PowerMax and VMAX All Flash Solutions Design (Auto-provisioning Groups Overview + Data Reduction) Flashcards
What is the purpose of auto-provisioning groups on PowerMax and VMAX?
unable users to group initiators, FE ports, and devices together and build masking views for them
reduces complexity and chance of error
What happens when a masking view is created with auto-provisioning groups?
necessary mapping and masking operations are performed automatically to provision storage
What happens after a masking view exits with auto-provisioning groups?
any changes to the grouping propagate throughout the view
What are the 4 kinds of auto-provisioning groups?
Initiator Group
Port Group
Storage Group
Masking View
What is an Initiator Group (IG)?
collection of host initiators (HBAs) that are used together
contains WWN and IQN of host initiators
What is a Cascaded IG?
an IG that contains other IGs
What is a Port Group (PG)?
collection of FE director ports that are used together
What is a Storage Group (SG)?
collection of storage volumes/devices grouped together
can contain either volumes/devices or child Storage Groups but not both simultaneously
What is a Masking View?
helpful in provisioning storage and making changes to storage presentation
making changes to any of the groups automatically updates the masking view making provisioning easier
How is auto-provisioning managed in PowerMax/VMAX?
Unisphere/SYMCLI
What does the symaccess command do?
performs all necessary auto-provisioning needs for IGs, PG, SGs, and Masking Views
What is the symsg command used for?
does a lot of what symaccess can do
used for auto-provisioning and w/ FAST operations
What are the limitations on IGs?
may contain either HBAs or child IGs but not both simultaneously
initiator can belong to one IG at a time but once initiator is added to IG that IG can become part of another IG
port flags are set at the IG level
FCID lockdown is set per initiator basis
What are the limitations on PGs?
FE ports may belong to more than one PG
before a port can be part of a PG ACLX flag must be enabled on it
What are the limitations on SGs?
can either contain devices or other SGs but not both
logical volume can belong to more than one SG
What is the limit on Initiator Groups?
up to 64 initiators or child IGs per IG
What is the limit on PGs?
up to 16,000 PGs per array
up to 32 ports per PG
up to 16,000 devices per director
up to 4,000 LUN addresses per director
What is the limit on SGs?
up to 16,000 SGs per array
up to 64 child SGs per SG
up to 4,000 devices per SG
What is the limit on Masking Views?
up to 16,000 masking views per array
What are the parts of a Cascaded Storage Group?
Parent SG
Child SG
What is a Parent SG?
SG that contains other SGs
cannot contain devices directly - inherits the child SG devices
cannot inherit same device from more than one child
masking view only permitted on parent SG
What is a child SG?
contain devices only
can only be contained in a single parent SG
used for setting SRP, SL and workload type
What does the symaccess command do for SGs?
creates SG and simultaneously adds devices or child SGs
What does the symsg command do for SGs?
enables you to create an empty SG
enables you to set SL, workload type, and Host IO limits
What is an SRP?
Storage Resource Group
When you create a SG w/ Unisphere what is the SRP assigned?
default SRP
When you create a SG w/ SYMCLI what is the SRP assigned?
SRP not assigned to SG unless specified
When you create a SG w/ Unisphere what is the SL assigned?
default is Diamond
When you create a SG w/ SYMCLI what is the SL assigned?
no SL is assigned unless specified
What is true about Workload Type feature?
no longer used w/ PMax or VMAX
What is true about compressed data on PowerMax/VMAX?
encrypted in real time
Where is data reduction applied in PMax/VMAX?
SG level
What is the data path for data reduction on PMax/VMAX?
data reduced as it moved from system cache to BE drives using data reduction IO module on each director
What happens if the hardware IO module fails on a PowerMax director?
software based reduction is used
How is PowerMax data reduction performance optimized?
ensures that most active data is not compressed
What is the difference between PowerMax and VMAX data reduction?
PowerMaxOS running on PowerMax give compression and deduplication
PowerMax OS running on VMAX gives compression only
What is a crucial sizing piece for data reduction?
cache sizing
What is the IO Data Reduction module on PowerMax?
Data Reduction Module
What is the IO Data Reduction module on VMAX?
Compression Module
What is the Adaptive Compression Engine (ACE)?
intelligent algorithms that learn from incoming workloads and dynamically create a customized BE
ACE changes BE compression pool layout as needed to ensure optimal performance levels
How do the algorithms work on ACE?
identify busiest data in the system and leaves it uncompressed, minimizing decompression overhead
What is the compression algorithm on PowerMax?
Deflate
What is the compression algorithm on VMAX?
LZS
Between PowerMax and VMAX which one has Extended Data Compression (EDC)?
PowerMax
What is Extended Data Compression (EDC)?
compresses data that is already compressed
What metrics must data meet to be further compressed by EDC ?
data that has not been accessed for 30 days qualifies for EDC
belong to a data reduction enabled SG
must not already have been compressed by EDC
What SRP Type supports data reduction on PowerMax/VMAX?
Open Systems (FBA) SGs
What SRP Type does not support data reduction on PowerMax/VMAX?
mainframe (CKD) SGs
mixed FBA and CKD SGs
SRPs w/ external flash (FAST .X)
When is CloudArray supported w/ data reduction?
when using VMAX as long as a different SRP is used
How does deduplication work on PowerMax?
pointers used to replace the redundant copies of data and provide access to multiple sources for subsequent requests
What is the Hardware Acceleration piece of PowerMax deduplication?
inline process using same methods as ACE
data passes through IO module to create unique Hash ID
What is the Deduplication Algorithm piece of PowerMax deduplication?
identifies identical patterns based on Hash ID
performs as data passes through data reduction module
What is the Hash Table piece of PowerMax deduplication?
unique Hash IDs stored in system memory
representation of data in a deduplication relationship
What is the Deduplication Management Object (DMO) piece of PowerMax deduplication?
only exists when deduplication relationship exists
stores and manages pointers between FE devices and the single instance of data stored on disk
What is the limit on DMOs?
5 pointers or less per DMO before new one is created
What is the write IO flow on PowerMax/VMAX?
writes stores in cache when immediately received from host
at later point destaged to disk
How is write IO flow on PowerMax w/ data reduction?
cache destages to data reduction IO module instead of disk
module compresses data and creates random Hash ID
IO engine check if Hash ID exists
compressed data written to BE TDAT and hash linked to it
What is duplicate write IO flow w/ on PowerMax?
duplicate write done to a different TDEV
data reduction module compresses data and generates dedupe hash
hash table checked and duplicate is recognized
data does not need to be destaged - pointer in the table is updated w/ new write
What is the default data reduction setting on PowerMax/VMAX?
compression on by default - can be turned off by SG
How is compression enabled on a cascaded SG?
enabling compression on the parent SG enables it on all the child SGs
can disable compression per child if user wishes