RAC - Test

RAC Best Practices on Linux

Kirk McGowanTechnical Director – RAC Pack

Server Technologies

Oracle Corporation

Session id: 40136

Roland KnappPrincipal Member Technical Staff – RAC Pack

Server Technologies

Oracle Corporation

Agenda Planning Best Practices

– Architecture– Expectation setting– Objectives and success criteria– Project plan

Implementation Best Practices– Infrastructure considerations– Installation/configuration– Database creation– Application considerations

Operational Best Practices– Backup & Recovery– Performance Monitoring and Tuning– Production Migration

Planning

Understand the Architecture– Cluster terminology– Functional basics

HA by eliminating node & Oracle as SPOFs Scalability by making additional processing capacity

available incrementally– Hardware components

Private interconnect/network switch Shared storage/concurrent access/storage switch

– Software components OS, Cluster Manager, DBMS/RAC, Application Differences between cluster managers

RAC Hardware Architecture

ClusteredDatabase

Servers

Mirrored Disk Subsystem

High Speed Switch or

Interconnect

Hub or Switch Fabric

NetworkCentralized Management

Console

Storage Area Network

Low Latency Interconnectie. VIA or Proprietary

Users

No SinglePoint Of Failure

Shared CacheShared Cache

RAC Software Architecture

Shared Disk DatabaseShared Disk Database

Shared Memory/Global Area

shared shared SQLSQL

log log bufferbuffer

. . .. . .Shared Memory/Global Area

shared shared SQLSQL

log log bufferbuffer

Shared Memory/Global Area

shared shared SQLSQL

log log bufferbuffer

Shared Memory/Global Area

shared shared SQLSQL

log log bufferbuffer

GES&GCSGES&GCS GES&GCSGES&GCS GES&GCSGES&GCS GES&GCSGES&GCS

Shared Data Model

RAC on Linux HW & SW Components

public network

Node1a

shared storage

redo log instance 1 …

redo log instance 2 …

control files

database files

Node2a

Unbreakable Linux Unbreakable Linux

ORACM ORACM

clusterinterconnect

cache to cache

N3 N4 Nn

Oracle 9i RACinstance 1

Oracle 9i RACinstance 2

DB cache DB cache

concurrent access from

every node = “scale out”

more nodes = higher

availability

Linux Cluster Hardware

Cluster interconnects – FastEthernet, Gigabit Ethernet

Public networks – Ethernet, FastEthernet, Gigabit Ethernet

Memory, swap & CPU Recommendations– Each server should have a minimum of 512Mb of memory,

at least 1Gb swap space, and two CPUs.

Fiber Channel, SCSI, or NAS storage connectivity

Unbreakable Linux Distributions

Red Hat Enterprise Linux AS and ES United Linux 1.0

– SuSE Linux Enterprise Server 8 (SuSE Linux AG)– Conectiva Linux Enterprise Edition (Conectiva S.A.)– SCO Linux Server 4.0 (The SCO Group)– Turbolinux Enterprise Server 8 (Turbolinux)

Oracle will support Oracle products running with other distributions but will not support the operating system.

RAC Certification for Unbreakable Linux

Certification – Enterprise class OS distribution (e.g. RH AS, United Linux

1.0)– Clusterware (Oracle Cluster Manager only)– Network Attached Storage (e.g. Network Appliance filers)– Most SCSI and SAN storage are compatible – 32 bit and 64 bit Itanium 2 Intel based servers are certified.

For more details on software certification: http://technet.oracle.com/support/metalink/content.html

Discuss hardware configuration with your HW vendor

Linux IA64 requirements

Operating System Requirements – Red Hat Linux Advanced Server 2.1 operating system with

kernel 2.4.18-e.14.ia64.rpm – glibc 2.2.4-29 – Gnu gcc 2.96.0 release – Linux Header Patch 2.4.18 (available from Intel) – asynch libraries libaio-0.3.92-1 – (Oracle9i Release Notes

Release 2 (9.2.0.2.0) for Linux Intel on Itanium (64-bit) Part No. B10567-02 )

Set Expectations Appropriately

If your application will scale transparently on SMP, then it is realistic to expect it to scale well on RAC, without having to make any changes to the application code.

RAC eliminates the database instance, and the node itself, as a single point of failure, and ensures database integrity in the case of such failures

Planning: Define Objectives

Objectives need to be quantified/measurable– HA objectives

Planned vs unplanned Technology failures vs site failures vs human errors

– Scalability Objectives Speedup vs scaleup Response time, throughput, other measurements

– Server/Consolidation Objectives Often tied to TCO Often subjective

Build your Project Plan Partner with your vendors

– Multiple stakeholders, shared success Build detailed test plans

– Confirm application scalability on SMP before going to RAC optimize first for single instance

Address knowledge gaps and training– Clusters, RAC, HA, Scalability, systems management– Leverage external resources as required

Establish strict System and Application Change control– Apply changes to one system element at a time– Apply changes to first to test environment– Monitor impact of application changes on underlying system components

Define Support mechanisms and escalation procedures

Agenda Planning Best Practices

– Architecture– Expectation setting– Objectives and success criteria– Project plan

Implementation Best Practices– Infrastructure considerations– Installation/configuration– Database creation– Application considerations

Operational Best Practices– Backup & Recovery– Performance Monitoring and Tuning– Production Migration

Infrastructure Considerations

Architecture/Design– Eliminate SPOFs (Single Points of Failure)– Workload Distribution (load balancing) strategy– Systems management framework for monitoring and managing to

SLAs Hardware/Software

– Processing nodes – sufficient CPU to accommodate failure– Scalable I/O Subsystem

Use S.A.M.E.– Private Interconnect

Gige, UDP, switched– Patch levels and certification

Impementation Flowchart

Configure HW

Configure private interconnect

Install Unbreakable Linux

Install cluster manager 9.2.0.1

Install Oracle 9.2.0.1

Install 9.2.0.3 cluster manager

Install Oracle 9.2.0.3

Create database

Configure storage and install OCFS

Boot

Choose Language

Select Keyboard & Mouse

Choose – Advanced Server Option

Use DRUID for Partition Setup

Select Boot Loader

Configure Network

Configure Timezone

Account Configuration

Select Graphic Mode

Boot Floppy Creation

Installation Complete / Reboot

Installation Flowchart for Red Hat Linux AS 2.1

Install tips for Red Hat Linux AS 2.1

As documented in:– “Tips and Techniques: Install and Configure Oracle9i on Red Hat

Linux Advanced Server” by Deepak Patel, Oracle http://otn.oracle.com/tech/linux/pdf/installtips_final.pdf

Boot options – Always use Advanced Server install. As needed install required

packages. CD 1 to 3 has all rpm packages. CD 3 and 4 has source packages. CD 5 includes docs.

Memory– Based on physical memory on machine smp or enterprise kernel

is installed. ( <= 4 GB smp kernel and > 4 GB enterprise kernel ) Post Installation

– Add users, configure network and other administrative tasks after installation.

Install tips for United Linux 1.0

You must install the latest UnitedLinux kernel update! Oracle was certified against an update kernel, the original UL-1.0 kernel is NOT certified!

After installing United Linux 1.0, install Service Pack 2a from:

ftp://suse.us.oracle.com/pub/suse/i386/unitedlinux-1.0-iso/

You will also need to have the basic developments tools installed, like make, gcc_old(2.95.3), and the binutils package.

Full installation instructions: ftp://ftp.suse.com/pub/suse/i386/supplementary/commercial/Oracle/docs/920_sles8_install.pdf

Install tips for United Linux 1.0

Install the orarun.rpm package from either the SP2 CD

– <mountpoint>/UnitedLinux/i586/orarun-1.8-18.i586.rpm

or from – ftp://ftp.suse.com/pub/suse/i386/supplementary/

commercial/Oracle/sles-8/orarun.rpm

orarun.rpm update kernel (ie. shmmax, shmmin) UDP settings (256K) Installs and configures hangcheck-timer

Prepare Linux Environment

Follow these steps on EACH node of the cluster

– Set Kernel parameters in /etc/sysctl.conf

– Add hostnames to /etc/hosts file

– Establish file system or location for ORACLE_HOME (writable for oracle userid)

– Setup host equivalence for oracle userid (.rhosts)

Install the rpm’s on all nodes

Run ocfstool as root(configures /etc/ocfs.conf)

on all nodes

Run load_ocfs(insmod will load ocfs.o)

on all nodes

Create partition on the primary node

Run ocfstool to format andmount your new filesystem

Mount the new filesystemon all nodes

Edit rc.local or equivalent addload_ocfs and ‘mount –t

ocfs <device> <mountpoint’

Download the latest OCFS rpm’s from

www.ocfs.org

Installation Flowchart for OCFS

OCFS and Unbreakable Linux

Redhat– currently ships 4 flavors of the AS 2.1 kernel, viz., UP, SMP,

Enterprise and Summit (IBM x440)– Oracle provides a separate OCFS module for each of the kernel

flavors – Minor revisions of the kernel do not need a fresh build of ocfs– e.g., ocfs built for e.12 will work for e.16, e.18, etc.

United Linux– United Linux ships 3 flavors of its kernel, for the 2.4.19-64GB-SMP,

the 2.4.19-4GB and the 2.4.19-4GB-SMP kernel– OCFS 1.0.9 is supported on UL 1.0 Service Pack 2a or higher– OCFS build is not currently upward compatible with kernel (pre

SP3) must ensure OCFS build exists for each new Kernel version prior to upgrading kernel

OCFS and RAC

Maintains cache coherency across nodes for the filesystem metadata only

– Does not synchronize the data cache buffers across nodes, lets RAC handle that

OCFS journals filesystem metadata changes only– Filedata changes are journalled by RAC (log files)

Overcomes some limitations of raw devices on Linux– No limit on number of files– Allows for very large files (max 2TB)– Max volume size 32G (4K block) to 8T (1M block)

Oracle DB performance is comparable to raw kernel e.25 is strongly recommended for use with OCFS 1.0.9

(remove old kernel tuning parameters)

Install Tips for OCFS

Ensure OCFS rpm corresponds to kernel version– uname –r (i.e. 2.4.19-4GB)

Remember to also download rpm’s for OCFS “Support Tools” and “Additional Tools”

Download the dd/tar/cp rpm that supports o_direct Use rpm –Uv to install all 4 rpm’s on all nodes Use OCFS for Oracle DB files only, not Oracle

binaries (OCFS 1.0.x was not designed as a general purpose filesystem).

Installation Flowchart for oracm and Oracle

Install the oracm from the 9.2.0.1 CD-ROM

Configure ocmargs.oraand cmcfg.ora

Load the softdog and start with ./ocmstart.sh the cluster

manager on both nodes

Install 9.2.0.1 softwarewith the RAC option

Kill the oracm and watchdog process

modify ocmargs.oraand cmcfg.ora

(remove watchdog)

Load the hangcheck-timermodule with lsmod

Install the oracm from the 9.2.0.3 patchset

Start with ./ocmstart.sh the cluster manager

Install the 9.2.0.3patchset

Configure private interconnect and quorum device

Fix empty directory bug

Oracle Instance

Cluster Manager (including Node Monitor)

Hangcheck-timer

User-mode

Kernel-mode

Oracm maintains both, node status view and instance status view.

The hangcheck-timer monitors the kernel for hangs, and resets the node if needed.

Hangcheck NM, and CM Flow(After V9.2.0.2)

Post Installation

To enable asynchronous I/O must re-link Oracle to use skgaioi.o

Adjust UDP send / receive buffer size to 256K Larger Buffer Cache

– Create an in-memory file system on the /dev/shm (mount -t shm shmfs -o size=8g /dev/shm)

– To enable the extended buffer cache feature, set the init.ora paramter USE_INDIRECT_DATA_BUFFERS = true

Increasing Address Space – Default 1.7 GB of address space for its SGA.– See Metalink Note: 200266.1 for details and a sample

program.

Create RAC database using DBCA

Create Database Use DBCA to simplify DB creation

– Start gsd ( global services daemon ) on all nodes, if it is not already running.

Set MAXINSTANCES, MAXLOGFILES, MAXLOGMEMBERS, MAXLOGHISTORY, MAXDATAFILES (auto with DBCA)

Create tablespaces as locally Managed (auto with DBCA) Create all tablespaces with ASSM (auto with DBCA) Configure automatic UNDO management (auto with DBCA) Use SPFILE instead of multiple init.ora’s (auto with DBCA)

Validate RAC Configuration Instances running on all nodes

SQL> select * from gv$instance RAC communicating over the private Interconnect

SQL> oradebug setmypid SQL> oradebug ipcSQL> oradebug tracefile_name

/home/oracle/admin/RAC92_1/udump/rac92_1_ora_1343841.trc – Check trace file in the user_dump_dest:

SSKGXPT 0x2ab25bc flags info for network 0 socket no 10 IP 204.152.65.33 UDP 49197 sflags SSKGXPT_UP info for network 1 socket no 0 IP 0.0.0.0 UDP 0 sflags SSKGXPT_DOWN

RAC is using desired IPC protocol: Check Alert.log... cluster interconnect IPC version:Oracle UDP/IP IPC Vendor 1 proto 2 Version 1.0 PMON started with pid=2 ...

Use cluster_interconnects only if necessary

Configure srvconfig / srvctl SRVCTL uses information from srvconfig

– Reads $ORACLE_HOME/srvm/config /srvConfig.loc information File can be a RAW Device or OCFS file

Srvconfig -init gsd must be running Add ORACLE_HOME

– $ srvctl add database -d db_name -o oracle_home [-m domain_name] [-s spfile]

Add instances (for each instance enter the command)– $ srvctl add instance -d db_name -i sid -n node

Application Deployment

Same guidelines as single instance– SQL Tuning– Sequence Caching– Partition large objects– Use different block sizes– Tune instance recovery– Avoid DDL– Use LMT’s and ASSM as noted earlier

Agenda Planning Best Practices

– Architecture– Expectation setting– Objectives and success criteria– Project plan

Implementation Best Practices– Infrastructure considerations– Installation/configuration– Database creation– Application considerations

Operational Best Practices– Backup & Recovery– Performance Monitoring and Tuning– Production Migration

Operations

Same DBA procedures as single instance, with some minor, mostly mechanical differences.

Managing the Oracle environment– Starting/stopping cluster services (ocmstart.sh)– Starting/stopping gsd– Managing multiple redo log threads

Startup and shutdown of the database– Use srvctl

Backup and recovery Performance Monitoring and Tuning Production migration

Operations: srvconfig / srvctl Use SRVCTL to administer your RAC database

environment. – OEM and the Oracle Intelligent Agent use the configuration

information that SRVCTL generates to discover and monitor nodes in your cluster.

Global Services Daemon (GSD) receives requests from SRVCTL to execute administrative job tasks, such as startup or shutdown.

– GSD must be started on all the nodes in your RAC environment so that the manageability features and tools operate properly. (GSDCTL)

Operations: Backup & Recovery

RMAN is the most efficient option for Backup & Recovery– Managing the snapshot control file location.– Managing the control file autobackup feature.– Managing archived logs in RAC – choose proper archiving scheme. – Node Affinity Awareness

RMAN and Oracle Net in RAC apply– you cannot specify a net service name that uses Oracle Net

features to distribute RMAN connections to more than one instance.

Oracle Enterprise Manager– GUI interface to Recovery Manager

Performance Monitoring and Tuning Tune first for single instance 9i Use Statspack:

– Separate 1 GB tablespace for Statspack– snapshots at 10-20 min intervals during stress testing, hourly during normal

operations– Run on all instances, staggered

Supplement with scripts/tracing – Monitor V$SESSION_WAIT to see which blocks are involved in wait events– Trace events like 10046/8 can provide additional wait event details– Monitor Alert logs and trace files, as on single instance

Oracle Performance Manager RAC-specific views Supplement with System-level monitoring

– CPU utilization never 100%– I/O service times never > acceptable thresholds– CPU run queues at optimal levels

Performance Monitoring and Tuning Obvious application deficiency on a single node can’t be solved by

multiple nodes.– Single points of contention. – Not scalable on SMP– I/O bound on single instance DB

Tuning on single instance DB to ensure applications scalable first– Identify/tune contention using v$segment_statistics to identify objects

involved– Concentrate on the top 5 Statspack timed events if majority of time is

spent waiting– Concentrate on bad SQL if CPU bound

Maintain a balanced load on underlying systems (DB, OS, storage subsystem, etc. )

– Excessive load on individual components can invoke aberrant behaviour.

Performance Monitoring and Tuning

Deciding if RAC is the performance bottleneck– Amount of Cross Instance Traffic

Type of requests Type of blocks

– Latency Block receive time buffer size factor bandwidth factor

Production Migration

Adhere to strong Systems Life Cycle Disciplines– Comprehensive test plans (functional and stress)– Rehearsed production migration plan– Change Control

Separate environments for Dev, Test, QA/UAT, Production

System AND application change control Log changes to spfile

– Backup and recovery procedures– Security controls– Support Procedures

Next Steps….

Recommended sessions– List 1 or 2 sessions that complement this session

Recommended demos and/or hands-on labs– List of or two demos or labs that will let them see this

product in action. See Your Business in Our Software

– Visit the DEMOgrounds for a customized architectural review, see a customized demo with Solutions Factory, or receive a personalized proposal. Visit the DEMOgrounds for more information.

Relevant web sites to visit for more information– List urls here.

Reminder – please complete the OracleWorld online session survey

Thank you.

Resources RedHat Linux

– http://www.redhat.com/oracle/

Linux Center - Technical White Papers & Documentation– http://otn.oracle.com/tech/linux/tech_wp.html

“Tips and Techniques: Install and Configure Oracle9i on Red Hat Linux Advanced Server” by Deepak Patel, Oracle Corporation

• http://otn.oracle.com/tech/linux/pdf/installtips_final.pdf

“Tips and Techniques: Install and Configure Oracle9i on SLES8 / United Linux 1.0

• http://www.suse.com/en/business/certifications/certified_software/oracle/db/9iR2_sles8.html

United Linux 1.0 Resources

United Linux– http://www.unitedlinux.com

SuSE– http://www.suse.com/us/business/products/server/sles/index.html

Connectiva– http://www.connectiva.com

SCO Group (Formerly Caldera System) - http://www.ebizenterprises.com/page1.asp?p=463

TurboLinux– http://www.turbolinux.com/

Recommended one-off patches

Bug 2820871 - ORA-29740 NODE EVICTION DESIGN ALGORITHM AND ABRUPT TIME CHANGEARU: 9.2.0.3 ARU 4161735 completed for LINUX Intel 

Bug 2420930 - GET ORA-600 [KSXPMPRP1] DURING STARTUP IN RAC MODE WITH LARGER BUFFERS. This was mysteriously included in 9.2.0.2, but not in 9.2.0.3. Bug 2875050 was opened for this issue. ARU: 9.2.0.3 ARU 4202164completed for LINUX Intel  

Bug 2420930 - GET ORA-600 [KSXPMPRP1] DURING STARTUP IN RAC MODE WITH LARGER BUFFERS Bug 2922471 – Fractured block found during crash/instance recovery. Not an Oracle bug. Do not use ‘intr’ for the mount option.

Recommended one-off patches

Bug:2844009 - MISSING LIBCXA.SO.3 LIBRARY ISSUE IN PSR 9203. ARU: 9.2.0.3 ARU 4046387 completed for LINUX Intel 

Bug 2779294 – node_list does not populated into oraInventory/ContentsXML/inventory.xml. opatch install will only apply to local node. Workaround is editing inventory.xml documented in bug 2742686.

  Bug 2646914, 2675090, 2706220 and 2695783 - ORA-600

[KCCSBCK_FIRST], [2] on linux and W2K platform after installing 9.2.0.2. Very important patch, missing from 9.2.0.3ARU: 9.2.0.3 ARU 4110670 completed for LINUX Intel ·       

Hangcheck-timer and Oracle Cluster Manager

Download Patch 2594820 from Metalink– #rpm -ivh <hangcheck-timer RPM name>

Detaching watchdogd from the Cluster Manager (Bug 2495915)

The removal of the watchdogd ORACLE_HOME/oracm/admin/cmcfg.ora

– WatchdogTimerMargin – WatchdogSafetyMargin

KernelModuleName=hangcheck-timer CMDiskFile from optional to mandatory

– CM quorum partition of cluster participation.

Hangcheck-timer and Oracle Cluster Manager

remove or comment out from the /etc/rc.local file:

/sbin/insmod softdog nowayout=0 soft_noboot=1 soft_margin=60

ADD to rc.local, execute as root to load

/sbin/insmod hangcheck-timer.o hangcheck_tick=30 hangcheck_margin=180

Hangcheck-timer and Oracle Cluster Manager

inclusion of the hangcheck-timer kernel module Parameter Service Value----------------- -----------------

---------------hangcheck_tick hangcheck-timer 30 seconds

hangcheck_margin hangcheck-timer 180 seconds

KernelModuleName oracm hangcheck-timer

MissCount oracm > hangcheck_tick

hangcheck_margin

(> 210 seconds)

Hangcheck-timer and Oracle Cluster Manager

cmcfg.ora example- HeartBeat=15000- ClusterName=Oracle Cluster Manager, version 9i- KernelModuleName=hangcheck-timer- PollInterval=1000- MissCount=215- PrivateNodeNames=int-node1 int-node2- PublicNodeNames=node1 node2- ServicePort=9998- CmDiskFile=/ocfsdisk1/quorum/quorumfile- HostName=int-node1

Hangcheck-timer and Oracle Cluster Manager

Parameters for ocmargs.ora- oracm- norestart 1800

Linux Monitoring and Configuration Tools

- Overall tools sar, vmstat- CPU /proc/cpuinfo, mpstat, top- Memory /proc/meminfo, /proc/slabinfo, free- Disk I/O iostat- Network /proc/net/dev, netstat, mii-tool - Kernel Version and Rel. cat /proc/version - Types of I/O Cards lspci –vv- Kernel Modules Loaded lsmod, cat /proc/modules - List all PCI devices (HW) lspci –v- Startup changes /etc/sysctl.conf, /etc/rc.local - Kernel messages /var/log/messages, /var/log/dmesg - OS error codes /usr/src/linux/include/asm/errno.h- OS calls /usr/sbin/strace-p

Post Installation

Increasing Address Space Default 1.7 GB of address space for its SGA. Shutdown all instances of Oracle cd $ORACLE_HOME/lib cp -a libserver9.a libserver9.a.org

– to make a backup copy cd $ORACLE_HOME/rdbms/lib genksms -s 0x15000000 >ksms.s

– lower SGA base to 0x15000000 make -f ins_rdbms.mk ksms.o

– compile in new SGA base address make -f ins_rdbms.mk ioracle (relink)

Post Installation

Increasing Address Space Cont.

sysctl –w kernel.shmmax=3000000000

Lower process base – Find out the pid of the process (shell) from where oracle will

be started using ps (Oracle - echo $$)

– changing /proc/$pid/mapped_base to 0x10000000 and restarting oracle

Metalink Note: 200266.1

Variable SGA

Reserved for kernel

DB Buffers(SGA)

Default

Code, etc.

0xFFFFFFFF

0xC0000000

0x500000000x40000000

0x00000000

Variable SGA

Reserved for kernel

DB Buffers(SGA)

After Relink

Code, etc.

0xFFFFFFFF

0xC0000000

0x150000000x100000000x00000000

mapped_base(/proc/<pid>/mapped_base)

sga_base(relink Oracle)

Lowering of mapped base

Post Installation

Post Installation

Larger Buffer Cache does buffer cache increase with larger SGA

Create an in-memory file system on the /dev/shm mount -t shm shmfs -o size=8g /dev/shm

To enable the extended buffer cache feature, set the init.ora paramter

USE_INDIRECT_DATA_BUFFERS = true

Don’t Use dynamic cache parameters DB_CACHE_SIZE DB_#K_CACHE_SIZE

Limitations apply to the extended buffer cache feature on Linux:You cannot change the size of the buffer cache while the instance is running.

You cannot create or use tablespaces with non-standard block sizes.

Post Installation

Adjust send / receive buffer size to 256K Tuning the default and maximum window sizes:- /proc/sys/net/core/rmem_default - default receive window- /proc/sys/net/core/rmem_max - maximum receive

window- /proc/sys/net/core/wmem_default - default send window - /proc/sys/net/core/wmem_max - maximum send window - sysctl -w net.core.rmem_max=262144 - sysctl -w net.core.wmem_max=262144 - sysctl -w net.core.rmem_default=262144- sysctl -w net.core.wmem_default=262144

Post Installation

To enable asynchronous I/O must re-link Oracle to use skgaioi.o

– cd to $ORACLE_HOME/rdbms/lib– make -f ins_rdbms.mk async_on – make -f ins_rdbms.mk ioracle

– set 'disk_asynch_io=true' (default value is true) – set 'filesystemio_options=asynch‘ (RAW Only)