Contents

Title and Copyright Information

About This Manual
Audience
Organization
Related Documents
Reader's Comments
Conventions

1    Introduction to High-Performance and High-Availability Systems
1.1    Terminology and Concepts
1.1.1    System Configuration
1.1.2    System Performance
1.1.3    Disk Performance
1.1.4    High Availability
1.2    Understanding High Availability
1.2.1    Eliminating Points of Failure
1.2.2    Increasing System Availability
1.2.3    Increasing Data Availability
1.2.4    Achieving High Availability and High Performance
1.3    Understanding High Performance
1.4    Planning Your Configuration
1.4.1    Understanding Your Workload
1.4.2    Determining Performance and Availability Goals
1.4.3    Choosing an Appropriate Configuration
1.5    Primary Configuration and Tuning Recommendations
1.6    Steps to Configure and Tune Systems

2    Diagnosing Performance Problems
2.1    Checking System Performance
2.2    Obtaining Performance Information
2.2.1    Configuring Event Logging
2.2.2    Setting up System Accounting and Disk Quotas
2.2.3    Choosing How to Monitor System Events
2.2.3.1    Using DECevent
2.2.3.2    Using Performance Manager
2.2.3.3    Using Performance Visualizer
2.2.4    Gathering Performance Statistics
2.3    Performance Monitoring Tools Overview
2.4    Gathering CPU and Virtual Memory Information
2.4.1    Using ps to Display CPU and Memory Usage
2.4.2    Using vmstat to Display Virtual Memory and CPU Statistics
2.4.3    Using uptime to Display the Load Average
2.4.4    Using swapon to Display Swap Space Usage
2.4.5    Checking CPU Usage With kdbx cpustat
2.4.6    Checking Lock Usage With kdbx lockstats
2.4.7    Checking Virtual Memory With dbx vm_perfsum
2.5    Gathering General Disk Information
2.5.1    Using iostat to Display Disk Usage
2.5.2    Checking the namei Cache With dbx nchstats
2.5.3    Checking the UBC With dbx vm_perfsum
2.5.4    Monitoring CAM Data Structures With dbx
2.6    Gathering UFS Information
2.6.1    Using dumpfs to Display UFS Information
2.6.2    Checking UFS Clustering With dbx ufs_clusterstats
2.6.3    Checking the Metadata Buffer Cache With dbx bio_stats
2.7    Gathering AdvFS Information
2.7.1    Using advfsstat to Display AdvFS Performance Information
2.7.2    Using advscan to Identify Disks in an AdvFS File Domain
2.7.3    Using showfdmn to Display AdvFS File Domain Information
2.7.4    Using showfile to Display AdvFS File Information
2.7.5    Using showfsets to Display AdvFS Filesets in a File Domain
2.8    Gathering LSM Information
2.8.1    Using volprint to Display LSM Configuration Information
2.8.2    Using volstat to Display LSM Performance Information
2.8.3    Using voltrace to Display LSM I/O Operation Information
2.8.4    Using volwatch to Monitor LSM Failures
2.8.5    Using dxlsm to Display LSM Configuration Information
2.9    Gathering Network Information
2.9.1    Using netstat to Display Network Information
2.9.2    Using nfsstat to Display Network and NFS Information
2.9.3    Checking Socket Listen Queue Statistics With sysconfig
2.9.4    Using ps to Display Idle Thread Information
2.10    Gathering Profiling and Debugging Information
2.11    Modifying the Kernel
2.11.1    Using dbx to Display and Modify Run-Time Kernel Variables
2.11.2    Using the Kernel Tuner to Display and Modify Attributes
2.11.3    Using the sysconfig Command to Display and Modify Run-Time Attributes
2.11.4    Using the sysconfigdb Command to Modify Attributes
2.11.5    Modifying Parameters in the System Configuration File

3    Optimizing Applications and CPU Performance
3.1    Configuring CPU Resources
3.2    Identifying CPU Bottlenecks
3.3    Optimizing CPU Resources
3.4    Identifying Application Bottlenecks
3.5    Improving Application Performance
3.6    Interprocess Communications Facilities

4    Configuring and Tuning Memory
4.1    Understanding Memory Management
4.2    Understanding Memory Hardware
4.3    Understanding Virtual Memory
4.3.1    Allocating Virtual Address Space to Processes
4.3.2    Translating Virtual Addresses to Physical Addresses
4.3.3    Page Faulting
4.3.4    Managing and Tracking Pages
4.3.5    Prewriting Modified Pages
4.3.6    Using Attributes to Control Paging and Swapping
4.3.7    Using Attributes to Control UBC Memory Allocation
4.3.8    Paging Operation
4.3.9    Swapping Operation
4.3.10    Using Swap Buffers
4.4    Understanding the Unified Buffer Cache
4.5    Understanding the Metadata Buffer Cache
4.6    Configuring Memory and Swap Space
4.6.1    Determining Your Physical Memory Requirements
4.6.2    Configuring Swap Space
4.6.3    Choosing a Swap Space Allocation Mode
4.7    Tuning Virtual Memory
4.7.1    Reducing the Number of Processes Running Simultaneously
4.7.2    Reducing the Static Size of the Kernel
4.7.3    Increasing the Available Address Space
4.7.4    Increasing the Available System Resources
4.7.5    Increasing the Number of Memory-Mapped Files
4.7.6    Increasing the Number of Pages With Individual Protections
4.7.7    Increasing the Size of a System V Message and Queue
4.7.8    Increasing the Size of a System V Shared Memory Region
4.7.9    Increasing the Minimum Size of a System V Shared Memory Segment
4.7.10    Reducing Application Memory Requirements
4.7.11    Reducing the Memory Available to the UBC
4.7.12    Changing the Rate of Swapping
4.7.13    Controlling Dirty Page Prewriting
4.7.14    Modifying the Size of the Page-In and Page-Out Clusters
4.7.15    Modifying the Swap I/O Queue Depth for Pageins and Swapouts
4.7.16    Modifying the Swap I/O Queue Depth for Pageouts
4.7.17    Modifying the UBC Write Device Queue Depth
4.7.18    Controlling Large File Caching
4.7.19    Increasing the Paging Threshold
4.7.20    Enabling Aggressive Task Swapping
4.7.21    Decreasing the Size of the Metadata Buffer Cache
4.7.22    Decreasing the Size of the namei Cache
4.7.23    Decreasing the Size of the AdvFS Buffer Cache
4.7.24    Reserving Physical Memory for Shared Memory
4.7.24.1    Tuning the Kernel to Use Granularity Hints
4.7.24.2    Modifying Applications to Use Granularity Hints
4.8    Tuning the UBC
4.8.1    Increasing the Maximum Size of the UBC
4.8.2    Decreasing the Amount of Borrowed Memory
4.8.3    Increasing the Minimum Size of the UBC
4.8.4    Using mmap in Your Applications
4.9    Tuning the Metadata Buffer Cache
4.9.1    Increasing the Size of the Metadata Buffer Cache
4.9.2    Increasing the Size of the Hash Chain Table

5    Configuring and Tuning Storage Subsystems
5.1    Understanding Storage Subsystems
5.2    Choosing How to Manage Disks and Files
5.2.1    Understanding RAID Levels and Products
5.2.1.1    Hardware RAID Subsystem Features
5.2.1.2    LSM Features
5.2.2    Understanding AdvFS
5.3    General Disk Storage Guidelines
5.3.1    High-Performance Hardware Guidelines
5.3.1.1    Using Fast Disks
5.3.1.2    Using Disks with Small Platters
5.3.1.3    Using Disks with Wide Data Paths
5.3.1.4    Using Solid-State Disks
5.3.1.5    Using High-Performance Host Bus Adapters
5.3.1.6    Using DMA Host Bus Adapters
5.3.1.7    Using Prestoserve
5.3.1.8    Using Write-Back Caches
5.3.2    Distributing the Disk I/O Load Guidelines
5.3.2.1    Distributing Swap Space Across Disks and Buses
5.3.2.2    Distributing I/O Across Disks and Buses
5.3.2.3    Distributing File Systems Across Disks
5.3.2.4    Placing Data at the Beginning of ZBR Disks
5.3.3    General File System Tuning Guidelines
5.3.3.1    Increasing the Maximum Number of Open Files
5.3.3.2    Increasing the Size of the namei Cache
5.3.3.3    Increasing the Size of the Hash Chain Table for the namei Cache
5.3.3.4    Allocating More Memory for the UBC
5.3.3.5    Using Prestoserve to Cache Only File System Metadata
5.3.3.6    Caching More Free vnodes
5.3.3.7    Increasing the Time vnodes Remain on the Free List
5.3.3.8    Delaying the Deallocation of vnodes
5.3.3.9    Accelerating the Deallocation of vnodes
5.3.3.10    Disabling vnode Deallocation
5.3.3.11    Modifying the Maximum Number of Open File Descriptors
5.3.3.12    Disabling Clearing of the DMA Scatter/Gather Map Registers
5.4    Using the Logical Storage Manager
5.4.1    Basic LSM Configuration Guidelines
5.4.1.1    Initializing LSM Disks as Sliced Disks
5.4.1.2    Increasing the Maximum Number of LSM Volumes
5.4.1.3    Sizing the rootdg Disk Group
5.4.1.4    Sizing Private Regions
5.4.1.5    Making Private Regions in a Disk Group the Same Size
5.4.1.6    Grouping Disks in Disk Groups
5.4.1.7    Choosing the Number and Size of the Database and Log Copies
5.4.1.8    Distributing the Database and Log Copies Across Buses
5.4.2    LSM Mirrored Volume Configuration Guidelines
5.4.2.1    Mirroring Volumes Across Different Buses
5.4.2.2    Choosing a Read Policy for a Mirrored Volume
5.4.2.3    Using Multiple Plexes in a Mirrored Volume
5.4.2.4    Using a Symmetrical Configuration
5.4.2.5    Using Multiple BCL Subdisks
5.4.2.6    Using a Write-Back Cache with LSM
5.4.2.7    Sizing BCL Subdisks
5.4.2.8    Placing BCL Logging Subdisks on Infrequently Used Disks
5.4.2.9    Using Solid-State Disks for BCL Subdisks
5.4.3    LSM Striped Volume Configuration Guidelines
5.4.3.1    Increasing the Number of Disks in a Striped Volume
5.4.3.2    Distributing Striped Volume Subdisks Across Different Buses
5.4.3.3    Choosing the Correct Stripe Width
5.4.4    LSM Tuning Guidelines
5.5    Hardware RAID Subsystem Configuration Guidelines
5.5.1    Distributing Storage Set Disks Across Buses
5.5.2    Using Disks with the Same Data Capacity
5.5.3    Choosing the Correct Chunk Size
5.5.4    Striping Mirrored Sets
5.5.5    Using a Write-Back Cache
5.5.6    Using Dual-Redundant Controllers
5.5.7    Using Spare Disks
5.6    Using the Advanced File System
5.6.1    AdvFS Configuration Guidelines
5.6.1.1    Using Multiple-Volume File Domains
5.6.1.2    Improving the Transaction Log Performance
5.6.1.3    Improving Bitmap Metadata Table Performance
5.6.1.4    Striping Files
5.6.1.5    Using AdvFS Quotas
5.6.2    AdvFS Tuning Guidelines
5.6.2.1    Modifying the Size of the AdvFS Buffer Cache
5.6.2.2    Defragmenting a File Domain
5.6.2.3    Increasing the Dirty Data Caching Threshold for a Volume
5.6.2.4    Decreasing the I/O Transfer Read-Ahead Size
5.6.2.5    Disabling the Flushing of Dirty mmapped Pages
5.6.2.6    Modifying the AdvFS Device Queue Limit
5.6.2.7    Consolidating I/O Transfers
5.6.2.8    Forcing Synchronous Writes
5.6.2.9    Moving the Transaction Log
5.6.2.10    Balancing a Multivolume File Domain
5.6.2.11    Migrating Files Within a File Domain
5.7    Using the UNIX File System
5.7.1    UFS Configuration Guidelines
5.7.1.1    Modifying the File System Fragment Size
5.7.1.2    Reducing the Density of inodes
5.7.1.3    Allocating Blocks Contiguously
5.7.1.4    Increasing the Number of Blocks Combined for a Read
5.7.1.5    Using a Memory File System
5.7.1.6    Using UFS Disk Quotas
5.7.2    UFS Tuning Guidelines
5.7.2.1    Defragmenting a File System
5.7.2.2    Delaying Full Write Buffer Flushing
5.7.2.3    Increasing the Number of Blocks Combined for Read Ahead
5.7.2.4    Increasing the Number of Blocks Combined for a Write
5.7.2.5    Increasing the Number of UFS or MFS Mounts
5.8    Tuning CAM

6    Tuning the Network Subsystem
6.1    Tuning Networks
6.1.1    Improving the Lookup Rate for TCP Control Blocks
6.1.2    Tuning the Socket Listen Queue Limits
6.1.3    Increasing the Maximum Number of Concurrent Nonreserved Dynamically Allocated Ports
6.1.4    Enabling TCP keepalive Functionality
6.1.5    Improving the Lookup Rate for IP Addresses
6.1.6    Decreasing the Partial TCP Connection Timeout Limit
6.1.7    Decreasing the TCP Connection Context Timeout Limit
6.1.8    Decreasing the TCP Retransmission Rate
6.1.9    Disabling Delaying the Acknowledgment of TCP Data
6.1.10    Increasing the Maximum TCP Segment Size
6.1.11    Increasing the Transmit and Receive Buffers for a TCP Socket
6.1.12    Increasing the Transmit and Receive Buffers for a UDP Socket
6.1.13    Allocating Sufficient Memory to the UBC
6.1.14    Disabling Use of a PMTU
6.2    Tuning the Network File System
6.2.1    Using Prestoserve to Improve Server Performance
6.2.2    Using the Appropriate Number of nfsd Daemons
6.2.3    Using the Appropriate Number of nfsiod Daemons
6.2.4    Increasing the Number of Threads
6.2.5    Modifying Cache Timeout Limits
6.2.6    Decreasing Network Timeouts
6.2.7    Using NFS Protocol Version 3.0

A    Tuning Special Configurations
A.1    Tuning Internet Servers
A.2    Tuning a Low-Memory Workstation
A.2.1    Attribute Settings for Low-Memory Workstations
A.2.2    Swap Space and Memory Tuning on Low-Memory Systems
A.2.3    X Window System Considerations for Low-Memory Workstations

B    Configuration Attribute Definitions
B.1    AdvFS Subsystem Attributes
B.2    TTY Subsystem Attribute
B.3    Configuration Manager Subsystem Attribute
B.4    DLI Subsystem Attributes
B.5    Generic Kernel Subsystem Attributes
B.6    Internet Subsystem Attributes
B.7    I/O Subsystem Attributes
B.8    IPC Subsystem Attributes
B.9    LSM Subsystem Attributes
B.10    Network Subsystem Attributes
B.11    Prestoserve Subsystem Attribute
B.12    Process Subsystem Attributes
B.13    Pseudoterminal Subsystem Attribute
B.14    Real-Time Subsystem Attributes
B.15    Security Subsystem Attributes
B.16    SNMP Information Subsystem Attribute
B.17    Socket Subsystem
B.18    STREAMS Subsystem Attributes
B.19    UFS Subsystem Attributes
B.20    VFS Subsystem Attributes
B.21    Virtual Memory Subsystem Attributes
B.22    XPR Subsystem Attributes

Glossary

Figures

Tables