PowerFlex 4.X Appliance and Rack Design Flashcards
What are the main 5 components of PowerFlex?
SDS
SDC
SDT
MDM
FSN
What is the SDS?
software that enables a server in the cluster to contribute local storage devices to a native storage pool
What is the SDC?
block device driver that exposes shared block volumes from the SDS to the operating system - runs on the same server as the application
What is the SDT?
manages host connections and controllers that are connected over NVMe/TCP
What is the MDM?
authority that controls and tracks data storage ownership, mapping, and protection.
What is the FSN?
file service node - software component that allows the PowerFlex cluster to make data available over file based protocols
supports SMB, NFS, FTP
How does the SDS work?
owns the contributing devices and forms a protected mesh from which storage pools are created
How does the SDC work?
an application issues an IO request and the SDC fulfills the request regardless of what SDS the requested blocks physically reside on
How does the SDT work?
sends adminstrative and IO commands forward to the SDS programmatically such that the SDS is oblivious to the source of the IO - makes traffic from an SDT look like it came from an SDC
How does the MDM work?
as volumes are created the MDM provides the information application servers need to connect to the cluster’s virtualized storage
What is a protection domain?
group of nodes/SDSs that provide data isolation, security and performance benefits
What is the rule of protection domains?
a node can only participate in one PD
separate PDs can be created for different node types with unequal performance profiles
What is a fault set?
logical entity that contains a group of SDSs within a PD
What is the benefit of fault sets?
SDSs have a higher chance of failing together
by grouping SDSs into a fault set PFlex mirrors the data for a fault set on separate SDSs that are outside of fault set
availability is thus assured even if all the servers within one fault set fail simultaneously
What are the main jobs of fault sets?
used to protect against a failure
performing maintenance tasks when all data in one fault set is mirrored onto a different fault set
What are storage pools?
subset of physical storage devices within a PD - each storage device belongs to a PD
What is the recommendation when it comes to storage pools?
recommend having the same type of storage devices within a pool to ensure that the volumes are distributed across the same type of storage within a PD
What do storage pools define on PowerFlex?
the performance of its volumes based on the type and capacity of drives in the SP
What is mesh communication in PowerFlex?
when the SDC communicates with each SDS simultaneously to maximize throughput
What are the QoS functions on PowerFlex?
limits bandwidth and IOPS for non-critical applications
What does the MDM do when it comes to data layout?
applies granularity rules and protection domain rules to evenly distribute blocks of data across every node
What are the two different granularity rules on PowerFlex?
fine granularity - 4KB block size geared toward compression and snapshots
medium granularity - 1MB block size that are geared toward speed and performance
What are the two different maintenance modes on PowerFlex?
IMM (instant maintenance mode) - minutes long (perfect for rolling upgrades and reboots)
PMM (protected maintenance mode) - available for maintenance windows longer for 30 minutes
What is CloudLink on PowerFlex?
optional application that manages the encryption and providing key management
What is the PowerFlex Management Platform (PFMP)?
toolset designed to automate IT operations and lifecycle management for the entire PowerFlex cluster
What is the architecture of the PFMP?
containerized M&O based on Infrastructure Management Services, Integration Services, and Kubernetes Platform Services
combined services along with Kubernetes pods called the Kubernetes Cluster
What is the PowerFlex Management Controller (PFMC)?
infrastructure underlying the PFlex management controller in a dedicated multi-node configuration
mandatory for PFlex rack and optional for the appliance
What is the software difference between 1.0 and 2.0 for PFMC?
1.0 based on vSAN
2.0 based solely on PowerFLex w/ Kubernetes
What What is the hardware difference between 1.0 and 2.0 for PFMC?
1.0 - 3 or 5 node ESXi cluster hosting PFMP
2.0 - either single management node w/ RAID or 3 plus management nodes on management cluster
What is PowerFlex Rack?
fully engineered rack scale platform with integrated networking
What is the benefit of PowerFlex Appliance over Rack?
allows customers to use a broad set of supporting networking options - limited to these supported options
What are PowerFlex custom nodes?
servers that are configured to support a PowerFlex system
ideal for customers who prefer to build their own environments and have their own management services
also allow for broader networking options that PowerFlex appliance
What networking does PowerFlex use?
TCP/IP and NVMe/TCP to offer block and file services
provides communication between the nodes and data transfer
How does a PowerFlex custom node networking work?
DIY networking and management system
How does a PowerFlex appliance networking work?
offers flexible networking options to suit different workloads and environments
How does a PowerFlex rack networking work?
provides integrated networking to simplify deployment and management of the infrastructure
What is the communication channel options on PowerFlex?
single IP network - all communication performed on same IP network
multiple separate IP network - separate networks for management, storage (only available for PowerFlex 4.5 or greater)
What are the two supported networking vendors you can use with PowerFlex?
Cisco or Dell
What are the two network options available for PowerFlex appliance?
Full Network Automation
Partial Network Automation
What is Full Network Automation (FNA)?
requires the customer to use fully integrated (supported) switches to which PowerFlex Manager can have write access
What is Partial Network Automation (PNA)?
used when customer may not want PowerFlex Manager to have write access to the switches - or customer may want to use other enterprise grade switches
What is the benefit of FNA over PNA?
preferred option since it’s a fully tested solution
customer can use PFlex Manager to help automate server-facing ports and node operating system networking
What would be the features available to the customer using PNA?
customer can leverage existing network investments - must make sure existing switches are in Combatability Matrix
use PFlex Manager to help automate node operating system networking only
How is each PowerFlex node connected?
each node connects to dedicated redundant switch paths that enable maximum bandwidth w/ fault tolerance
nodes also connected to OOB management switch
What network connections does each PowerFlex node have?
2x links for customer/management traffic - 25/100G based on network design
2x links for storage traffic - 25/100G based on network design
1x Gbps for OOB Management (iDRAC/PMI)
traffic separation/QoS/Class of Service preconfigured to maintain systems availability/performance
What is the access aggregation topology?
three tier topology for appliance and rack
when multicabinet configuration is needed the aggregation switches are used to allow data flow between nodes in each cabinet
uses two ToR switches known as access switches per cabinet
includes one OOB management switch per cabinet
What are the automation options for access aggregation topology?
Appliance - FNA or PNA
Rack - FNA only
What is the spine-leaf aggregation topology?
two layer networking topology that is composed of spine and leaf switches
includes an OOB management switch
recommended for high performance/large scale environments
each leaf switch connects back to the spine using 100GbE
What are the switch options for leaf to node connectivity?
10G or 25G Cisco Nexus 93240YC-FX2 switch
100G Cisco Nexus 9336-fX2, 9364C, or 9364-GX switches
What is mandatory in an FNA spine/leaf aggregation topology?
Cisco spine/leaf switches
What is a logical network?
an overlay network that groups a subset of devices that share a physical network
isolates the traffic for each group
What does a logical network require?
an IP address for every each device in the network
node/switch communication requires a VLAN and data transfer protocol
What networking configuration was introduced in PowerFlex 4.5?
in rack and appliance the system can be implemented to use a switch and node multi-subnet or a multi-subnet and multi-VLAN network
What is the most common multi-subnet or a multi-subnet and multi-VLAN network implementation in PowerFlex?
when the PowerFlex management node or vMotion network spans multiple racks
each rack is its own subnet or VLAN
can also be used per a row of racks in a data center
What is the recommended maximum subnet size for each rack network?
27 subnets
What is the recommended maximum subnet size for each row network?
21 subnets
What is a subnet?
a logical division of an IP network into smaller, more efficient networks.
Subnets are used to make networks more efficient by reducing the distance network traffic needs to travel to reach its destination
What is multi-subnet?
the support of two or more subnets of the same network type with the same logical VLAN ID
What is a multi-subnet and multi-VLAN?
support of two or more subnets of the same network type with a different logical VLAN ID for each subnet
What is a caveat with multi-subnet or multi-subnet and multi-VLAN implementation?
A multi-subnet or multi-subnet and multi-VLAN implementation is supported in all PowerFlex 4.5 networks except for the PowerFlex Data, vSAN, NSX Overlay, and NSX external networks.
How does deployment work when using multi-subnet or multi-subnet and multi-VLAN implementation?
when deploying resource groups engineer associates multi-subnet or multi-subnet and multi-VLAN
nodes in the resource groups have different IPs from the subnets for hypervisor management, hypervisor migration, replication, and node management networks
How does deployment of new nodes for an existing 4.5 cluster work when using multi-subnet or multi-subnet and multi-VLAN implementation?
new nodes are added to the existing resource group
since the resource group is multi-subnet or multi-subnet and multi-VLAN that PowrFlex manages, the newly added nodes are assigned IPs from different networks
What changed in PFMC 2.0 in relation to multi-subnet or multi-subnet and multi-VLAN?
2.0 is reliant on a series of management VMs to control the PowerFlex cluster and provide the block services for each node
new VLAN subnet called pfmc-mgmt-130 (VLAN 130) is used in 4.5 for all management VMs when implementing multi-subnet
Why is VLAN 130 required on multi-subnet PFMC node implementations?
b/c of an issue w/ use of VLAN 105 on management VMs
on standard single network implementation existing VLAN 105 node management is shared between ESXi Management VMKernal and the management VMs
in multi-subnet implementation the ESXi Management is configured in different racks/rows
What is the cause of multi-subnet implementation the ESXi Management is configured in different racks/rows when using VLAN 105?
will cause the management VMs in PFMC nodes to break after vMotion migration
use of VLAN 130 remediates issue
for all other node types VLAN 105 can be used
What are the customer VLAN options for standard single network implementation?
can use either VLAN 105 or 130 for management VMs
recommended to use 130 for the PFMC nodes in a standard implementation regardless
When is it best to implement software defined networking (hybrid network)?
excellent for large/complex networks that require high levels of availability
What are the hybrid network configurations available for PowerFlex?
Cisco Application Centric Infrastructure (ACI)
VMware NSX-T Ready
What is required for a Cisco ANI hybrid network configuration?
for rack configurations only
two new customer access switches added to either the access aggregation or spine-leaf topology
two additional ports to the HCI/compute only nodes to connect ot the new customer access switches
new distributed virtual switch is added to the logical configuration to funnel customer traffic into the ACI SDN
storage nodes remain connected to standard PFlex access-aggregation or spine-leaf network
known as a dual network design (FE = SDN / BE = PowerFlex)
What is required for a NSX-T Ready hybrid network configuration?
available for both appliance and rack
includes one extra NSX-T management controller and two or more NSX-T Edge nodes
customer provides the NSX-T Edge nodes and NSX-T controller or the nodes/controller are port of the PowerFlex appliance deployment
NSX-T edge nodes can run as virtual appliances or bare metal nodes (PFlex rack only support NSX-T eduge nodes as virtual appliances)
edge nodes connect directly to the aggregation switches
What are the networking speed options on PowerFlex?
25G or 100G
What are the drive size options on PowerFlex?
3.84 or 7.68
What are the drive types on PowerFlex?
SATA, SAS, PCI, FIPS_SAS, NVMe
What is system reliability in PowerFlex?
uptime - likelihood of not having a failure event - absence of failure
What is system availability in PowerFlex?
amount of time system is nominally operating
described as a percentage
What is the availability of PowerFlex?
6 9s of availability (31.56 seconds of downtime a year)
some configurations can go up to 7 9s
How does drive size impact a system failure?
solutions require fewer drives the larger the drive capacity
but drives with larger times take longer to rebuild
is fewer drives or faster rebuilds more important to the customer
After a node failure how quickly can the node recover?
more nodes are required for smaller capacity drives
a greater number of nodes w/ smaller drives rebuilds more quickly than fewer nodes w/ larger drives
What is happening in a PowerFlex system during a rebuild?
since there is redundancy across nodes - single node or drive failures don’t result in data loss
using spare capacity built into system data residing on failed node data on failed node is rebuilt to other nodes/drives
after a node or drive is replaced the data is rebalanced
What does remote replication affect on PowerFlex?
journal capacity
factors like bandwidth between sights, RPO, and rate of change of source data may require additional capacity beyond default 10%
What does protective maintenance mode do?
creates a third copy of data on the node to protect it during a maintenance window
uses both free and spare capacity when creating a copy
recommended for maintenance windows longer than 30 minutes
What is the sizing adjustment needed for PMM?
add one node of storage to accomodate for the PMM space
What is Enterprise Project Services (EMS)?
end to end deployment platform that contains the configuration details for the installation and implementation of a PowerFlex appliance
What happens after a PowerFlex sale is complete?
project manager creates a project in EPS online and imports the order from Financial Force
What is Network Validation Tool (NVT)?
portable application that automates the validation of a customer’s network readiness
PowerFlex appliance networks that are customer provided must be validated
What does NVT do?
verifies the IP addresses of the network components and infrastructure services
run after the design is captured and before PowerFlex is deployed
