Oracle on Nutanix - Best Practices
The following
Summary is taken from:-
Oracle on Nutanix Best Practice Guide
The Nutanix
Distributed File System (NDFS) has a built in Information Lifecycle Management
process which ensures all the hot data is automatically written to the highest
performance SSD tier.
An in-memory read
cache is leveraged to cache frequently accessed data from all tiers.
Hot and Warm data is
kept via the NDFS SSD storage tier, whilst cold data is kept on larger and
slower HDD.
General Best Practices.
Use benchmarks to
regularly check performance is matching or exceeding expectations.
Spend time upfront
to architect the required solution for both current and future needs.
Design to deliver
consistent performance, reliability and scalability.
Use only Oracle
certified operating systems and Oracle RDBMS versions.
Use Oracle validated
packages, for database software pre-installation.
Configure Oracle
parameters
Parallel_threads_per_cpu
= 1
Db_file_multiblock_read_count
= 512
SGA
= 50 - 70% of allocated RAM (OLTP) Less for Data Warehousing.
PGA
= 10 - 20% of allocated RAM (OLTP) more for Data Warehousing.
Utilize multiple
disks for Redo Log, Archive Log and Database Tablespaces and monitor I/O wait
for any performance issues.
Use Oracle ASM for
database files, redo and archive storage each group of files being on a
different ASM disk group.
Utilize a 4MB ASM
Allocation unit size for ASM disk groups.
Configure ASM disk
groups for External Redundancy ( except Quorum groups)
Use disk mode
independent persistent for all ASM disks.
Split Redo, Archive
and Tablespaces over separate vSCSCI controllers.
Set the Linux
Maximum IO Size to match ASM AU Size in rc.local - see documentation for
details.
Enable Oracle Huge
Memory Pages
Where appropriate
use Oracle Automatic Shared Memory Management - see documentation for details.
Scale number of
database VM's vs large number of database instances and schemas per VM.
More Memory = higher
performance and less IO, avoid swapping at all costs.
Size Linux swap
partition to be big enough to handle unexpected load, monitor swapping and
increase VM memory if required.
ORACLE DB AVAILABILITY.
vSphere HA will
provide adequate level, usually 99.9% and uptime for NON-MISSION
critical/tier-1 applications.
For MISSION
critical/tier-1 applications use:-
Oracle
Data Guard or archive log shipping
Oracle
Golden Gate
Oracle
RAC
Use Oracle RMAN
backup/restore integrated solutions.
Take consistent
database snapshots/backups
When using multiplexed Redo Logs ensure the same number
of disks are provisioned for the primary and sencondary redo logs ASM disk
groups and that the disk groups are on different vSCSI controllers.
ORACLE DB MANAGEABILITY
Standardise, monitor
and maintain.
Leverage enterprise
manager and any other oracle DB monitoring solutions and virtualisation
monitoring tools.
Create standardised
data file sizes and data file growth sizes.
Pre-allocate data
files and manages size and growth pro-actively.
Comments