BEST PRACTICES GUIDE: Nimble Storage Best Practices For Scale-Out - NDM

BEST PRACTICES GUIDE :Nimble Storage Best Practices for Scale-Out

ContentsIntroduction. 3Terminology . 3Planning Scale-Out Clusters and Pools. 3Cluster Arrays Based on Management Boundaries. 3Adding, Removing and Merging Scale-out Cluster Arrays . 4Pooling Arrays within a Scale-out Cluster . 4Pool Arrays for Consolidation and Fault Isolation . 5Pool Arrays with the Same Storage Capacity . 5Pool Rebalancing. 6Non-Disruptive Pool Rebalancing Use Cases . 6Networking Best Practices. 7SAN Bandwidth Availability. 8Cluster Arrays on the same Subnets . 8Choosing iSCSI Switches . 8Use Nimble CIM and PSP Provider for VMware Hosts . 9Use Nimble Connection Manager on Windows Hosts . 9Non-Windows/VMware Host Scale-out Connectivity .10Getting Started .10BEST PRACTICES GUIDE: NIMBLE STORAGE BEST PRACTICES FOR SCALE-OUT2

IntroductionNimble Storage scale-out cluster functionality allows scaling of performance and capacity beyondthe limits of a single array, while maintaining the management simplicity of a single array. Addingmultiple arrays into the cluster linearly increases both performance and capacity, while leveragingthe resources of all arrays in the cluster. Scale-out storage is ideal for IT environments looking toeliminate individual silos of storage across their infrastructure. Multiple types and generations ofarrays can be combined in a single Nimble Storage cluster, protecting your investment.This document walks through the planning, hosting, and networking best practices associatedwith Nimble Storage’s scale-out cluster.TerminologyScale-Out Cluster: Multiple Nimble Storage arrays that are combined together and managed as asingle logical storage entity.Storage Pool: A pool is a logical collection of one or more arrays in a cluster containing volumesand their respective snapshots and clones. Volumes within a pool are striped across all the poolmember arrays.Storage Array: Arrays are a highly available purpose-built storage appliance that consists ofredundant storage controllers, disks and solid state flash storage.Planning Scale-Out Clusters and PoolsNimble Storage provides a rich set of features associated with scale-out clustering around datarebalancing, migration and host integration to get the most out of your investment. The followingsection features best practices and suggestions to follow when planning your deployment.Cluster Arrays Based on Management BoundariesA Nimble Storage scale-out cluster or “group” contains one or more individual arrays that aremanaged as a single logical storage array (up to 4 in the 2.x releases). This means that there is asingle management domain and single “pane of glass” management view. This provides theability to scale beyond a single array, managing all of the pools of storage within the cluster.BEST PRACTICES GUIDE: NIMBLE STORAGE BEST PRACTICES FOR SCALE-OUT3

Cluster boundaries can also make sense for organizational boundaries to keep data assetsseparate. Creating separate scale-out clusters of arrays might also make sense for sensitive datasuch as finance and human resources to meet security audit requirements. If you require moregranular array grouping then you should use separate Storage Pools.Adding, Removing and Merging Scale-Out Cluster ArraysChanging cluster membership is a quick process that makes storage performance and capacityavailable immediately, or removes it entirely. Moreover, it is transparent to hosts – i.e. there is nodowntime for hosts or applications. However, it is important to understand that changing poolmembership can take time while the existing data volume, snapshots and clones are rebalancedon the resulting pool arrays. For example, if you would like to remove an array from a cluster, thenyou should first evacuate it from its’ current storage pool to migrate data to the remaining poolmembers. Merging scale-out clusters is a quick process because it doesn’t require any immediatedata rebalancing; however this process may be followed by migrating volumes between pools ormerging pools which does result in data rebalancing across the member arrays.For further information about the rebalancing process, please see the following section titled“Pool Rebalancing” to fully understand the time associated with cluster and pool management.Pooling Arrays within a Scale-Out ClusterNimble Storage scale-out architecture provides pooling of arrays for horizontal scalability of bothperformance and capacity to existing storage arrays. Scale-out architecture grows linearly as yourSAN needs grow which allows you to continue to leverage older hardware, and eliminates thepenalty associated with incorrectly forecasting the future performance and capacity needs of aSAN. Pooling arrays within a cluster allows you to distribute I/O across multiple arrays whichincreases performance and allows capacity to scale across all arrays in the pool. A Nimble Storagepool consists of whole arrays and arrays are members of a single storage pool at a time.BEST PRACTICES GUIDE: NIMBLE STORAGE BEST PRACTICES FOR SCALE-OUT4

Pool Arrays for Consolidation and Fault IsolationNimble Storage scale-out clusters allow you to pool whole arrays to consolidate storage based onlogical boundaries. Pooling storage can be helpful for departmental chargebacks, fault isolationbetween production and non-production environments, or by application. Nimble’s scaling andmigration technology eliminates planned downtime when storage pools are changed.Pool Arrays with the Same Storage CapacityNimble Volumes are distributed between each array in a storage pool relative to the storagecapacity of the array. This data distribution storage architecture also has the effect of distributingI/O operations based on the data distribution. Pooling arrays of the same capacity will evenlydistribute I/O operations across all arrays in the storage pool and permit maximum performanceBEST PRACTICES GUIDE: NIMBLE STORAGE BEST PRACTICES FOR SCALE-OUT5

across all arrays. Placing arrays with significantly differing capacity within the same pool may notfully utilize the performance resources available in the pool. By creating pools that span arrays ofsimilar capacity, performance is balanced equally across all arrays in the pool which maximizesutilization of all performance resources.Pool RebalancingNimble Storage pools are rebalanced when storage objects such as arrays, pools and volumes areadded, removed or merged. Rebalancing is a non-disruptive low-impact process that allowsapplication I/O to continue uninterrupted even to the data sets being migration. Pool rebalancinggives highest priority to active data I/O and performs the rebalancing process with a lower priority.The relative speed of rebalancing is dependent upon the use case being performed and howquickly a typical user would expect the procedure to complete.Non-Disruptive Pool Rebalancing Use CasesVolume Migration - Volumes can beselectively migrated between poolsincluding their associated snapshots.Volumes support read/write during themigration process.Adding Arrays to a Pool - Addperformance and storage capacity toan existing pool.Evacuating Pools - Remove an arrayfrom a pool when it’s no longer neededor when it is being replaced by anotherarray. There must be enough capacityBEST PRACTICES GUIDE: NIMBLE STORAGE BEST PRACTICES FOR SCALE-OUT6

within the remaining pool members toaccommodate the volumes andsnapshots migrated from theevacuating pool member.Merging Pools - Useful when combiningstorage resources that were previouslyin separate pools thus increasingperformance scaling across multiplearrays.Networking Best PracticesNimble Storage Scale-out architecture allows you to manage a Storage Cluster that spreadsvolumes and their I/O requests between multiple arrays. A host cannot assume that a volume canbe accessed through paths to one specific array or another. Instead of advertising all of the iSCSIinterfaces on the array, Nimble Storage scale-out clusters advertise one IP address (iSCSIDiscovery). Volume I/O requests are redirected to the appropriate array by leveraging deepintegration with the Microsoft and VMware host operating system platforms or using iSCSIredirection. Storage network architecture is of paramount importance to ensure high-speed lowlatency communication from hosts to arrays and from array to array. The following network bestpractices will help you to create the most robust iSCSI storage network for your Nimble scale-outenvironment.BEST PRACTICES GUIDE: NIMBLE STORAGE BEST PRACTICES FOR SCALE-OUT7

SAN Bandwidth AvailabilityNimble Storage scale-out clusters perform communication between arrays to coordinateproduction I/O in-addition to performing rebalancing of data between arrays. The intra-SANcommunication and data migration is designed to be a low-impact process that won’t interferewith production I/O. However, the amount of available SAN bandwidth will impact the speed withwhich volume migration and pool rebalancing can take place.Cluster Arrays on the same SubnetsAll of the arrays within the same Group must be able to see the same sub-networks. This ensuresthat host connectivity follows the proper network pathing as well as to ensure that volume andpool migration works properly and with maximum performance.Choosing iSCSI SwitchesNetwork switches provide a critical part of an iSCSI storage area network. There are manydifferent classifications of switches and it is important to understand the characteristics thatmake a switch good for supporting iSCSI storage traffic. Use the following table when evaluatingnetwork switches:Non-blocking BackplaneA switch used for iSCSI data communicationshould have a backplane that provides enoughbandwidth to support full duplex connectivity forall ports at the same time. For example, a 24port Gigabit switch backplane should provide atleast 48 Gigabits per second of bandwidth or (1Gbps * 2 for Full Duplex * 24 Ports).Flow Control (802.3x)Flow control provides a mechanism fortemporarily pausing the transmission of data onEthernet networks when a sending nodetransmits data faster than the receiving nodecan accept it. You should enable flow control onall hosts, switch, and array ports to ensuregraceful communication between networknodes. Nimble Storage array NICs have flowcontrol enabled by default.Buffer Space per Switch PortSwitches are used to provide communicationbetween hosts and arrays and for NimbleStorage scaling communication between arrays.Each switch port should have at least 512Kilobytes of buffer memory per port to ensureBEST PRACTICES GUIDE: NIMBLE STORAGE BEST PRACTICES FOR SCALE-OUT8