Oracle Clinic – Maclean Liu的个人技术博客

Owner: 刘相兵(Maclean Liu), Mail&Gtalk: liu.maclean@gmail.com, Mobile: 13764045638, QQ: 47079569, QQ群: 23549328

ROI,HA,Consolidation,Scalability and Manageability:How Dell IT DBAs can do “MORE with Less”

作者: Maclean Liu , post on December 17th, 2010 , English Version
【本站文章除注明转载外，均为本站原创编译】
转载请注明：文章转载自: Oracle Clinic – Maclean Liu的个人技术博客 [http://www.oracledatabase12g.com/]
本文标题: ROI,HA,Consolidation,Scalability and Manageability:How Dell IT DBAs can do “MORE with Less”
本文永久地址: http://www.oracledatabase12g.com/archives/roi-ha-consolidation-scalability-and-manageability-how-dell-it-dbas-can-do-more-with-less.html

No SQLPLUS or Export privileges to DBAs
- Created Client only home
SRVCTL commands are very slow.
- Remove NFS mount parameter “actimeout =0“, not documented but advised by oracle
- actimeo=0  no caching of file and directory attributes.
- This setting causes binary startups to slow down significantly if LD_LIBRARY_PATH points to a directory on a NFS mount with the actimeo=0 flag.
ASM Instance crashes
- Note:359515.1 – Mount Options for Oracle files when used with NAS devices
- Note:279393.1 – Linux/NetApp: RHEL/SUSE Setup Recommendations for NetApp Filer Storage
Environment variables
- ORA_CRS_HOME: cluster utilities, crs_stat etc.
- SRVCTL utility for database related activities need t o point to ORACLE_HOME
- Created MGCTL utility that takes care of this
  - MGCTL utility that starts the parallel operations based on the operation
  - Move instances on demand
  - Start services after a bounce
House Keeping Scripts
- Universal script that takes case of trace files, log rotation
Patching and multiple database versions
- Create New home, path the home and then move the databases to new home
Grid agent automation script that adds the databases to the EM
Keep track of database consumption via automated reporting

Return on Investment & Consolidation
- About $1.5m hard savings on this Grid alone
High Availability
- If a RAC node is down, all the databases are moved to another node
- Move the instances to another node if the performance is an issue
- Databases are monitored via Enterprise Manager
- Escalated support for the infrastructure
Scalability
- New instances can be brought up on other nodes
- More storage can be allocated
- Never databases can be created without having to wait for provisioning
Manageability
- Well defined processes, SLAs
- Scripts for database provisioning, instance movement, house keeping
- Complete control on provisioning and de-provisioning