Under the Hood and Best Practices in one deck highlighting the most important changes you need to be aware of before getting started with Oracle RAC 12c Rel. 2. This deck was created for and first presented as part of the OTN LAD 2017 and will be updated as needed.

2. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Oracle
Real
Applica@ons
Cluster
(RAC)
12c
Release
2
–
Under
the
Hood
&
Best
Prac@ces
Markus
Michalewicz
Senior
Director
of
Product
Management,
Oracle
RAC
Development
Markus.Michalewicz@oracle.com
@OracleRACpm
hSp://www.linkedin.com/in/markusmichalewicz
hSp://www.slideshare.net/MarkusMichalewicz

3. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Safe
Harbor
Statement
The
following
is
intended
to
outline
our
general
product
direc@on.
It
is
intended
for
informa@on
purposes
only,
and
may
not
be
incorporated
into
any
contract.
It
is
not
a
commitment
to
deliver
any
material,
code,
or
func@onality,
and
should
not
be
relied
upon
in
making
purchasing
decisions.
The
development,
release,
and
@ming
of
any
features
or
func@onality
described
for
Oracle’s
products
remains
at
the
sole
discre@on
of
Oracle.
3

4. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
“Under
the
Hood”
vs.
“Best
Prac@ces”
Under
the
Hood
–
You
need
to
know
Best
Prac5ces
–
You
need
to
do
4
MPG:
25-­‐28
Front-­‐wheel
Dr.
178-­‐hp,
2.5-­‐liter
I-­‐4

5. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
“Under
the
Hood”
vs.
“Best
Prac@ces”
Under
the
Hood
–
You
need
to
know
Best
Prac5ces
–
You
need
to
do
5
MPG:
25-­‐28
Front-­‐wheel
Dr.
178-­‐hp,
2.5-­‐liter
I-­‐4

6. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Program
Agenda
Under
the
Hood
Installa@on
Changes
Under
the
Hood
Opera@onal
Changes
Installa@on
Best
Prac@ces
Opera@onal
Best
Prac@ces
Oracle
RAC
Support
in
the
Cloud
1
2
3
4
5
6

7. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Program
Agenda
Under
the
Hood
Installa@on
Changes
Under
the
Hood
Opera@onal
Changes
Installa@on
Best
Prac@ces
Opera@onal
Best
Prac@ces
Oracle
RAC
Support
in
the
Cloud
1
2
3
4
5
7

8. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
8
Simple
Flow
Oracle
Grid
Infrastructure
(GI)
&
Database
(DB)
Deployment
SoAware
installa5on
1
GI
Configura5on
2
3
DB
Crea5on
4

9. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
9
Reality
–
mul5ple
installa5on
steps
Oracle
Grid
Infrastructure
(GI)
Deployment
✔
4
t
Base
SoAware
installa5on
e.g.
12.1.0.2
1
B
PSU
applica5on
e.g.
12.1.0.2.x
2
PSU
OneOff
applica5on
e.g.
PatchXYZ
3
OneOff

10. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
10
Requirement:
gold
image
support
Oracle
Grid
Infrastructure
(GI)
Deployment
✔
Base
SoAware
installa5on
e.g.
12.1.0.2
1
4
t
PSU
applica5on
e.g.
12.1.0.2.x
2
OneOff
applica5on
e.g.
PatchXYZ
3
B
PSU
OneOff

11. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Oracle
Grid
Infrastructure
Installer
12c
Release
2
• Star@ng
with
Oracle
Grid
Infrastructure
12c
Rel.
2,
the
Grid
Installer
uses
a
gold
image
instead
of
the
tradi@onal
shiphome
as
installa@on
media
• The
zip-­‐file
available
for
download
is
therefore
a
12.2.0.1
base
gold
image
• The
launch
script
is
gridSetup.sh
or
gridSetup.bat
• User-­‐created
(customized)
gold
images
are
supported
– Customiza5on
can
include
PSUs
and
/
or
addi5onal
patches
• The
new
deployment
process
has
no
impact
on
any
post-­‐install
sodware
lifecycle
management
opera@ons;
e.g.
adding
nodes,
cloning,
patching
11
Gold
image-­‐based
installa5on

12. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
12
Requirement:
gold
image
support
Oracle
Grid
Infrastructure
(GI)
Deployment
✔
Base
SoAware
installa5on
e.g.
12.1.0.2
1
4
t
PSU
applica5on
e.g.
12.1.0.2.x
2
OneOff
applica5on
e.g.
PatchXYZ
3
B
PSU
OneOff
Create
Customized
Gold
Image

13. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Oracle
Database
Installer
will
Follow
13
Future
OUI
enhancement
More
informa@on
about
OUI
New
Features:
hSps://www.slideshare.net/MarkusMichalewicz/new-­‐in-­‐oracle-­‐universal-­‐installer-­‐oui

14. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Install
Steps
in
detail
14

15. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
• The
depreca@on
of
Oracle
Restart
has
been
withdrawn
effec@ve
July
10th
2017.
• MOS
Note
1584742.1,
formerly
@tled
“Support
Impact
of
the
Depreca@on
Announcement
of
Oracle
Restart”,
has
been
renamed
accordingly:
• Withdrawn:
Depreca@on
Announcement
of
Oracle
Restart
with
Oracle
Database
12c
15
Oracle
Restart
is
Back
–
For
Good!

16. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
• Oracle
Grid
Infrastructure
12
Rel.
2
offers
two
cluster
architectures:
1. Standalone
Cluster
• This
is
the
tradi@onal
cluster
as
offered
with
previous
versions
2. Domain
Services
Cluster
1. with
Member
Clusters
for
Databases
2. with
Member
Clusters
for
Applica@ons
• Also
supported:
• Oracle
Extended
Clusters
• hSps://www.slideshare.net/MarkusMichalewicz/
oracle-­‐extended-­‐clusters-­‐for-­‐oracle-­‐rac
16
More
Cluster
Installa@on
Choices

17. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
17
Any
New
Oracle
12c
Rel.
2
Cluster
will
be
a
Flex
Cluster
[GRID]>
crsctl
get
cluster
name
CRS-­‐6724:
Current
cluster
name
is
'SolarCluster'
[GRID]>
crsctl
get
cluster
class
CRS-­‐41008:
Cluster
class
is
'Standalone
Cluster'
[GRID]>
crsctl
get
cluster
type
CRS-­‐6539:
The
cluster
type
is
'flex'.

18. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Introduced
during
OOW
2013:
hSp://www.slideshare.net/
MarkusMichalewicz/
understanding-­‐oracle-­‐rac-­‐12c-­‐
internals-­‐oow13-­‐con8806
18
Recommended
during
OOW
2014:
hSp://www.slideshare.net/
MarkusMichalewicz/oracle-­‐
rac-­‐12102-­‐opera@onal-­‐best-­‐prac@ces
The
standard
going
forward
(every
Oracle
12c
Rel.
2
cluster
is
a
Flex
Cluster
by
default,
as
an
all-­‐Hub
Flex
Cluster
is
equivalent
to
the
previous
model
)
Oracle
Flex
Cluster
-­‐
A
Brief
Review

19. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
• Oracle
Flex
Cluster
in
12c
Rel.
1
required
the
configura@on
of
the
GNS
in
general
• One
GNS
IP
with
name
resolu@on
• GNS
in
a
Flex
Cluster
is
required
to
operate
Leaf
Nodes
as
part
of
the
cluster
• Oracle
Grid
Infrastructure
12c
Rel.
2
only
requires
a
GNS
when
Leaf
Nodes
are
added
to
the
cluster
• Either
during
installa@on
or
later
19
Oracle
Flex
Cluster
–
With
or
Without
GNS?

20. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
• Star@ng
with
Oracle
Grid
Infrastructure
(GI)
12c
Rel.
2,
all
GI-­‐related
files
need
to
be
managed
by
Oracle
ASM:
• hSps://docs.oracle.com/database/122/CWLIN/changes-­‐in-­‐
this-­‐release-­‐for-­‐oracle-­‐grid-­‐infrastructure-­‐installa@on-­‐
guide.htm#GUID-­‐8D8FBD22-­‐8744-­‐4580-­‐A711-­‐6395B0F815CF
• Desupport
of
Direct
File
System
Placement
for
Oracle
Cluster
Registry
(OCR)
and
Vo@ng
Files
• If
your
Oracle
Database
files
are
stored
on
a
shared
file
system,
then
you
can
con@nue
to
use
shared
file
system
storage
for
database
files,
instead
of
moving
them
to
Oracle
ASM
storage.
• This
affects
NFS
and
Cluster
File
System
(CFS)
customers
only.
• Transi@on
is
smooth:
• For
new
installa@ons,
provide
file
system
(FS)
and
installer
will
create
files
on
the
FS
to
be
used
for
ASM
disks.
• For
upgrades,
manual
prepara@on
is
required.
20
Oracle
ASM
is
Required
to
Manage
Clusterware
Files

21. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
• Star@ng
with
Oracle
Grid
Infrastructure
(GI)
12c
Rel.
2,
the
GIMR
can
be
installed
into
a
different
disk
group
than
vo@ng
files
and
the
OCR.
• The
size
requirements
for
the
GIMR
have
increased
significantly.
21
Oracle
Grid
Infrastructure
Management
Repository
(GIMR)
External
Redundancy

22. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
• GIMR
is
the
heart
of
the
Autonomous
Health
Framework
(AHF)
• In
the
AHF,
GIMR
is
used
as
a
centralized
diagnos@c
data
repository,
used
by
various
tools
• For
example:
• Cluster
Health
Monitor
(CHM)
• Cluster
Health
Advisor
(CHA)
• GIMR
data
can
be
used
to
op@mize
the
system
and
prevent
failures
at
run@me
using
CHA-­‐based
analysis
and
for
post-­‐mortem
failure
analysis.
22
GIMR
–
Storage
Space
Well
Worth
Inves@ng
In
Diagnos@c
Data

23. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
• Flex
Diskgroups
enable
• Quota
Management
-­‐
limit
the
space
databases
can
allocate
in
a
diskgroup
and
thereby
improve
the
customers’
ability
to
consolidate
databases
into
fewer
DGs
• Redundancy
Change
–
u@lize
lower
redundancy
for
less
cri@cal
databases
• Shadow
Copies
(“split
mirrors”)
to
easily
and
dynamically
create
database
clones
for
test/dev
or
produc@on
databases
23
Oracle
ASM
–
Flex
Diskgroup
and
Redundancy
Flex
Diskgroup
DB1
File
1
File
2
File
3
DB2
File
1
File
2
File
3
File
4
DB3
File
1
File
2
File
3
Quota
DB3
File
1
File
2
File
3
hSps://www.slideshare.net/MarkusMichalewicz/whats-­‐
new-­‐and-­‐coming-­‐in-­‐oracle-­‐asm-­‐12c-­‐rel-­‐2-­‐by-­‐jim-­‐williams

24. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Program
Agenda
Under
the
Hood
Installa@on
Changes
Under
the
Hood
Opera@onal
Changes
Installa@on
Best
Prac@ces
Opera@onal
Best
Prac@ces
Oracle
RAC
Support
in
the
Cloud
1
2
3
4
5
24

25. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
25
• Pre-­‐12.2,
node
evic@on
follows
a
rather
“ignorant”
paSern
– Example
in
a
2-­‐node
cluster:
The
node
with
the
lowest
node
number
survives.
• Customers
must
not
base
their
applica@on
logic
on
which
node
survives
the
split
brain.
– As
this
may(!)
change
in
future
releases
Node
Evic@on
Basics
hhp://www.slideshare.net/MarkusMichalewicz/oracle-­‐clusterware-­‐node-­‐management-­‐and-­‐vo5ng-­‐disks
✔
1
2

26. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
26
• Node
Weigh@ng
is
a
new
feature
that
considers
the
workload
hosted
in
the
cluster
during
fencing
• The
idea
is
to
let
the
majority
of
work
survive,
if
everything
else
is
equal
– Example:
In
a
2-­‐node
cluster,
the
node
hos@ng
the
majority
of
services
(at
fencing
@me)
is
meant
to
survive
Node
Weigh@ng
in
Oracle
RAC
12c
Release
2
Idea:
Everything
equal,
let
the
majority
of
work
survive
✔
1
2

27. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
A
three
node
cluster
will
benefit
from
“Node
Weigh@ng”,
if
three
equally
sized
sub-­‐clusters
are
built
as
s
result
of
the
failure,
since
two
differently
sized
sub-­‐clusters
are
not
equal.
27
Secondary
failure
considera5on
can
influence
which
node
survives.
Secondary
failure
considera@on
will
be
enhanced
successively.
A
fallback
scheme
is
applied
if
considera@ons
do
not
lead
to
an
ac@onable
outcome.
Let’s
Define
“Equal”
✔
Public
network
card
failure.
“Conflict”.

28. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
CSS_CRITICAL
can
be
set
on
various
levels
/
components
to
mark
them
as
“cri@cal”
so
that
the
cluster
will
try
to
preserve
them
in
case
of
a
failure.
28
CSS_CRITICAL
will
be
honored
if
no
other
technical
reason
prohibits
survival
of
the
node
which
has
at
least
one
cri@cal
component
at
the
@me
of
failure.
A
fallback
scheme
is
applied
if
CSS_CRITICAL
sewngs
do
not
lead
to
an
ac@onable
outcome.
CSS_CRITICAL
–
Fencing
with
Manual
Override
crsctl
set
server
css_cri5cal
{YES|NO}
+
server
restart
srvctl
modify
database
-­‐help
|grep
cri@cal
…
-­‐css_cri@cal
{YES
|
NO}
Define
whether
the
database
or
service
is
CSS
cri@cal
✔
Node
evic@on
despite
WL;
WL
will
failover.
“Conflict”.

29. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
• Leaf
nodes
require
at
least
one
Hub
node
in
the
cluster
to
which
they
can
connect.
• If
a
Hub
node
fails,
all
Leaf
nodes
connected
to
the
failed
Hub
node
re-­‐connect
to
another
Hub.
– Failover
is
transparent
…
• …
on
cluster
level.
• …
on
the
Leaf
nodes.
• …
for
instances
running
on
the
Leaf
nodes.
Last
but
Not
Least
–
Leaf
Node
Failover
EarthVenus
Oracle GI | HUB Oracle GI | HUB
Oracle RAC Oracle RAC
29
Mars
Oracle GI | Leaf
Oracle RAC

30. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Scalability
30

31. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
RAC
Scalability
is
Con@nuously
“Upgraded”
31
Tuning
under
the
hood
Supercharger
Performance
Exhaust
System
Air
Intake
System
Each
step
alone
provides
marginal
improvements.
Together
they
make
a
(huge)
difference.

32. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
32
Oracle
RAC
Scalability
–
A
Brief
Review
Oracle
RAC
scalability
• is
independent
of
the
number
of
nodes
• does
not
require
applica@on
changes
(unlike
sharding)
Oracle
RAC
scales
• most
of
the
enterprise
solu@ons
used
today
✔
Oracle
RAC
scales
• Oracle
Mul@tenant
Oracle
RAC
scales
• Oracle
Database
In-­‐Memory
hhp://www.slideshare.net/MarkusMichalewicz

33. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Cache
Fusion
33
A
quick
refresher
• Maximum
3-­‐way
communica@on
• Dynamic
Resource
Management
(DRM)
aSempts
to
op@mize
down
to
2-­‐way
communica@on
by
moving
the
master
to
the
instance
where
the
resource
is
frequently
accessed

34. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Small
Changes
Making
a
Huge
Difference
–
Part
1
Introducing
LMS
slaves
• Before
12c
Rel.
2,
LMS
works
on
incoming
consistent
read
requests
sequen@ally
– Sessions
reques@ng
consistent
blocks
that
require
applying
a
lot
of
undo
may
cause
LMS
to
be
busy
• With
Oracle
RAC
12c
Release
2,
LMS
offloads
work
to
‘CR
slaves’
– if
the
amount
of
UNDO
to
be
applied
exceeds
a
certain,
dynamic
threshold
• Default
is
1
slave
and
addi@onal
slaves
are
spawned
as
needed
34

35. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Small
Changes
Making
a
Huge
Difference
–
Part
2
SSD-­‐op5mized
data
retrieval
35
• RAC
determines
the
op@mal
path
to
serve
blocks
–
network
or
disk
• SSDs
and
NVMe
storage
technology
con@nue
to
drive
down
latency
• e.g.
flash
storage
may
provide
beSer
access
@mes
to
data
than
the
private
network
under
high
load
• RAC
takes
those
sta@s@cs
into
account
Query
Block
Block
Network
conges@on

36. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
1010101000010011100101001
1011100101001110010100101
0010010100001111010100101
36
• Using
Oracle
Mul@tenant,
PDBs
can
be
opened
as
singletons
(in
one
database
instance
only),
in
a
subset
of
instances
or
all
in
instances
at
once.
• If
certain
PDBs
are
only
opened
on
some
instances,
Pluggable
Database
Isola@on
– improves
performance
by
• Reducing
DLM
opera@ons
for
PDBs
not
open
in
all
instances.
• Op@mizing
block
opera@ons
based
on
in-­‐memory
block
separa@on.
– improves
availability
by
• Ensuring
that
instance
failures
of
instances
only
hos@ng
singleton
PDBs
will
not
impact
other
instances
of
the
same
RAC-­‐based
CDB.
Op@mized
Singleton
Workload
Scaling
Pluggable
Database
/
Service
Isola@on
NEW
IN
12.2

37. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
37
• Service-­‐oriented
Buffer
Cache
Access
over
@me
determines
the
data
(on
database
object
level)
accessed
by
the
service.
This
informa@on
– Is
persisted
in
the
database.
– Is
used
to
improve
data
access
performance
(e.g.
do
not
manage
data
of
a
service
in
an
instance
that
does
not
host
the
service).
– Can
be
used
to
pre-­‐warm
an
instance
cache
prior
to
a
service
startup
(fresh
start
or
reloca5on).
Op@mized
Singleton
Workload
Scaling
Service-­‐oriented
Buffer
Cache
Access

38. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Availability
38

39. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
RAC
Availability
–
Reconfigura@on
Times
MaSer
39
Tuning
under
the
hood

40. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
40
• Recovery
Buddies
• Track
block
changes
on
buddy
instance
• Quickly
iden@fy
blocks
requiring
recovery
during
reconfigura@on
• Allow
rapid
processing
of
transac@ons
ader
failures
Near
Zero
Reconfigura@on
Time
with
Recovery
Buddies
A.k.a.
Buddy
Instances

41. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
41
• Buddy
Instance
mapping
is
simple
(random)
– e.g.
I1
à
I2,
I2
à
I3,
I3
à
I4,
I4
à
I1
• Recovery
buddies
are
assigned
during
startup
• RMS0
on
each
recovery
buddy
instance
maintains
an
in-­‐memory
area
for
redo
log
change
• The
in-­‐memory
area
is
used
during
recovery
– Eliminates
the
need
to
physically
read
the
redo
Near
Zero
Reconfigura@on
Time
with
Recovery
Buddies
How
it
works
under
the
hood
Instance
I1
Instance
I2
Instance
I3
Instance
I4
Recovery
Buddy
I3
Recovery
Buddy
I4
Recovery
Buddy
I1
MyCluster
Recovery
Buddy
I2

42. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
How
Recovery
Buddies
Help
Reducing
Recovery
Time
Without
Recovery
Buddies
With
Recovery
Buddies
42
Detect
Evict
Elect
Recovery
Read
Redo
Apply
Recovery
Detect
Evict
Elect
Recovery
Read
Redo
Apply
Recovery
Up
to
4x
faster

43. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Overlooked
and
Underes@mated
–
Hang
Manager
• Customers
experience
database
hangs
for
a
variety
of
reasons
– High
system
load,
workload
conten@on,
network
conges@on,
general
errors,
etc.
• Before
Hang
Manager
was
introduced
with
Oracle
RAC
11.2.0.2
– Oracle
required
quite
some
informa@on
to
troubleshoot
a
hang
-­‐
e.g.:
• System
state
dumps
• For
RAC:
global
system
state
dumps
– Customer
usually
had
to
reproduce
“the”
hang
with
addi@onal
events
to
analyze
it
43
Why
having
a
Hang
Manager
is
useful

44. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
44
• Always
on,
as
enabled
by
default
• Reliably
detects
database
hangs
• Autonomically
resolves
them
• Considers
QoS
policies
for
Hang
Resolu@on
• Logs
all
detected
hangs
&
their
resolu@ons
Introduc@on
to
Hang
Manager
How
it
works
Session
DIAG0
EVALUATE
DETECT
ANALYZE
Hung?
VERIFY
Vic5m
QoS
Policy

45. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
45
• Hang
Manager
auto-­‐tunes
itself
by
periodically
collec@ng
instance-­‐and
cluster-­‐wide
hang
sta@s@cs
• Metrics
like
Cluster
Health/Instance
health
is
tracked
over
a
moving
average
• This
moving
average
is
considered
during
resolu@on
• Holders
wai@ng
on
SQL*Net
break/reset
are
fast
tracked
Hang
Manager
Op@miza@ons
with
Oracle
RAC
12c
Tuning
under
the
hood

46. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
46
• Early
Warning
exposed
via
(V$
view)
• Sensi@vity
can
be
set
higher
– If
the
default
level
is
too
conserva@ve
• Hang
Manager
considers
QoS
policies
and
data
during
the
valida@on
process
DBMS_HANG_MANAGER.Sensi@vity
A
new
SQL
interface
to
set
Hang
Manager
sensi@vity
Hang
Sensi5vity
Level
Descrip5on
Note
NORMAL
Hang
Manager
uses
its
default
internal
opera@ng
parameters
to
try
to
meet
typical
requirements
for
any
environments.
Default
HIGH
Hang
Manager
is
more
alert
to
sessions
wai@ng
in
a
chain
than
when
sensi@vity
is
in
NORMAL
level.

47. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Program
Agenda
Under
the
Hood
Installa@on
Changes
Under
the
Hood
Opera@onal
Changes
Installa@on
Best
Prac@ces
Opera@onal
Best
Prac@ces
Oracle
RAC
Support
in
the
Cloud
1
2
3
4
5
47

48. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
“Under
the
Hood”
vs.
“Best
Prac@ces”
Under
the
Hood
–
You
need
to
know
Best
Prac5ces
–
You
need
to
do
48
MPG:
25-­‐28
Front-­‐wheel
Dr.
178-­‐hp,
2.5-­‐liter
I-­‐4

49. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
“Under
the
Hood”
vs.
“Best
Prac@ces”
Under
the
Hood
–
You
need
to
know
Best
Prac5ces
–
You
need
to
do
49
MPG:
25-­‐28
Front-­‐wheel
Dr.
178-­‐hp,
2.5-­‐liter
I-­‐4

50. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Oracle’s
Goal
Best
Prac5ces
–
You
need
to
do
Best
Prac5ces
–
Applied
for
you
50

51. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Installa@on
Best
Prac5ces
51

52. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
• Oracle
Grid
Infrastructure
12
Rel.
2
offers
two
cluster
architectures:
1. Standalone
Cluster
• This
is
the
tradi@onal
cluster
as
offered
with
previous
versions
2. Domain
Services
Cluster
1. with
Member
Clusters
for
Databases
2. with
Member
Clusters
for
Applica@ons
• Also
supported:
• Oracle
Extended
Clusters
• hSps://www.slideshare.net/MarkusMichalewicz/
oracle-­‐extended-­‐clusters-­‐for-­‐oracle-­‐rac
52
Cluster
Architecture
–
Your
Choice

53. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Different
Architectures
–
Different
Purposes
Standalone
Cluster
Tradi@onal
architecture
for
dedicated
and
high
performance
systems;
Flex
Cluster
support
Cluster
Domain
Architecture
Domain
Services
Cluster
(DSC)
&
Member
Clusters
for
large
cluster
estate
and
storage
consolida@on
Extended
Cluster
Site-­‐awareness
for
Standalone
Clusters
and
DSC;
Member
Clusters
not
supported
53

54. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Best
Prac@ces
(BP)
Across
Architectures
54
• BP
are
designed
for
Standalone
Cluster
Deployments
• A
DSC
is
a
Standalone
Cluster
that
offers
addi@onal
services
to
Member
Clusters
• BP
for
Standalone
Clusters
apply
to
DCSs
• Member
Clusters
come
with
a
simplified
deployment.
Basic
BP
apply
to
Member
Clusters
• Extended
Clusters
for
RAC
come
with
an
addi@onal
set
of
BP:
• hSps://www.slideshare.net/MarkusMichalewicz/
oracle-­‐extended-­‐clusters-­‐for-­‐oracle-­‐rac
=
+

55. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
How
to
Apply
Best
Prac@ces?
Read
&
apply
Oracle
Documenta@on
and
My
Oracle
Support
Notes
Run
&
apply
Cluster
Verifica@on
U@lity
(CVU)
and
Oracle
Universal
Installer
(OUI)
Run
&
follow
ORAchk
55

56. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Each
“Tool”
has
a
Different
Purpose
Establish
the
base
system;
e.g.:
hardware
and
OS
minimum
requirements
Check
and
confirm
baseline;
e.g.
check
OS
for
minimum
parameter
requirements;
supports
diff
comparison
Check
for
and
then
implement
recommended
best
prac@ces
56

57. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Recommended
Usage
–
Automated
Risk
No@fica@on
1) Schedule
ORAchk
to
run
in
daemon
mode
weekly
or
daily
and
email
report
2) Iden5fy
ac@ons
easily
by
viewing
automated
comparison
of
previous
runs
3) Act
on
recommenda@ons
57

58. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
View
ORAchk
Report
• Health
score
• Summary
of
ORAchk
run
• Table
of
content
• Controls
for
report
features
• Findings
• Recommenda.ons
58

59. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
View
ORAchk
Findings
• Check
status
• Type
of
Check
• Check
Message
• Where
the
check
was
run
• Link
to
expand
details
59

60. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
• What
to
do
to
solve
the
problem
• Links
to
relevant
Knowledge
docs
• Where
recommenda@on
applies
• Where
problem
doesn’t
apply
• Example
of
data
the
recommenda@on
is
based
on
View
Recommenda@ons
60

61. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Cluster
Domain
Installa@on
Installa5on
Steps
61

62. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Step
1:
Domain
Services
Cluster
(DSC)
Installa@on

63. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Status:
DSC
–
Installed
✔
63
Domain
Services
Cluster
ASM
IO
Service
ACFS
Services
ASM
Service
Trace
File
Analyzer
(TFA)
Service
Mgmt
Repository
(GIMR)
Service
Shared
ASM
Addi5onal
Op5onal
Services
Rapid
Home
Provisioning
(RHP)
Service

64. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
64
The
Manifest
File
defines
the
characteris5cs
of
the
Member
Cluster
Step
2:
Member
Cluster
(MC)
Manifest
File
Crea@on
$ssh <DSC_host>
$cd $GRID_HOME/bin
$./crsctl create -h
Usage:
crsctl create policyset -file <filePath>
$./crsctl create member_cluster_configuration
salesmb –file $HOME/sales_mb.xml
$scp $HOME/sales_mb.xml MC_host:/tmp
Define
which
DSC
services
the
MC
can
use
Create
MC-­‐specific
Manifest
File
Copy
Manifest
File
to
MC
installa@on
node
(host)

65. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
Step
3:
Member
Cluster
(MC)
Installa@on
/tmp/sales_mb.xml
MyScan

66. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
66
Domain
Services
Cluster
ASM
IO
Service
ACFS
Services
ASM
Service
Database
Member
Cluster
Uses
ASM
Service
Trace
File
Analyzer
(TFA)
Service
Mgmt
Repository
(GIMR)
Service
Shared
ASM
Addi5onal
Op5onal
Services
Rapid
Home
Provisioning
(RHP)
Service
Private
Network
SAN
Storage
ASM
Network
Storage
Status:
MC
–
Installed
✔

67. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
67
Domain
Services
Cluster
Cluster
Domain
ASM
IO
Service
ACFS
Services
ASM
Service
Database
Member
Cluster
Uses
ASM
Service
Database
Member
Cluster
Uses
ASM
IO
Service
of
DSC
Trace
File
Analyzer
(TFA)
Service
Mgmt
Repository
(GIMR)
Service
Applica5on
Member
Cluster
GI
only
Database
Member
Cluster
Uses
local
ASM
Shared
ASM
Addi5onal
Op5onal
Services
Rapid
Home
Provisioning
(RHP)
Service
Private
Network
SAN
Storage
ASM
Network
Storage
1 2 3 4

68. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Cluster
Domain
Summary
68
• A
cluster
is
a
member
cluster,
if
it
uses
the
GIMR
service
provided
by
Domain
Services
cluster
• The
architecture
requires
a
fresh
install
– No
upgrade
from
a
12.1
GI
typical
install
to
the
Cluster
Domain
model;
conversion
planned
• Storage
configura@on
types:
– Member
Cluster
has
storage
that
is
shared
amongst
its
nodes
but
local
to
that
MC
– Member
Cluster
and
use
the
shared
storage
on
the
Domain
Services
Cluster
Domain
Services
Cluster
ASM
IO
Service
ASM
Service
TFA
GIMR
Shared
ASM
RHP

69. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
• GIMR
is
the
heart
of
the
Autonomous
Health
Framework
(AHF)
• In
the
AHF,
GIMR
is
used
as
a
centralized
diagnos@c
data
repository,
used
by
various
tools
• For
example:
• Cluster
Health
Monitor
(CHM)
• Cluster
Health
Advisor
(CHA)
• GIMR
data
can
be
used
to
op@mize
the
system
and
prevent
failures
at
run@me
using
CHA-­‐based
analysis
and
for
post-­‐mortem
failure
analysis.
• In
the
Cluster
Domain
architecture
the
GIMR
is
installed
centrally
on
the
Domain
Services
Cluster
69
GIMR
–
Centralized
Diagnos@c
Data
in
the
Cluster
Domain
Diagnos@c
Data

70. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Standalone
Cluster
Installa@on
Like
DSC
–
Consider
GIMR
Loca5on
70

71. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
GIMR
Placement
Considera@ons
Try
to
place,
not
to
relocate
the
GIMR
• For
fresh
installa@ons,
GIMR
can
be
placed
in
dedicated
diskgroups
– Default
in
12.2:
Vo@ng
Disk
diskgroup
• For
upgrades,
loca@on
of
the
GIMR
is
determined
by
the
current
loca@on
– Usually
OCR
containing
diskgroup,
unless
the
GIMR
or
the
OCR
were
relocated
• Excep@ons:
–
If
upgrading
from
a
GI
version
that
is
meant
to
contain
a
GIMR,
but
does
not,
the
GIMR
will
be
put
in
the
Vo@ng
Disk
containing
disk
group
71
External
Redundancy

72. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Program
Agenda
Under
the
Hood
Installa@on
Changes
Under
the
Hood
Opera@onal
Changes
Installa@on
Best
Prac@ces
Opera@onal
Best
Prac@ces
Oracle
RAC
Support
in
the
Cloud
1
2
3
4
5
72

73. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Cloud
Challenge
–
Scale
73
Server
Server
Server
Server
Server
Server
Server
Server
Server
Server
Server
Server
Server
Server
Server
Server
Server
Server
Server
Server
Server
Server
Server
Server
Server
Server

74. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Handling
Scale
Through
Focus
74
Server
Server
Server
Server
Server
Server
Server
Server
Server
Server
Server
Server
Server
Server
Server
Server
Server
Server
Server
Server
Server
Server
Server
Server
Server
Server
The
challenge
is
to
iden@fy
troubled
servers
in
@me
to
react
appropriately
given
human
reac@on
@me.

75. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Oracle
Autonomous
Health
Framework
• Integrates
next
genera@on
tools
running
as
components
-­‐
24/7
• Discovers
Poten@al
Issues
and
No@fies
or
takes
Correc@ve
Ac@ons
• Speeds
up
Issue
Diagnosis
and
Recovery
• Preserves
Database
and
Server
Availability
and
Performance
• Autonomously
Monitors
and
Manages
resources
to
maintain
SLAs
75
Working
for
You
Con5nuously

76. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
76
Establishes
Baseline
and
Maintains
Best
Prac5ce
Configura5ons
Oracle
Cluster
Verifica@on
U@lity
(CVU)
• Always
on
-­‐
Enabled
by
default
• Checks
O/S,
GI
and
DB
mandatory
compliance
and
best
prac@ces
with
healthcheck
op@on
• Runs
as
daemon
every
6
hours
• ASM
Best
Prac@ces
Check
• Create
Baseline
collec@ons
• Add
user-­‐specified/disable
problem
checks
• New
user-­‐friendly
report
format

77. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Schedule
ORAchk
?
Hour
(0
–
23)
?
?
?
Day
of
month
(1
–
31)
Month
(1
–
12)
Day
of
week
(0
–
6)
(0
to
6
are
Sunday
to
Saturday)
example:
orachk
-­‐set
'AUTORUN_SCHEDULE=8,20
*
*
2,5'
will
schedule
runs
on
Tuesday
and
Friday
at
08:00
&
20:00
AUTORUN_SCHEDULE
=
77
77

78. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
78
Generates
Diagnos5c
Metrics
View
of
Cluster
and
Databases
Cluster
Health
Monitor
(CHM)
• Always
on
-­‐
Enabled
by
default
• Provides
Detailed
OS
Resource
Metrics
• Assists
Node
evic@on
analysis
• Locally
logs
all
process
data
• User
can
define
pinned
processes
• Listens
to
CSS
and
GIPC
events
• Categorizes
processes
by
type
• Supports
plug-­‐in
collectors
(ex.
traceroute,
netstat,
ping,
etc.)
• New
CSV
output
for
ease
of
analysis
GIMR
ologgerd
(master)
osysmond
12c
Grid
Infrastructure
Management
Repository
OS
Data
osysmond
osysmond
OS
Data
OS
Data

79. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Introducing
Oracle
12c
Cluster
Health
Advisor
(CHA)
• Real
@me
monitoring
of
Oracle
RAC
database
systems
and
their
hosts
• Early
detec@on
of
impending
as
well
as
ongoing
system
faults
• Diagnoses
and
iden@fies
the
most
likely
root
causes
• Provides
correc@ve
ac@ons
for
targeted
triage.
• Generates
alerts
and
no@fica@ons
for
rapid
recovery
79
Proac5ve
Health
Prognos5cs
System
hSps://www.slideshare.net/MarkusMichalewicz/cluster-­‐health-­‐advisor-­‐cha-­‐deep-­‐dive-­‐by-­‐mark-­‐scardina

80. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Cluster
Health
Advisor
(CHA)
Architecture
Overview
80
OS
Data
GIMR
ochad
DB
Data
CHM
Node
Health
Prognos5cs
Engine
Database
Health
Prognos5cs
Engine
OS
Model
DB
Model
• cha
–
Cluster
node
resource
• Single
Java
ochad
daemon
per
node
• Reads
Cluster
Health
Monitor
data
directly
from
memory
• Reads
DB
ASH
data
from
SMR
w/o
DB
connec@on
• Uses
OS
and
DB
models
and
data
to
perform
prognos@cs
• Stores
analysis
and
evidence
in
the
GI
Management
Repository
• Sends
alerts
to
EMCC
Incident
Manager
per
target
EMCC
Alert

81. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Scalability
81

82. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Scale
Read-­‐only
Workload
Reliably
with
RAC
Reader
Nodes
Use
Read-­‐Only
workload
(WL)
on
read-­‐mostly
Leaf
node
instances
for
ad
hoc
data
analysis
scaled
across
hundreds
of
nodes
with
no
delay
in
accessing
updated
data,
without
any
impact
on
OLTP
performance
and
with
beSer
HA*
82
*
A
Leaf
node
failure
does
not
impact
any
other
node.

83. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Introduced
during
OOW
2013:
hSp://www.slideshare.net/
MarkusMichalewicz/
understanding-­‐oracle-­‐rac-­‐12c-­‐
internals-­‐oow13-­‐con8806
83
Recommended
during
OOW
2014:
hSp://www.slideshare.net/
MarkusMichalewicz/oracle-­‐
rac-­‐12102-­‐opera@onal-­‐best-­‐prac@ces
The
standard
going
forward
(every
Oracle
12c
Rel.
2
cluster
is
a
Flex
Cluster
by
default,
as
an
all-­‐Hub
Flex
Cluster
is
equivalent
to
the
previous
model
)
Oracle
Flex
Cluster
-­‐
A
Brief
Review
Reminder

84. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Connect
Leaf
nodes
to
storage
Leaf
nodes
for
applica@ons
do
not
require
direct
storage
access;
running
database
instances
on
Leaf
nodes
does.
84
Install
Oracle
Database
Home
on
all
nodes
and
as
needed.
If
you
ever
want
to
run
a
database
instance
on
a
Leaf
node,
it
needs
a
database
home
as
any
other
node.
Extend
public
network
to
Leaf(s)
For
RAC
Reader
Nodes
use
case
only,
enable
a
public
network
connec@on
on
Mars
by
extending
the
network
and
listener
resources
to
the
leaf.
Run
a
Database
Instance
on
a
Leaf
Node
–
Prepara@on

85. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Create
a
Policy-­‐Managed
RAC
DB
RAC
Reader
Nodes
require
a
Policy-­‐Managed
database.
Admin-­‐managed
DBs
cannot
be
extended
to
Leafs.
85
Create
database
on
HUB
nodes
the
addi@on
of
database
instances
on
Leaf
nodes
is
dynamic
and
managed
via
command
line.
Run
a
Database
Instance
on
a
Leaf
Node
–
DB
Crea@on
Policy
management
allows
for
an
easy
re-­‐
assignment
of
a
Leaf
nodes
to
other
tasks.
Serverpool
OLTP
was
pre-­‐created
using
the
oracle
user.

86. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Add
a
“Reader
Farm”
(RF)
pool
to
the
system
using
the
“add
service”
command
(dynamic).
86
Summary
Connect
Run
RAC
Reader
Nodes
–
Finaliza@on
(Re-­‐)star@ng
the
OLTPWL
Service
finalizes
the
DWHWL
service
setup.

87. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Oracle
Database
In-­‐Memory
–
ideal
for
RAC
Reader
Nodes
87
It’s
easy
to
run
Oracle
Database
In-­‐Memory
on
Leaf
Nodes:
alter system set
inmemory_size=100M
scope=spfile sid=‘*’;
Emphasizing
Leaf
Node
Usage
by
using
instance-­‐specific
sewngs
is
“work
in
progress”
Run
Oracle
Database
In-­‐Memory
on
Leaf
Node
Instances
The
IMDB
Colum
Store
will
be
ac@vated
ader
instance
restart.
A
min.
100MB
Column
Store
size
is
required.
Column
Stores
need
to
be
equally
sized
across
all
instances.
select
INST_ID,
pool,
alloc_bytes,
alloc_bytes,
used_bytes
from
GV$INMEMORY_AREA;

88. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Availability
88

89. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Improving
Planned
Maintenance
• Registered
databases
do
not
use
resources
• Star@ng
databases
may
temporarily
use
more
resources
than
running
databases
– The
same
applies
to
stopping
databases
• A
system
to
host
a
maximum
amount
of
running
databases
may
s@ll
have
issues
with
all
of
them
star@ng
or
stopping
• Solu@on:
stagger
startup
and
shutdowns
– Addi@onally:
drain
workload
prior
to
stopping
89
Drain
and
stagger
startups
as
well
as
shutdowns
Registered:
•
n
DB
instances
are
defined
to
run
on
a
machine
(poten@ally)
Running:
•
Registered
databases
and
instances
are
(concurrently)
running
(ac@ve
workload)
Star5ng:
•
Registered
databases
and
instances
start
•
Default
is
“star@ng
at
the
same
@me”.

90. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Program
Agenda
Under
the
Hood
Installa@on
Changes
Under
the
Hood
Opera@onal
Changes
Installa@on
Best
Prac@ces
Opera@onal
Best
Prac@ces
Oracle
RAC
Support
in
the
Cloud
1
2
3
4
5
90

91. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Oracle
RAC
is
fully
cer5fied
in
the
Oracle
Cloud
via
the
Exadata
Service
and
as
an
DBaaS
offering.
91
Azure
provides
an
ability
to
run
Oracle
RAC
over
a
min.
of
three
Azure
IaaS
instances.
AWS
provides
an
ability
to
run
Oracle
RAC
over
a
min.
of
three
EC2
IaaS
instances.
Support
Basics
for
Oracle
RAC
“in
the
Common
Clouds”
hSps://www.slideshare.net/MarkusMichalewicz/how-­‐to-­‐use-­‐oracle-­‐rac-­‐in-­‐a-­‐cloud-­‐a-­‐support-­‐ques@on

92. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
92
It
describes
the
ability
to
u5lize
compute
instances
as
shared
storage
in
some
way
What
does
“ability
to
run
RAC
over
3
IaaS
instances”
mean?
• Most
clouds
do
not
provide
shared
storage
na@vely.
• “Na@vely”
means
“as
part
of
their
standard
offer”.
• RAC
requires
shared
storage.
Per
RTCM:
• iSCSI
and
NFS
are
supported
for
Oracle
RAC
• Restric@on:
“The
accessed
storage
must
be
supported
by
the
system
and
storage
vendors.”
• Using
one
instance
as
an
NFS
/
iSCSI
server,
which
provides
shared
storage
to
“RAC
compute
nodes”
is
a
“supportable”
setup
following
RTCM
regula@on.
• As
long
as
the
OS,
network
and
other
layers
are
supported,
such
a
configura@on
is
supportable.
• Special
ruling
s@ll
needs
to
be
considered.

93. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
93
Strongly
recommended
to
operate
Oracle
RAC
in
any
cloud
Na@vely
Provided
Shared
Storage
• Cloud
vendors
are
asked
by
their
customers
to
support
Oracle
RAC
as
part
of
their
cloud
offerings.
• Other
than
the
Oracle
Cloud,
no
other
cloud
offering
currently
provides
shared
storage
support
as
required
by
Oracle
RAC.
• Mul@ple
non-­‐na@ve
solu@ons
have
been
used
by
different
cloud
vendors
to
provide
shared
storage.
• Such
solu@ons
include
virtual
shared
storage
as
well
as
physical
third
party
storage
solu@ons.
• As
ease
of
support
and
manageability
decreases
with
every
addi@onal
non-­‐na@ve
layer
used
to
operate
an
Oracle
RAC
database,
na@vely
provided
shared
storage
is
strongly
recommended
to
operate
Oracle
RAC
in
any
cloud.

94. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
94
Puwng
“1+1”
Together
–
Example
Can
I
use
the
FlashGrid
sodware
to
enable
Oracle
RAC
in
either
AWS
or
the
Azure
Cloud?
Is
it
supported?
Determining
Support:
1. FlashGrid
SoAware
• FlashGrid
Cloud
Area
Network
provides
a
high-­‐speed
network
overlay
with
mul@cast
and
QoS
• Local
drives
(elas5c
block
storage
or
local
SSDs)
are
shared
between
all
nodes
in
the
cluster
2. Oracle
RAC
on
Amazon
EC2
• AWS
supports
the
AWS
infrastructure
components
on
which
FlashGrid's
solu5on
relies.
• For
ques@ons
about
FlashGrid
sodware
and
support
capabili@es,
please
see
the
FlashGrid
website…
NO,
it’s
not
supported,
as
AWS
and
Azure
as
the
storage
and
server
vendors
do
not
support
FlashGrid
as
a
storage
solu@on
as
required
by
the
RTCM:

95. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
95
Special
Ruling
Affec@ng
Oracle
RAC
Support
in
the
Cloud
• The
“Licensing
Oracle
Sodware
in
the
Cloud
Compu@ng
Environment”
document
states:
• hSp://www.oracle.com/us/corporate/pricing/cloud-­‐licensing-­‐070579.pdf
• This
policy
applies
to
cloud
compuLng
environments
from
the
following
vendors:
Amazon
Web
Services
–
Amazon
Elas.c
Compute
Cloud
(EC2),
Amazon
Rela.onal
Database
Service
(RDS)
and
MicrosoF
Azure
PlaHorm
(collecLvely,
the
‘Authorized
Cloud
Environments’)
This
policy
applies
to
these
Oracle
programs.
• These
Oracle
programs
do
not
include
Oracle
RAC
as
well
as
a
few
other
op@ons.
For
more
details
see:
• hSp://www.oracle.com/us/corporate/pricing/
authorized-­‐cloud-­‐environments-­‐3493562.pdf

96. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
Oracle
con@nues
to
support
Oracle
RAC
in
the
Oracle
Cloud
via
Exadata
Service
and
as
an
DBaaS
offering,
which
is
planned
to
scale
beyond
the
current
2-­‐nodes.
96
General
rule:
As
long
as
an
IaaS
Cloud
meets
Oracle
RAC
requirements
(e.g.
network
and
storage),
this
cloud
is
in
principle
“supportable”
for
Oracle
RAC.
“Special
ruling”
regulates
most
Public
Cloud
and
IaaS
solu@ons.
Check
for
such
ruling
prior
to
considering
any
solu@on.
Oracle
RAC
in
the
Cloud
–
Where
is
it
Heading?

97. Copyright
©
2016,
Oracle
and/or
its
affiliates.
All
rights
reserved.
|
• Virtual
Machines
• Test,
Development,
Departmental
Applica@ons
• Oracle
Apps
Unlimited,
PaaS
• Compute
Shapes
–
by
OCPU,
Standard
or
High
RAM
• Block
Storage
–
by
the
GB
• Up
to
7.8TB
database
• Intensive
Test,
Development,
Departmental
Applica@ons
• Custom
Applica@ons
• Bare
Metal
Compute
Shape
–
2-­‐node
RAC
DB
System
• Local
SSD
drives
storage
• Up
to
8.6TB
database
97
• Mission
Cri@cal,
Intensive
OLTP
and
Decision
Support
• Oracle
and
Custom
Apps
• ¼,
½
and
Full
Rack
Shapes
• 2/4/8-­‐node
Exadata
DB
System
• Fixed
Storage
and
Memory
by
Shape
• Up
to
168TB
database
Oracle
RAC
Database
Cloud
Service
Infrastructure
Choice
Elas@c
Compute
Bare
Metal
Engineered
Systems