ONTAP 9.2: new functionality
In this article I would like to review the innovations that appeared in the firmware version of ONTAP 9.2 which came out nine months after the release of version 9.0 . As usual, NetApp is pleased with a large number of new features and capabilities in its storage systems, let's analyze them.
ONTAP 9.1 and newer releases will now be released regularly every 6 months. Previously, releases were being delayed in order to accommodate an incredibly large number of new features, now there will be fewer of them in each release, but they will start coming out more regularly and will not delay the output of other features. There were releases with long-term and short-term support. Understand correctly, both releases are fully production-ready, general avalability, tested, debugged and workable. Since some new systems had the opportunity to buy support for 7 or more years, in order to be able to extend the support period for ONTAP releases, a compromise was made to release some releases with long-term support, and some with short ones, as happens, for example, in Ubuntu.
')
NetApp started making their systems simple as a toaster that made sandwiches. Then the sandwich grew and began to intervene only in the oven, and now ONTAP is a full-fledged whole kitchen, with many built-in features that are somewhere, sometimes hidden and hard to find. ONTAP was once a very simple OS, but now it is very sophisticated and has an incredible mass of functionality, so NetApp continues this constant downward movement to make it cool and sophisticated to be easy to use.
Balance placement (QoS) - when creating a LUN, the system will offer where to place it.
QoS minimum, this functionality is also called Throughput Floors (TP), it can work together with the QoS Maximum functionality (throughput ceilings) that was previously available. Throughput Floors take care that IOPS performance does not fall below the previously specified. A cluster must have at least one AFF node in order for the feature to work for iSCSI / FCP protocols.
Minimum QoS appeared in version 9.2, which is set in group policy performance. How much the system can produce with the current load is dynamically taken from the monitoring Headroom built into the storage system and selected from the optimum. The optimum is the portion of the system load graph in IOPS / Latensie, as long as it goes linearly. After the linear one, the graph begins, where with increasing load, a greater number of IPOs causes a disproportionate, geometric growth of latency with a linear increase of IPOs. Thus, Throughput Floors ensures that the moon is given at least, if necessary, taking part of the performance from those moons that have not yet reached its minimum. In other words, for all other moons, a “temporary or dynamic maximum” will be set, which will be less than their Throughput Ceilings, but larger than the Throughput Floors, so that the moon will be given its minimum.
Inline Aggregate-Wide Deduplication on SSD (or Cross Volume Deduplication). The feature is enabled by default for all SSD aggregates in ONTAP 9.2. Yes, of course I wanted to have a “global cluster dedup”, but there is no limit to perfection and while it works only at the level of the unit, but even with it the current ONDAP SSD units are so large that they are larger than the whole systems of some competitors with global-dedupas. Therefore, the aggregate-dedup will give enough to compete with other All Flash Array (AFA) systems in terms of efficiency. The maximum aggregate size for All Flash FAS (AFF) systems has been increased to 800TiB (previously there were 400)
In version 9.2, a new command line “storage aggregate show-efficiency” is available for the command line, which can show the effect of savings on each individual technology, be it clones, compression, dedup, etc.
FabricPool is a technology implemented as part of the DataFabric strategy, which, in turn, should expand the integration of all NetApp products and enhance data mobility. When NetApp makes a new feature, it doesn’t look at others and doesn’t try to add a tick feature. Instead, a technology is released that is really well thought out and will be in demand by consumers. This also applies to the new Tiering feature FabricPool. With increasing demand for SSDs, the number of installations and interest in this technology increases, so NetApp has released the FabricPool functionality. FabricPool is a cold data offset technology with SSD aggregate (AFF, FAS, ONTAP Cloud systems) to a cold object level - to Amazon S3 cloud or to NetApp StorageGRID object storage system .
Implementing FabricPool Data Tiering on Object Storage will consist of two stages. In 9.2, the ability to shift only snapshots will be available. And the next stage will be the possibility of shifting cold data from the active volum file system to the cold level. So far, FlexFroup, MetroCluster, ONTAP Select, SnapLock, etc. are not supported with FabricPool.
Licensed per terabyte on FAS / AFF systems with Amzon S3. Adding a license immediately adds 10TB. Then you can add by terabyte.
For FabricPool functionality on FAS / AFF storage systems, a license is only necessary if you are using Amazon S3. In the case of NetApp StorageGRID, a license for running FabricPool is not needed. For ONTAP Cloud systems living in Amazon, there is also no ONTAP loss-by-turn licensing to enable FabricPool — you only pay for the actual cold-level space consumed in AWS S3 directly by Amazon itself .
Wolum encryption can now be enabled from the GUI. SnapLock Volumes can now be encrypted with NetApp Volume Encryption. FlexGroup now also works with NVE encryption (only when creating FG).
Online Certificate Status Protocol (OCSP) - this functionality is used for applications that use TLS encryption, for example, LDAP over TLS. Since 9.2, it has become possible to set the maximum number of unsuccessful SSH connection attempts in order to prevent passwords from being searched. iSCSI Endpoint Isolation - allows you to specify a range of IP addresses that are allowed to login to the storage system.
ADP for AFF: Ability to stretch the unit, built on partitions ( RD2 ) on the disks not only in the first shelf (or in the head), but also in the next shelf (up to 48 disks).
ADP Root-Data Partitioning is now supported on hybrid systems: FAS80 / FAS8200 / FAS9000, which will allow a little more efficient use of disks due to the ability to do without dedicated disks for root aggregates.
A new item has appeared in the boot menu (Option 9) for working with partitions (Appeared with ONTAP 9):
OnCommand Unified Manager 7.2 now includes OnCommand Performance Manager right away.
I ask to send messages on errors in the text to the LAN . Comments, additions and questions on the article on the contrary, please in the comments.
ONTAP 9.1 and newer releases will now be released regularly every 6 months. Previously, releases were being delayed in order to accommodate an incredibly large number of new features, now there will be fewer of them in each release, but they will start coming out more regularly and will not delay the output of other features. There were releases with long-term and short-term support. Understand correctly, both releases are fully production-ready, general avalability, tested, debugged and workable. Since some new systems had the opportunity to buy support for 7 or more years, in order to be able to extend the support period for ONTAP releases, a compromise was made to release some releases with long-term support, and some with short ones, as happens, for example, in Ubuntu.
Constant movement towards simplification
')
NetApp started making their systems simple as a toaster that made sandwiches. Then the sandwich grew and began to intervene only in the oven, and now ONTAP is a full-fledged whole kitchen, with many built-in features that are somewhere, sometimes hidden and hard to find. ONTAP was once a very simple OS, but now it is very sophisticated and has an incredible mass of functionality, so NetApp continues this constant downward movement to make it cool and sophisticated to be easy to use.
- Expanding the cluster by adding compatible nodes can now be performed from the GUI, and Auto Discovering of new nodes is now available from there.
- Automatic switchless-cluster detection - a cluster of two or more nodes will automatically determine if there is no longer a need to manually switch
- The detect-switchless-cluster option is enabled by default on new and updated to 9.2 systems.
- OnCommand System Manager Guided Setup (GUI) - At the very beginning, you need to connect to the storage console and set the node-IP address, and further initial cluster configuration can now be performed from the graphical interface
- Updated Dashboard (since ONTAP 9.1)
- Export and Import Cluster Configuration for Initial Setup - allows you to upload and download cluster settings from a csv-template, which can be edited in a regular text editor or Excel
Advanced QoS
Balance placement (QoS) - when creating a LUN, the system will offer where to place it.
QoS minimum, this functionality is also called Throughput Floors (TP), it can work together with the QoS Maximum functionality (throughput ceilings) that was previously available. Throughput Floors take care that IOPS performance does not fall below the previously specified. A cluster must have at least one AFF node in order for the feature to work for iSCSI / FCP protocols.
Minimum QoS appeared in version 9.2, which is set in group policy performance. How much the system can produce with the current load is dynamically taken from the monitoring Headroom built into the storage system and selected from the optimum. The optimum is the portion of the system load graph in IOPS / Latensie, as long as it goes linearly. After the linear one, the graph begins, where with increasing load, a greater number of IPOs causes a disproportionate, geometric growth of latency with a linear increase of IPOs. Thus, Throughput Floors ensures that the moon is given at least, if necessary, taking part of the performance from those moons that have not yet reached its minimum. In other words, for all other moons, a “temporary or dynamic maximum” will be set, which will be less than their Throughput Ceilings, but larger than the Throughput Floors, so that the moon will be given its minimum.
Cross Volume Deduplication
Inline Aggregate-Wide Deduplication on SSD (or Cross Volume Deduplication). The feature is enabled by default for all SSD aggregates in ONTAP 9.2. Yes, of course I wanted to have a “global cluster dedup”, but there is no limit to perfection and while it works only at the level of the unit, but even with it the current ONDAP SSD units are so large that they are larger than the whole systems of some competitors with global-dedupas. Therefore, the aggregate-dedup will give enough to compete with other All Flash Array (AFA) systems in terms of efficiency. The maximum aggregate size for All Flash FAS (AFF) systems has been increased to 800TiB (previously there were 400)
In version 9.2, a new command line “storage aggregate show-efficiency” is available for the command line, which can show the effect of savings on each individual technology, be it clones, compression, dedup, etc.
Tiering or FabricPool
FabricPool is a technology implemented as part of the DataFabric strategy, which, in turn, should expand the integration of all NetApp products and enhance data mobility. When NetApp makes a new feature, it doesn’t look at others and doesn’t try to add a tick feature. Instead, a technology is released that is really well thought out and will be in demand by consumers. This also applies to the new Tiering feature FabricPool. With increasing demand for SSDs, the number of installations and interest in this technology increases, so NetApp has released the FabricPool functionality. FabricPool is a cold data offset technology with SSD aggregate (AFF, FAS, ONTAP Cloud systems) to a cold object level - to Amazon S3 cloud or to NetApp StorageGRID object storage system .
Implementing FabricPool Data Tiering on Object Storage will consist of two stages. In 9.2, the ability to shift only snapshots will be available. And the next stage will be the possibility of shifting cold data from the active volum file system to the cold level. So far, FlexFroup, MetroCluster, ONTAP Select, SnapLock, etc. are not supported with FabricPool.
Licensed per terabyte on FAS / AFF systems with Amzon S3. Adding a license immediately adds 10TB. Then you can add by terabyte.
For FabricPool functionality on FAS / AFF storage systems, a license is only necessary if you are using Amazon S3. In the case of NetApp StorageGRID, a license for running FabricPool is not needed. For ONTAP Cloud systems living in Amazon, there is also no ONTAP loss-by-turn licensing to enable FabricPool — you only pay for the actual cold-level space consumed in AWS S3 directly by Amazon itself .
Wolum encryption can now be enabled from the GUI. SnapLock Volumes can now be encrypted with NetApp Volume Encryption. FlexGroup now also works with NVE encryption (only when creating FG).
Online Certificate Status Protocol (OCSP) - this functionality is used for applications that use TLS encryption, for example, LDAP over TLS. Since 9.2, it has become possible to set the maximum number of unsuccessful SSH connection attempts in order to prevent passwords from being searched. iSCSI Endpoint Isolation - allows you to specify a range of IP addresses that are allowed to login to the storage system.
ADP for AFF: Ability to stretch the unit, built on partitions ( RD2 ) on the disks not only in the first shelf (or in the head), but also in the next shelf (up to 48 disks).
ADP Root-Data Partitioning is now supported on hybrid systems: FAS80 / FAS8200 / FAS9000, which will allow a little more efficient use of disks due to the ability to do without dedicated disks for root aggregates.
A new item has appeared in the boot menu (Option 9) for working with partitions (Appeared with ONTAP 9):
- remove partitioning disks and partitioning;
- clear the configuration and reinitialize all disks using partitioning;
- clear the configuration and reinitialize all disks with whole, non-partitioned disks.
OnCommand Unified Manager 7.2 now includes OnCommand Performance Manager right away.
ONTAP Select now supports
- 2-Node HA for ROBO customers
- ONTAP Select now supports FlexGroup
- vNAS is an external storage for vSAN users, they naturally don’t have an industrial storage system and do not have enough NASa.
- inline deedup for all-falsh ONTAP Select
FAS and All Flash FAS (AFF) hardware platform update
- DS460C Disk Shelf Support in Metro Cluster - Only fully clogged shelves are available. Requires 9.2
- ONTAP 9.2 Metro Cluster now supports up to 8 ISL links
- AFF-based Metrocluster with ONTAP 9.2 supports Inline aggregate dedup (Cross Volume Deduplication), Metrocluster with SSD aggregates based on FAS systems (hybrid), as in conventional cluster systems does not support Inline aggregate dedup
- Upgradeability of the 2620 from a single-node configuration to a two-node one
- Ability to convert 2620 into DS122C regiment and 2650 into DS224C regiment
- Hybrid SSD shelves will appear for Metro Klastara based on 8200/9000 / A300 / A700, where 18 SSD disks of larger size and 6 small SSD for root partitions are used (so as not to waste large disks for root partitions).
- 2620 now supports up to 132 disks (previously 120)
I ask to send messages on errors in the text to the LAN . Comments, additions and questions on the article on the contrary, please in the comments.
Source: https://habr.com/ru/post/328444/
All Articles