DBA Consulting Blog

RAC in Containers

2012-01-20T12:48:00.002+01:00

Virtualizing Oracle RACwith Solaris Containers

Oracle Solaris Containers

An integral part of the OracleSolaris 10 Operating System, Oracle Solaris Containers isolate

software applications and servicesusing flexible, software-defined boundaries and allow many private executionenvironments to be created within a single instance of Oracle Solaris 10. Eachenvironment has its own identity, separate from the underlying hardware. Eachbehaves as if it is running on its own operating system making consolidationsimple, safe, and secure. Containers can all share CPU resources, can each havededicated CPU resources, or can each specify a guaranteed minimum amount ofresources as well as a maximum. Memory can be shared among all Oracle SolarisContainers, or each can have a specified memory cap. Physical I/O resourcessuch as disk and network can be dedicated to individual Oracle SolarisContainers, shared by some, or shared by all. Regardless of what is shared ordedicated, each virtualized environment will have isolated access to local filesystem and networking, as well as system and user processes. Ideal forenvironments, that consolidates a number of applications on a single server.The cost and complexity of managing numerous machines make it advantageous toconsolidate several applications on larger, more scalable servers. Enable moreefficient resource utilization on your system. Dynamic resource reallocationpermits unused resources to be shifted to other Oracle Solaris Containers asneeded. Fault and security isolation mean that poorly behaved applications donot require a dedicated and under-utilized system. There are two ways tocreate Oracle Solaris Containers. One of them is 'sparse root'Container, where the root file system of the Container is mounted as read-onlyfrom the global zone, occupies less space on the file system and its quick tocreate. Second one is 'whole root' Container, where the root file systemis mounted in read-write mode, all the packages required for Container areinstalled inside it. The whole root Container is like a typical system. Thereare two types of networking available for Containers. One of which is'shared-IP' where the NIC is shared with the global zone, with the shared-IPcan not plumb IP address within Container, only at the time of bootingconfigured IPs of Container are brought up. For 'exclusive-IP' type of networka dedicated NIC is assigned and the IP to this NIC can be configured within aContainer.

Oracle RAC database

Oracle RAC database is a RDBMSdatabase, which is highly available and scalable. It can dynamically adjust todynamic resource changes in the environment. Oracle RAC's data files are hostedon shared storage, be it NAS or SAN based storage connected to all theavailable systems. It's recommended to use low latency high throughput privateinterconnect for Oracle RAC inter node communication. Minimum 2 physicalsystems are required to provide high availability, to mitigate failure of onenode. However, for increased scalability and availability add additionalphysical server or node.

To host Oracle RAC in Oracle SolarisContainers environment, following points to be considered,

• Provision one Oracle Solaris Container pernode or server.

o By this high availability is achieved per instance or node level, asif one node

goes down the other instance isavailable from other node and not the same.

o Consolidate multiple such Oracle RAC databases within one set ofphysical

servers.

o In addition to that Oracle Solaris Container is being highly secureand isolated

environment, these environments couldbe assigned as independent nodes to

different DBAs managing differentOracle RAC database installations with the

Oracle Solaris Container root access.

• Name space isolation offers, configurationof different Containers of different Oracle

RAC database to be in different time zones on a given server, however,it's practiced

that the same time zone is configured across all the other nodes hostingthat particular

Oracle RAC database.

o Name space isolation offers, toconfigure these contained environments as

independent systems by configuringthem to different name server, be it DNS

or NIS etc.

o Independent process tree per Container offer process level isolation,there by

a processes running in oneContainer will not have any impact on the other

Container in the event of a crashand reboot by that process on that

environment reboots otherContainers are untouched.

• To host Oracle RAC database binaries use ZFS file system. Or bycreating Containers

environment on ZFS and host database binaries on it. Gives greatadvantages like:

o Faster deployment of Oracle binariesand setting up of the Container

environment by cloning the ZFS filesystem hosting it.

o Dynamically increase the file systemsize transparently, when the file system

is near full.

• To host the Oracle RAC database data files leverage ASM, as it offers

o Greater availability of the LUNsachieved by using redundancy. ASM does

mirror disks or LUNs acrosscontrollers there by the availability increases.

o Most optimum way to retrieve andstore the data, as Oracle has better

understanding of data.

• Flexibility to change the CPU resources based on the requirement couldbe achieved

by the host administrator, to accommodate workload needs.

Deploy Oracle RAC withoutVirtualization

To host Oracle RAC withoutvirtualization, following is considered,

• Need minimum 2 to3 nodes for high availability, Shared storage to host data files,

and the required private and public networking for this Oracle Clusterenvironment.

• However there isno scope of hosting other version or different patch level of Oracle

RAC on the

same set of systems.

• Thoughthe over-all CPU and other resource utilization might be just 20 to 25%. Tohost

another

setup with a different version another set of systems needs to beconfigured, installed

and managed.

The below drawing 1 shows that,

• There are 4 * SunSPARC T5220 servers, configured in the RAC environment, hosts

Oracle 10gR2 RAC database with 2 different databases on global zone orhost’s

operating system environment.

• Thereare 2 'Sun StorageTek 6140' array withredundant controllers. Each redundant

controller is connected to redundant HBAs on the systems. On the system,multipath IO

(MPxIO) is used to

provide high availability of connectivity to the LUNs on the disk array.

• There are 2 colorcodes used to show 2 different controllers,

• From each storagecontroller, say storage1, cables would be connected to port1 on

slot4 and slot5. Same with storage2 however port02 is used.

• MPxIO enables HAamong slot4 and slot5 and one disk name is provided instead of

two.

• Oracle AutomaticStorage Management (ASM) normal disk group is used for normal

redundancy among 2 LUNs of port1 and port2, it provides HA among thefailure of

entire storage array.

· Public networkcables are connected to switch1 and switch2 from e1000g0 and e1000g1, an IPMPgroup ipmp0 provides HA for the public network to hosts IP and VIP addresses.

· These switchesare connected to the public network, in the lab environment they can have anuplink port configured

· Privatenetwork cables are connected to switch3 and switch4 from e1000g2 and e1000g3,an IPMP group ipmp1 provides HA for the private network on the host, and the privateIP address is configured and brought up as the host comes up, and the ipmp1group also comes-up along with that.

· OracleRAC leverages this IP address as its private IP address, when the Oracleclusterware comes up.

· Theseswitches have an up-link port configured as trunk port to pass the traffic incase of a failure of one of the switches components the link fails-overtransparently to other switch.

Oracle RAC Deployment withVirtualization

The virtualized environment is OracleSolaris Containers, a built-in technology as explained earlier. To deployOracle RAC inside Oracle Solaris Containers extend the learning from theprevious chapter as the high level deployment inside this virtual environmentis seamless. Let us consider the following while deploying Oracle RAC insideOracle Solaris Containers.

The below drawing 2 is a typicaldeployment scenario, that shows,

• There are 4nodes, hosting 2 Container environments, running 2 different versions

of Oracle RAC database 10gR2 and 11gR1 on 4 different Oracle clusterwareinside

Containers.

• This shows thatone container is created per node for a given Oracle Clusterware

environment, there by HA of instances is achieved.

• At the same time,other container environment on a given failed node also goes

down, will not have major impact as only oneContainer or virtual node per cluster is

not available.

• These cont10g01to cont10g04, and cont11g01 to cont11g04, are hosted on physical

nodes port01 to port04 respectively.

• Two cores areassigned (16 threads) per Container.

• The public andprivate networking shares the same set of hardware NICs among 2

Oracle Clusterware environments without impacting each other by usingVLAN tagged

NICs. Since Oracle clusterware plumbs the VIP, exclusive-IP typeContainers are

created and VLAN tagged NICs are assigned instead of physical NICs.

• NICs e1000g0 ande1000g1 are connected to switch1 and switch2.

o Ports of these switches are configured as trunk ports to allow VLANtraffic with

VLAN tags 131,132 for Oracle 10gR2 and Oracle 11gR1 environments.

o Are public NICs, VIP is hosted on it. Plumbed by vip service of Oracleclusterware.

o For example, inside cont10g01 Container these NICs are e1000g131000and

e1000g131001, an IPMP group ipmp0 is created to provide HA at thisnetwork layer.

• NICs e1000g2 ande1000g3 are connected to switch3 and switch4.

o Ports of these switches are configured as trunk ports to allow VLANtraffic with

VLAN tags 111,112,113 for Oracle 10gR2,Oracle 11gR1, and Oracle 11gR2

environments.

o Private IPs are configured by Container and Oracle clusterware.

o For example, inside cont10g01 Container these NICs are e1000g111002and

e1000g111003, an IPMP group ipmp1 iscreated to provide HA at this network

layer.

• Forsimplification, same IPMP group names could be used across all the Container

environments of a given Clusterware.

• A set of dedicated StorageLUNs are assigned per Oracle clusterware environment, and the same set

of LUNsare configured for other Containers across all the nodes. Physical connectivityis same as

explained in the previous chapter.

Conclusion

• Oracle SolarisContainers is the best choice for consolidating various Oracle RAC

databases with different versions or patchlevels. And could be hosted on the same

operating system kernel.

• Oracle SolarisContainers enable the end users to manage resources well, pool the

resources from underutilized Containers and provision the same to therequired

Containers.

• In addition tothat the accounting feature of SRM could be leveraged to bill the end

users based on the CPUs consumed, probably a grid environment couldleverage

this feature for appropriate billing based on resource utilization.

• Dynamic resourcemanagement like memory and CPU changes would offer greater

flexibility.

• VLAN tagged NICsovercome the hard limitation of physical NICs on a server. To

support this, network switches also posses this feature.

• IPMP takes careof availability at the network stack, apart from Oracle RAC's monitoring

and timeouts to detect network failure.

• MPxIO offersseamless availability by offering single disk name for different paths of the

same disk/LUN. It manages the availability of different paths ofLUNs/disks from

storagearrays to hosts.

• ASM offers thebenefits of a cluster volume manager where disks/LUNs from different

storage arrays could be mirrored to offer HA among two different storagearrays.

Sample Container configuration

Hardware:

Sun SPARC Enterprise T5220 (M-seriesor x86 based system could also be used to host Container virtual environment).

4 * Sun SPARC Enterprise T5220 servereach configured with 1.6GHz 8 cores or 64 threads, and

64GB RAM.

Storage:

Sun StorageTek 6140 Array with dualcontroller.

OS:

Solaris10 10/09 SPARC with kernelpatch 142900-14 and it's dependent patch 143055-01

Oracle database:

• Oracle RAC 10gR210.2.0.4 with latest patch set 9352164

• Oracle RAC 11gR111.1.0.7 with latest patch set 9207257 + 9352179

Configuration files

Host system configurations files:

1. Limit ZFS cache to 1GB

edit /etc/system # add these lines

* set this value for limiting zfscache to 1G

set zfs:zfs_arc_max = 0x3E800000

2. To override the system wide limitof 1/4 of physical memory by default on S10, the

shminfo_shmmax tunable would need to be configured in /etc/system toremove that

limit. Edit /etc/system and set the value of shminfo_shmmax to the valuethat suites

the requirement.

* set the max shared memory to 24G

set shmminfo_shmmax = 0x600000000

3. Enable MPxIO

edit /kernel/drv/fp.conf # thefollowing entry

mpxio-disable="no";

If the entry is present, make sureit's set as mentioned above

to enable the MPxIO or add the entry.

For all the /etc/system values totake affect reboot the node.

Inside Containers configurationfiles:

1. Oracle shared memory configurationusing SRM facility /etc/project.

root@cont11g01:/# cat /etc/project

system:0::::

user.root:1::::

noproject:2::::

default:3::::

group.staff:10::::

user.oracle:100:Oracleproject:::process.max-sem-nsems=(privileged,4096,deny);project.max-shmmemory=(

privileged,16106127360,deny)

2. IPMP configuration

a. Public network

root@cont10g01:/# cat/etc/hostname.e1000g131000

cont10g01 group pub_ipmp0

root@cont10g01:/# cat /etc/hostname.e1000g131001

group pub_ipmp0 up

root@cont10g01:/#

b. Private network

root@cont10g01:/# cat/etc/hostname.e1000g111002

cont10g01-priv group priv_ipmp0

root@cont10g01:/# cat/etc/hostname.e1000g111003

group priv_ipmp0 up

root@cont10g01:/#

Container configuration file:

Create a file with the followingcontent to create the Containers.

#save the below content as:

config_template_to_create_cont10g01.cfg

create -b

set zonepath=/zonespool/cont10g01

set autoboot=false

setlimitpriv=default,proc_priocntl,proc_lock_memory

set scheduling-class=TS,RT,FX

set ip-type=exclusive

add net

set physical=e1000g111002

end

add net

set physical=e1000g111003

end

add net

set physical=e1000g131000

end

add net

set physical=e1000g131001

end

add capped-memory

set physical=24G

end

add dedicated-cpu

set ncpus=16

end

add rctl

set name=zone.max-swap

add value(priv=privileged,limit=25769803776,action=deny)

end

add rctl

set name=zone.max-locked-memory

add value (priv=privileged,limit=12884901888,action=deny)

end

add device

setmatch=/dev/rdsk/c5t600A0B800011FC3E00000E074BBE32EAd0s6

end

add device

setmatch=/dev/dsk/c5t600A0B800011FC3E00000E074BBE32EAd0s6

end

add device

set match=/dev/rdsk/c5t600A0B800011FC3E00000E194BBE3514d0s6

end

add device

setmatch=/dev/dsk/c5t600A0B800011FC3E00000E194BBE3514d0s6

end

add device

setmatch=/dev/rdsk/c5t600A0B800011FC3E00000E234BBE720Cd0s6

end

add device

setmatch=/dev/dsk/c5t600A0B800011FC3E00000E234BBE720Cd0s6

end

# Copy paste the above content, change the diskpaths and NIC names to suite your configuration.

root@port01:~/# zonecfg -z cont10g01 -fconfig_template_to_create_cont10g01.cfg

# This will create the zone with the name“cont10g01”.

root@port01:~/# zoneadm -z cont10g01 install

#Installation is complete, boot the Container andconfigure it.

root@port01:~/# zoneadm -z cont10g01 boot

#Login at the Container console to configure itfor the first time, set the root passwd, configure networking, configure timezone, etc. And zone/Container will reboot.

root@port01:~/# zlogin –C cont10g01

# Follow the steps on the screen to configure thezone.

Seasons Greetings

2011-12-21T11:13:00.001+01:00

Is a smartphone really a smart choice?

2011-12-09T13:17:00.000+01:00

I recently ordered a free smartphone with a very cheap subscription (sounds really Dutch). Is it a smartchoice? I can not afford an iPhone subscription or other more expensivesubscription, but after a few seconds after turning on my smart phone someother person was already making menu choices for me! Looks funny when someonegoes through the menu of my phone without me touching any part of the touchscreen. Looks like it is already hacked. It is Phone of the year in my country(what were the criteria, certainly not security).

Is there a really secure smartphone then? NO!

Do you value your privacy? Thenhere is a sad small article from InSecure Magazine. Insecure phones top 10.

The most vulnerable smartphones

Bit9 highlighted the most vulnerable popular

smartphones in use today. The devices on the

list pose the most serious security and privacy

risk to consumers and corporations.

Fifty six percent of Android phones in the

marketplace today are running out-of-date and

insecure versions of the Android operating

system software. The study found that

smartphone manufacturers such as Samsung,

HTC, Motorola and LG often launch new

phones with outdated software out of the box,

and they are slow to upgrade these phones to

the latest and most secure versions of

Android. In some cases, the phones are not

updated at all, as the manufacturers shift their

focus to newer models, leaving existing

customers stranded with insecure software.

Here is a picture with some of the common

Bugs on Mobile Apps.

The “Dirty Dozen” list includes:

1. Samsung Galaxy Mini

2. HTC Desire

3. Sony Ericsson Xperia X10

4. Sanyo Zio

5. HTC Wildfire

6. Samsung Epic 4G

7. LG Optimus S

8. Samsung Galaxy S

9. Motorola Droid X

10. LG Optimus One

11. Motorola Droid 2

12. HTC Evo 4G

Which linux is good for you? A small history of Novell and RedHat with IBM!

2011-12-05T16:25:00.004+01:00

Linux on the System z: Which Linux isRight for You?

A profile of Linux distribution partners.

April 2006 | by Ed Gauthier

In early 2000, customers became awarethat Linux* was able to run on IBM* mainframe computers. With theirreliability, security, and scalability, coupled with their suite of virtualizationfeatures and functions, customers saw mainframes as an ideal place to run multipleLinux servers to make more efficient use of their computing resource.

Customers began experimenting withearly adopter, no charge versions of the operating system that were availablefrom Marist College in Poughkeepsie N.Y. However, when customers wanted to useLinux on their mainframes to support business applications, they startedlooking for commercial distributions, backed by quality vendors to entrust withtheir mission critical workloads.

IBM System z* is fortunate to havethe support of two premier Linux distributors. Novell* SuSE and Red Hat* offerSystem z customerscustomers quality engineering, a variety of support options,and timely updates to their versions of Linux to help customers get the mostout of their mainframes while supporting Linux applications.

Which brings us to the question thatmost customers ask when they are planning to use Linux on a mainframe to host abusiness application: "Which one is right for me?"." With thatin mind, IBM Systems Magazine, Mainframe Edition, approached IBM distribution partnersNovell SuSE and Red Hat and asked them to prepare a short description of their companies,their commitment to Linux on mainframes, features, and anything else they thoughtIBM customers would like to know to make their decision easier. We were pleasedto get responses from both. We'll begin with a few words about Red Hat fromMark Spencer - a Red Hat solutions architect - and we'll move on to RayLazarine - product marketing manager for NovellSuSE.

Mark Spencer on Red Hat

Red Hat has long been a player onIBMon IBM mainframe platforms. Since the turn of the Milenium, Red Hat has been committedto providing a stable open source platform on the mainframe. From the S/390*platform, and continuing on to today's System z architecture, Red Hat has continued to provide aplatform for the delivery of mission-critical applications. This is manifest in Red Hat'scommitment to a unified code-base, a cornerstone of Red Hat Enterprise Linuxstability and predictability. This means that Red Hat uses the same source code to deliver Red Hat EnterpriseLinux on a System z9* as it does on an X86 or X86_64 platform.

Red Hat is dedicated to delivering a predictableLinux platform for mission-critical application deployment. Through it'spartnership with IBM, as well as the ISV community, Red Hat has become identified with Linux. Thispartnership provides a proven Linux computing environment clients have come toexpect from IBM and Red Hat.

By utilizing the same Linux operating system asused in the distributed computing environment, clients can further lower theirtotal cost of ownership (TCO) by taking advantage of the fact that they onlyneed to train systems administrators on Red Hat Enterprise Linux, rather thanmultiple operating systems. This focus on a single operating systems allows clients to develop highly-skilledsystems administrators who are capable of running Red Hat Enterprise Linux onwhichever platform their business requires.

In addition to its hardware commitment, Red Hatdelivers a full solution stack to System z platforms, including software thatallows for the virtualization of storage. Red Hat's Global File System (GFS) allows multiple nodes to sharea single file-system, in with both read and write capabilities. This allows forthe deployment of clustered database-based applications, as well as enablingthe clustering of other services such as NFS.

Not long ago - and even continuing today - thebig buzz-term in our industry was was "open source." Conventional wisdomstated that, as a consultant, you asked your clients not if they were usingopen source software in their environment, but where. Today, clients are askingabout virtualization. They are keenly aware of virtualized servers and arestarting to become more aware of the possibilities of virtualized storage aswell. They strive to maximize their technology investments while lowering theircomputational costs.

Many Red Hat customers are examining ways tocontinue to scale their environments. Again, conventional wisdom drives them to takeadvantage of grid computing methodology. This is the so-called"scale-out" or "distributed" model. In such a model, thereasoning goes, if you need more computational horse-power, you simply add morenodes to your grid. The obvious rebuttal to this methodology is that eachadditional node requires incrementally more power and cooling, as well as addingto the administrative overhead. These three items contribute directly to a risein the TCO of the solution, while keeping the return on investment (ROI)constant.

The opposing methodology of "scale-out"is "scale-up" where, by adding more workloads to existingcomputational resources, clients can raise the ROI of their computinginvestments. The more work a single node can accomplish with the same power,cooling and administrative overhead, the TCO lowers, but the ROI rises.However, many clients have valid concerns about "having all of their eggsin one basket".

Fortunately, IBM and Red Hat have the a solutionwhich will allow clients to take advantage of both "scale up" and"scale out" methodologies, with the combined effect of lowering theclient's TCO, while raising their ROI.

Taking advantage of IBM's built-in virtualizationtechnology, z/VM on IBM's System z architecture allows clients to not only"scale out", implementing many virtual nodes within the z/VM*environment, but it allows them to "scale up", driving up theutilization rates of their existing hardware investment. This has thenet combined effect of lowering the TCO of the solution, all-the-whileincreasing ROI.

A good example leveraging mainframevirtualization strength is the use of IBM's HiperSockets that allows that allows fastcommunications between the virtualized nodes, without having to change theguest kernel at all. This allows for the creation of solutions with massive throughput, allowing for thecreation of highly-scalable deployments of enterprise applications.

Ray Lazarine on Novell

Novell, Inc., founded in 1983, is a provider ofinfrastructure software and services to over 50,000 customers in 43 countries.With more than 20 years of experience in data center, workgroup and desktopsolutions, the 6,000 Novell employees, 5,000 partners and support centersaround the world are meeting customer requirements for identity-drivencomputing and Linux solutions. By providing enterprise-class software andsupport for commercial and open source software, Novell delivers increasedoperating flexibility and choice at a lower total cost of ownership.

Novell entered the Linux market through itsacquisition of Ximian (Linux desktop software vendor) in 2003, and SUSE* Linux,one of the leading enterprise Linux companies, in January of 2004, andultimately bringing SUSE Linux Enterprise Server to the market.

Novell began shipping SUSE LINUX EnterpriseServer 9 in August of 2004. Since then, we've been hearing from our customersand partners that SUSE Linux is a strong foundation for open source computingin the enterprise.

Some of the key features of SUSE LINUX EnterpriseServer 9 include:

· CommonCriteria Evaluation Assurance 4+ certification, the highest security certificationlevel available today

· SUSELINUX Enterprise Server 9 also Enablingenables you customers to deploymission-critical applications by including scalability, reliability,performance, and clustering services

· AllComplete supported on of IBM's System z9 platforms

· Our Linuxplatform also includes Rrobust platform support for Linux applications deliveredby third party ISVs

Novell and IBM have been working together foryears on enabling the IBM mainframe IBM mainframe platform to run Linux. Ourpartnership has resulted in a growing acceptance of SUSE Linux as a primary operating system ofchoice for mainframe customers. In fact, customers are now able to deploy SUSELinux Enterprise Server on IBM's zSeries System z platform, to leveragecapabilities in the areas of:

· BusinessResilliency, including high availability, disaster recovery, serviceability, reliability,Hiperswap, XRC and PPRC

· Security,including privacy, regulatory requirements, identity management, zSeries* hardwarequality of services, ethical hacking by research, HiperSockets and, as mentioned,Common Criteria Certification

Novell and IBM are now working together tosupport the launch of Novell's next release of the SUSE Linux platform, SUSELinux Enterprise Server 10, scheduled for availability in the second quarter of2006. Planning is now under way between Novell and the IBM zSeries System zteam to execute joint marketing campaigns that will drive further adoption of SUSELinux and the zSeries platform in the datacenter.

SUSE Linux Enterprise Server 10 is a secure,reliable and flexible server operating system for enterprise computing. Backed by Novell with aseven-year lifecycle guarantee, SUSE Linux Enterprise Server 10 offers improvedperformance and scalability, comprehensive opensource functionality, enhanced virtualization capabilities and support for abroad range of software applications and hardware platforms, including IBM'sSystem z.

The SUSE Linux Enterprise Serverplatform is created using a Novell development and testing methodology - Autobuild- which ensures that all Linux software code that Novell releases meeting meetsstrict technical standards for reliability and integrity. Using AutoBuild,Novell is able to create and deliver important software updates in hours,rather than days or weeks. AutoBuild also allows Novell to create SUSE LinuxEnterprise for multiple platforms simultaneously. As a result, you can create amore efficient IT infrastructure because you will run the same Linux operatingsystem on your mainframes, grids, clusters, workstations and desktop systems.

Novell's AutoBuild methodologyensures that you can deploy SUSE Linux Enterprise Server 10 on all of the mostimportant hardware architectures used in business with confidence, includingthe System z platform from IBM.

SUSE Linux Enterprise Server 10 alsoprovides open application programming interfaces (APIs) and other developmenttools that simplify Linux integration and customization. As a result,businesses can lower operational costs across servers, increase computingutilization, and protect corporate data.

SUSE Linux Enterprise Server 10 helpsenterprises reduce cost, improve system performance and increase flexibilitythrough a combination of software and services designed for today's mainframerequirements in the data center. It also provides an open foundation for avariety of server workloads, including enterprise database deployment, line ofbusiness applications, and mission-critical software applications.

With SUSE Linux Enterprise Server andIBM System z9, customers can build the data center of the future - today.

For Additional Information

Novell and Red Hat are IBM'sStrategic Alliance Partners for Linux and - in collaboration with IBMdevelopers and our Linux Technology Center - have each earned the respect of

IBM customers and have establishedthemselves as leaders in delivering commercial Linux distributions. For furtherinformation about Red Hat Linux, visit www.redhat.com, and for

more information about Novell SUSELinux, visit www.novell.com.

IBM Systems Magazine is a trademarkof International Business Machines Corporation. The editorial content of IBMSystems Magazine is placed on this website by MSP TechMedia under license fromInternational Business Machines Corporation.

My external RAC memory

2011-11-24T11:38:00.009+01:00

RAC Cheat sheet

This is a quick and dirty cheat sheet on OracleRAC 10g, as my experience with RAC grows I will update this section, below is abeginners guide on the commands and information that you will require toadminister Oracle RAC.

Acronyms

Files and Directories

Useful Views/Tables

Useful Parameters

Processes

General Administration

CRS Administration

Voting Disk

End of the year approaching, let use help you modernize you IT Infrastructure!

2011-11-22T15:25:00.001+01:00

It will help you save lots of money and runs your Applications much more efficient!

Website: http://www.dbaconsulting.nl

Contact: info@dbaconsulting.nl

IBM System p

2011-11-21T10:38:00.002+01:00

IBM System p for Oracle Data Warehousing

Main Advantages

■ Reduces the complexity and risk of Oracle Data Warehousing deployments

■ Incorporates IBM System p™ andSystem Storage™ solution options into IBM and Oracle Data Warehousing referenceconfigurations developed collaboratively as part of the Oracle InformationAppliance Initiative

■ Offers a family of validatedand tuned pre-sized configuration building blocks to support a range of rawdata, user and query workload requirements

■ Provides repeatable,balanced building blocks to scale-out the data warehouse

■ Seamlessly incorporatesExtract, Transform and Load (ETL) and application loads into the data warehouse

IBM System p

A reliable, scalable, resilientserver infrastructure is critical for application deployments that rely on anOracle database. IBM System p technology is a smart choice for UNIX® and Linux®operating system-based Oracle database deployments and for businesses thatdemand powerful, flexible, reliable and secure computing solutions. Anextensive system family from 1- to 64-corescalability is competitively pricedwith64-bit IBM POWER5™, POWER5+™, and POWER6™ technology designed to lowersoftware, energy and space costs through leadership performance and unique IBMcapabilities that can enable a dramatic increase in individual systemutilization. A choice of more than 10,000 IBM AIX® and Linux applicationssupports a broad array of business requirements. And models can be selected forspecific workloads, from front-end business intelligence (BI) applications toback-end data warehouse (DW) infrastructure.

System p technology leadership

IBM System p models are based onadvanced IBM POWER™ Dual- Core chip technology and deliver outstandingprice/performance, mainframe-inspired reliability features and innovativevirtualization capabilities. Add to that IBM’s uniquesimultaneous-multithreading technologyallowing two application threadsto be run at the same time, and it’s easy to see why the IBM System p serverdelivers outstanding performance.

System pvirtualization

System pvirtualization technology can dramatically increase server utilization allowingworkloads to be easily consolidated and enabling you to do more than everbefore with a single server. Available as an option on all System p models,Advanced POWER Virtualization allows you to run multiple applications on AIXand Linux operating systems at the same time in separate, secure partitionsusing as little as one-tenth of a processor — allowing a reduction in thenumber of servers and peripheral devices needed. And the System p Capacity onDemand function available on selected models is designed to scaleup andscale-down processing power and memory as business needs fluctuate with nodisruption to your business.

IBM System Storage

Disk storage is acritical element in an Oracle Data Warehousing environment. Multiple solutionoptions are available including:

• IBM System StorageDS8000™ family - designed to deliver robust, flexible, highly available, andcost-effective disk storage to support continuous operations formission-critical workloads

• IBM System StorageDS4000™ family - scalable, modular Fibre Channel disk storage designed with growth,reliability, and availability in mind, from entry-level to enterprise environmentsand performance oriented to capacity-oriented applications

For more informationregarding IBM System Storage product compatibility with the Oracle databaseplease visit:

IBM storage Oracle compatibility

IBM and the OracleInformation

Appliance Initiative

IBM has developeddata warehouse reference configurations, called Information ApplianceFoundations, as part of the Oracle Information Appliance Initiative. Thesefoundations combine Oracle Data Warehousing components with IBM System p andSystem Storage products. These Oracle Information Appliance Foundations providevalidated, balanced configurations for data warehouses that combine systemresources such as hardware, storage, I/O and networking into data warehousebuilding blocks that can be combined to address different scalability needs ina linear fashion. They’re designed to support varying loads based on varyingraw data size, concurrent user load and varying query complexity.

The development and use of repeatablebuilding blocks allows high-performance data warehouses to scale through theuse of a modular design approach to the data warehouse, ETL functionality,business intelligence tools and applications. The modular building blocks aredesigned to deploy into existing client business intelligence and OLTP (On LineTransaction Processing) infrastructures. For more information about the OracleInformation Appliance Initiative please visit:

Oracle Information Appliance Initiative

Business intelligence referencearchitecture for Oracle on IBM Systems.

Figure 1 outlines the IBM referencearchitecture for the deployment of Oracle business intelligence components onIBM Systems and IBM System Storage products. The architecture is a high-levelsystem design. It is free of implementation details and provides a high leveldescription of the solution components.

The key elements of the referencearchitecture are the:

• Software Architecture Components –these define the overall structure and relationships among the key functionalelements of the data warehouse, the infrastructure software and BI serverrepository

• Systems Architecture - this definesa proven approach for insuring that a balanced set of system resources are inplace to deliver expected performance based on the primary drivers of datawarehouse performance – including computer power, network bandwidth and storagecapacity and bandwidth

The software architecture is made upof three primary components or groups. The first group is made up of those componentsthat comprise the data warehouse. These include the relational database(Oracle’s latest Enterprise Edition release with Oracle Partitioning), storagemanagement (Oracle Advanced Storage Management product for logical volumemanagement of the database objects) and ETL functionality (Oracle WarehouseBuilder). For existing deployments the architecture is flexible enough to allowthird party ETL products such as IBM Ascential, Informatica, SAS or BusinessObjects to be used as the ETL driving mechanism.

The second group within the softwarearchitecture comprises the infrastructure including administration andmanagement (Oracle Enterprise Manager), cluster control (Oracle RealApplication Clusters (RAC)) as well as hardware (IBM System p) and storage (IBMSystem Storage).

The third software architecture group consistsof the BI server repository (Oracle BI Enterprise Edition) and the individualapplications that the client uses to access the data warehouse. Theseapplications can be specific Oracle applications or third party tools fromother vendors such as SAS, Business Objects, Cognos or Microstrategy.

The systems architecture that IBM has developedwhich supports the Oracle Information Appliance Foundation and the IBM businessintelligence reference architecture for Oracle on IBM Systems will be describedin the next section.

Functional components of the business intelligencereference architecture for Oracle on IBM Systems

The logical flow of data from the datasource to delivery can be defined as a series of processes that accommodatedata integration, data warehousing and data analytics. Each of these areasexhibit different workload and resource characteristics. These characteristics canbe defined and resource requirements articulated through best practices andworkload sizing and capacity planning methodologies.

Figure 2 outlines these different areasand defines the approach of repeatable building blocks (called Oracle nodes)for ETL, data warehousing and analytics applications. These building blocks ornodes relate back to the reference architecture and comprise the systemsarchitecture described earlier. The use of nodes in this fashion allows theclient to integrate existing ETL and application deployments into the datawarehouse using appropriate sizing techniques. Each of the nodes are designed usingIBM System p and System Storage hardware with associated interconnecttechnology to scale in an Oracle RAC-managed environment. The nodes provideflexibility allowing a deployment to start in an SMP environment and then toscale-out in an Oracle RAC deployment as the data warehouse grows. Eachfunctional node (Figure 3) is sized based on workload characteristics in termsof I/O bandwidth, memory requirement and CPU utilization. The holistic design isthat data integration nodes are balanced not only across the internal resourcesavailable but also balanced in terms of performance with the other nodes in theentire business intelligence solution. The Oracle data warehouse node is designedto provide balanced, scalable performance of the datawarehouse as multiple nodes are connected together into a RAC-enabled datawarehouse deployment.

The individual nodes arethe core components of the Oracle Information Appliance Foundation (Figure 4).The business intelligence solution is deployed based on a detailed sizingexercise which defines the number of Oracle data integration nodes, Oracle datawarehouse nodes and Oracle analytics nodes that are required. The deploymentuses the infrastructure components outlined in the reference architecture basedon IBM System p, System Storage, Oracle RAC, Oracle Partitioning and EnterpriseManager.

Sizing functional buildingblocks How to size OracleBusiness Intelligence and Oracle DataWarehousing solutions

IBM offers a process forsizing future hardware requirements when a client is looking to run OracleBusiness Intelligence and data warehouse solutions on IBM hardware. Thisprocess is based on performance data and other information gathered from theclient’s existing environment. This input is used to estimate the resourcesrequired to support one or more of the following scenarios:

• New Oracle BI or DWinstallations

• Additional applicationsfor an existing Oracle BI or DW production environment

• Migrations to new IBMhardware platforms, such as System p or System x™ servers

In order to start thesizing estimate process, follow the instructions on the cover page of theOracle Database sizing questionnaire. The questionnaire provides information onwhat needs to be completed and where to send the completed document forprocessing. Please work with your IBM representative or IBM Business Partner(DBA Consulting) in order to obtain a sizing estimate. To access the Oracledatabase sizing questionnaire please visit:

Oracle Database sizing questionnaire

Once there, select the Oracle Databasesizing questionnaire from the list of supported applicationsystems. On the sameweb page just mentioned you can access the Oracle Business Intelligence quicksizer from the same list. This is a simplified tool that provides generalsizing guidelines on possible IBM hardware configurations when running Oracle BI.It does not replace the standard IBM sizing process but can be an excellentcomplement or starting point for the sizing estimate.

The Oracle BI quick sizer is a tool developedby IBM and Oracle. It is used as an initial sizing reference with clients whowish to build a new, or extend an existing Oracle Database data warehouse. TheBI quick sizer is designed to provide a reference point for discussion aroundthe deployment of Oracle Database 10g and, where appropriate Oracle RACon IBM System p or IBM System x server technology. The Oracle BI quick sizerprovides a discussion structure whereby the client can consider a number offactors:

• Comparisons of growth scenarios aroundscale-out or scale-up strategies

• Comparisons between System x withLinux and System p with AIX solutions.

• Comparisons of servers for differentquery workloads

• Storage configurations based on amountof usable disk and Host Bus Adapters (HBAs aka Fiber Channel adapters)

In addition to sizing methodologies forOracle Data Warehousing, ETL and business intelligence applications, IBM hassizing methods and practices for many of the third party tools that can be seenin a complex heterogeneous business intelligence deployment including SAS,Business Objects, Cognos and Microstrategy. These sizing practices and methodsare used to design the analytics nodes as part of the complete deployment. Theavailability of all these sizing paths provides a comprehensive sizing solutionfor the client sizing needs in this market segment.

Oracle data warehouse node for IBM Systemp and System Storage

As noted, the Oracle data warehouse nodeis the core component in constructing a scalable data warehouse which is theunderlying driver of the business intelligence solution. Table 1 identifies theinfrastructure elements of an Oracle data warehouse node that can be found inthe Oracle BI quick sizer. The node provides 5TB of raw data warehouse spaceproviding the flexibility for a moderate deployment to grow in a scale upfashion to 5TB before then scaling outwards in an Oracle RAC deployment fordata warehouses greater than 5TB. The node is designed to provide balanced memory,I/O and processing power to efficiently execute application queries. The nodeis designed to service between 15 and 90 concurrent users depending on thecomplexity of the user queries.

Summary

One of the majorareas of concern for business is the effort involved in architecting,developing and deploying complex data warehouses and the associated components ofthe business intelligence solution around ETL and analytic applications. IBMand Oracle have developed a powerful architectural model and approach tosimplifying the deployment and reducing the risk of data warehouses based on theIBM System p for Oracle Data Warehousing. This solution fully supports theOracle Information Appliance Initiative and provides a family of optimized andvalidated presized configuration building blocks for data warehouses to supporta range of raw data, user and query workload requirements.

For more information

To explore otherSystem p and Oracle Data Warehousingsolutions or to find out more about other joint solutions from IBM and Oracle,please contact an IBM sales representative at 1 866 426-9989, or visit us at:

http://ibm.com/solutions/oracle

For more informationabout the IBM System p please visit:

http://ibm.com/systems/p

For more informationabout IBM System Storage please visit:

http://ibm.com/storage

Or contact IBMPartner: DBA Consulting

Email: info@dbaconsulting.nl

Website : http://www.dbaconsulting.nl

Blog: http://drsalbertspijkers.blogspot.com/