PowerMax and VMAX All Flash Solutions Design (SRDF Asynchronous Concepts and Operations + SRDF/A Resiliency Features)

Question 1

What must you do before adding devices to an active SRDF/A session?

Answer 1

existing devices don’t need to be suspended

new devices are excluded from consistency recalculation until they are synchronized

Question 2

What are devices that have the Consistency Exempt attribute assigned to them?

Answer 2

can be controlled independently of the other devices in the active SRDF/A session

Question 3

What are operational problems that interfere w/ SRDF/A resilience?

Answer 3

unbalanced configurations

IO spike

transient network outages

constant writes to the same location

Question 4

What is the maximum cache utilization for SRDF/A?

Answer 4

set as a percentage of the system write pending limit for the individual SRDF/A sessions

if the threshold is reached or the write pending limit is exceeded, cache resources become exhausted and the SRDF/A session drops

Question 5

What can users do to avoid cache utilization limits for SRDF/A sessions?

Answer 5

define a period to keep the session alive during which an effort may be made to free up resources

if resources continue to be exhausted the session is dropped at the expiration period

Question 6

What happens if the storage system approaches the maximum SRDF/A cache utilization threshold?

Answer 6

Delta Set Extension (DSE) offloads some or all of the delta data set

DSE is configured an independently enabled or disabled on storage systems (recommended both use same)

Question 7

What does SRDF/A DSE do?

Answer 7

extends the cache space available for SRDF/A session cycles by offloading cycle data from the cache to a preconfigured DSE storage pool

Question 8

What is the benefit of SRDF/A DSE?

Answer 8

helps SRDF/A ride through larger and longer throughput imbalances than cache based buffering alone

Question 9

For data sitting in cache when is it data sent to DSE disk?

Answer 9

data paged to disk when array write pending count crosses the DSE threshold

default threshold is 50% of the array write pending limit

when conditions become normal data is is read back from the DSE storage to cache and transmits to target array

Question 10

What is Transmit Idle in SRDF/A?

Answer 10

a feature that allows SRDF/A sessions to remain active during transient link outages

engages if feature is enabled and link limbo timer expires

Question 11

What is link limbo?

Answer 11

allows users to set a specific length of time for Enginuity to wait when a link goes down before updating the link status

Question 12

How can you extend response time on PowerMax/VMAX?

Answer 12

through RA-group level pacing

Question 13

What does RA-group level pacing do?

Answer 13

adjusts the pace of incoming host IO write rates to match the SRDF link bandwidth and transfer rates

user enabled or disabled setting

Question 14

What happens if there’s a link loss during SRDF/A sessions?

Answer 14

requires a resynchronization of the volumes in the session once the link is restored

R1 volumes continue to be written during the link loss - results in invalid mirrors of R2

Question 15

What does enabling SRDF/A immediately after link resumption causes a surge of traffic on the link?

Answer 15

due to shipping accumulated invalid mirror tracks and new data that is added w/ host writes to the SRDF/A cycles

could lead to excessive cache consumption

Question 16

When should resynchronization take place on SRDF/A?

Answer 16

during times of low activity

Question 17

What is Failover in SRDF/A?

Answer 17

operation that is used to recover from disasters and perform DR testing

Question 18

How does Failover work in SRDF/A?

Answer 18

places R1 volumes that are identified in a device group in a write-disabled state

R2 mirrors are made read/write enabled in the remote array

Question 19

What must occur before starting the Failover operation?

Answer 19

all applications running on the R1 volumes must be stopped and unmounted

Question 20

What is Failback in SRDF/A?

Answer 20

may be run on the local or remote host

workloads able to be restarted on R1 device

Question 21

What must occur before starting the Failback operation?

Answer 21

copy of R2 volumes should be snapped and unmounted

Question 22

How are SRDF/A sessions managed?

Answer 22

either individually or in groups to ensure consistent data in remote mirrors

Question 23

What is a Single Session Consistency (SSC)?

Answer 23

individual managment of one or more sessions

individual sessions are associated w/ an RA Group

Question 24

What happens during the loss of links for an SRDF/A Group?

Answer 24

causes the SRDF Pair State to reflect a condition of Transmit Idle

Question 25

How does a Multi-Session Consistency work (MSC)?

Answer 25

multiple SRDF/A Sessions

single SRDF/A daemon (storrdfd) process controls the cycle switching in the application host

Question 26

What is a limit of MSCs?

Answer 26

if one or more R1 devices in an MSC enabled consistency group cannot mirror data to their R2 devices the MSC process suspends data mirroring