark-VM-HC: 5 storage & 18 compute nodes using data stores & VVols
  2. IOmark-VDI-HC: 5 storage nodes & 12 compute nodes with only data stores

Total 1,440 VMs with 3,200 VDI desktops.

Notice how asymmetrical number of storage nodes compared to compute nodes are, and in “Real” HCI architectures with local drives you have to have more equipment, while with NetApp HCI you can choose how much storage and how much compute resources you need and scale them separately. Dedup & compression were enabled in the tests.

Disclaimer

This article is for information purposes only, may contain errors and personal opinions. This text neither authorized nor sponsored by NetApp. If you have spotted an error, please let me know.

What you might miss about NetApp from Aug-Nov 2019, including Insight in Las Vegas? Content

Some competitors might say NetApp do not innovate anymore. Well, read this article and answer yourself whether it is true, or it is just yet another shameless marketing.

Part 1

E-Series

Performance

NVMe in EF600

E-Series Performance Analyzer

New TR docs about EF & DB

Part 2

MAX Data 1.5

Previously in 1.4

Part 3

NetApp & Rubrik

NetApp & Veeam

Part 4

Active IQ 2.0

Active IQ Unified Manager 9.7

Part 5

AFF & FAS

AFF & NVMe

ONTAP AI with containers

ASA

ONTAP

ONTAP Select

ONTAP SDS is in embedded non-X86 systems for edge devices

FlexGroup

SnapMirror Sync (SM-S)

NDAS

SnapCenter 4.2

New with VMware & VVOLs:

Virtual Storage Console (VSC)

FlexCache

MCC

MetroCluster IP

MCC-FC

ONTAP Mediator instead of tie breaker

Part 6

StorageGRID v11.3

Part 7

Keystone

Complete Digital Advisors as part of Support Edge:

Part 8

Lab on demand

Lab on demand for Customers

There are more labs for current NetApp customers

Part 9

NAbox

Harvest 1.6

Part 10

SaaS Backup

SaaS backup for Salesforce

Cloud Volumes

Cloud Volumes On-Premises

Cloud Compliance

Cloud Insights

Cloud Secure

NetApp Kubernetes Services (NKS)

HCI

Part 11

New Solutions

Part 12

Containers

NetApp Trident

Ansible

Part 13

Technical Support

How to collect logs before open a support ticket

How to measure storage performance

Gartner Magic Quadrant for Primary Array

Will NetApp adopt QLC flash in 2020?

Continue to read

All announcements from Aug-Nov 2019

Am I missing something?

Please let me know in the comments below!

If you spotted an error, please let me know personally 😉

Disclaimer

Opinions & observations are my own, and not official NetApp information. This post contains future looking statements and may contain errors. If you have spotted an error, please let me know.

New NetApp platform & Licensing improvements in ONTAP 9.6 (Part 1)

A320

All flash A320 2U platform introduced, here are a few important details for this new AFF system:

  • From the performance perspective of view most notable is ~100 microseconds latency on SQL SLOB workload. If true, that is a notable improvement because previously we’ve seen only sub 1 millisecond (1,000 microseconds) latency and new latency basically a few times (in the best-case scenario ~10 times) faster
    • About 20% better IOPS performance than A300
  • NVDIMM instead of traditional NVRAM in high-end/mid-range platforms. This is the second NetApp AFF platform after A800 system which adopted NVDIMM instead of PCIe-based NVRAM. Strictly speaking, NVDIMM has been around in entry FAS/AFF systems for an extended period of time, but only because of luck of PCIe slots & space in the controllers
  • No disk drives in the controller chassis
  • No RoCE support for hosts. Yet
  • End to End NVMe
  • Rumors from Insight 2018 confirmed about new disk shelves
    • NS224 directly connected over RoCE
    • 2 disk shelves maximum
    • 1.9 TB, 3.6 TB, and 7.6 TB drives supported
    • With an upcoming ONTAP release disk shelves connected to controllers over a switch will be supported and thus more disk shelves than just two
  • Not very important to customers, but interesting update from engineer theoretical perspective: with this new platform HA and Cluster Interconnect connectivity now combined, unlike in any other appliances before.
  • 8x Onboard 100 GbE ports per controller:
    • 2 for cluster interconnect 100 GbE ports (and HA)
    • 2 for the first disk shelf and optionally another 2 for the second disk shelf
    • it leaves 2 or 4 100 GbE ports for host connection
  • 2 optional PCIe cards per controller with next ports:
    • FC 16/32 Gb ports
    • RoCE capable 100/40 GbE
    • RoCE capable 25 GbE
    • Or 10 Gb BASE-T ports

Entry Level Systems

Previously released A220 system now available with 10G BASE-T ports, thanks to increase popularity of 10G BASE-T switches.

MCC IP for low-end platforms

MCC IP becomes available for low-end platforms: A220 & FAS2750 (not for 2720 though) in ONTAP 9.6 and requires a 4-node configuration (as all MCC-IP configs). New features made in a way to reduce cost for such small configurations.

  • All AFF systems with MCC-IP supports partitioning, including A220
  • Entry-level systems do not require special iWRAP cards/ports like other storage systems
  • Mixing MCC IP & other traffic allowed (with all the MCC-IP configs?)
    • NetApp wants to ensure customers to get great experience with their solutions so there will be some requirements your switch must meet to maintain high performance to be qualified for such MCC IP configuration.

Brief history of MCC IP:

  • In ONTAP 9.5 mid-range platforms FAS8200 & A300 added support for MCC IP
  • In ONTAP 9.4 MCC IP becomes available on high-end A800
  • And initially MCC IP introduced in ONTAP 9.3 for high-end A700 & FAS9000 systems.

New Cluster Switches

Two new port-dense switches from Cisco and Brocade with 48x 10/25 GbE SFP ports and a few 40 GbE or 100GbE QSFP ports. You can use same switches for MCC IP. Here is Brocade-based BES-53248 which will replace CN1610:

And new Cisco Nexus 92300YC with 1.2U height.

NVMe

New OS supported with ONTAP 9.6: Oracle Linux, VMware 6.7, and Windows Server 2012/2016. Previously in ONTAP 9.5 were supported SUSE Linux 15 and RedHat Enterprise Linux 7.5/7.6, RedHat still doesn’t have ANA support. New FlexPod config with A800 connected over FC-NVMe to SUSE Linux. Volume move now available with NVMe namespaces.

NVMe protocol becomes free. Again

In ONTAP 9.6 NVMe protocol become free. It was free when firstly introduced in 9.4 without ANA (Analog for SAN ALUA multipathing), and then it became not free in 9.5.

SnapMirror Synchronous licensing adjusted

In 9.6 simplified licensing, SM-S included in Premium Bundle. NetApp introduced SM-S in ONTAP 9.5 and previously licensed it by TB. If you not going to use a secondary system as the source to another system, SM-S do not need licensing on the secondary system.

New services

  • SupportEdge Prestige
  • Basic, Standard and Advanced Deployment options
  • Managed Upgrade Service

Read more

Disclaimer

All product names, logos, and brands are property of their respective owners. All company, product and service names used in this website are for identification purposes only. No one is sponsoring this article.