| Lustre 1.6 Operations Manual |     |
Lustre 1.6 Operations Manual
820-3681-10
|
Contents |
1.3 Files in the Lustre File System
1.3.1 Lustre File System and Striping
1.6 Lustre Failover and Rolling Upgrades
1.7 Additional Lustre Features
2. Understanding Lustre Networking
3.1 Preparing to Install Lustre
3.1.2 Supported Configurations
3.2 Using a Pre-Packaged Lustre Release
3.2.1 Choosing a Pre-Packaged Kernel
3.3 Environmental Requirements
3.3.4 Choosing a Proper Kernel I/O Scheduler
3.3.5 Changing the I/O Scheduler
3.4.1 Determining the MDS’s Memory
4.2 Quick Configuration of Lustre
4.2.2 More Complex Configurations
4.2.3 Other Configuration Tasks
4.3.1 Building Your Own Kernel
4.4 Building a Lustre Source Tarball
4.4.1 Lustre Source Tarball from Lustre Source RPM
4.4.2 Lustre Source Tarball from CVS
5. Configuring the Lustre Network
5.1 Designing Your Lustre Network
5.1.1 Identify All Lustre Networks
5.1.2 Identify Nodes to Route Between Networks
5.1.3 Identify Network Interfaces to Include/Exclude from LNET
5.1.4 Determine Cluster-wide Module Configuration
5.1.5 Determine Appropriate Mount Parameters for Clients
5.2 Configuring Your Lustre Network
5.2.2 Module Parameters - Routing
5.3 Starting and Stopping LNET
6. Configuring Lustre - Examples
6.1.1 Lustre with Combined MGS/MDT
6.1.2 Lustre with Separate MGS and MDT
7. More Complicated Configurations
7.3 Load Balancing with Infiniband
7.4 Multi-Rail Configurations with LNET
8.1.1 The Power Management Software
8.1.4 Connection Handling During Failover
8.1.5 Roles of Nodes in a Failover
8.4 Configuring MDS and OSTs for Failover
8.4.1 Starting/Stopping a Resource
8.4.2 Active/Active Failover Configuration
8.4.3 Hardware Requirements for Failover
8.5 Setting Up Failover with Heartbeat V1
8.7 Setting Up Failover with Heartbeat V2
8.7.2 Configuring the Hardware
8.8 Considerations with Failover Software and Solutions
9.1.2 Creating Quota Files and Quota Administration
9.1.5 Known Issues with Quotas
10.1 Considerations for Backend Storage
10.1.2 Selecting Storage for the MDS and OSS
10.1.3 Understanding Double Failures with Software and Hardware RAID5
10.1.4 Performance Considerations
10.2 Insights into Disk Performance Measurement
10.3 Creating an External Journal
11.2 Lustre Setup with Kerberos
11.2.1 Configuring Kerberos for Lustre
11.2.2 Types of Lustre-Kerberos Flavors
12.3 Using Lustre with Multiple NICs versus Bonding NICs
12.4 Bonding Module Parameters
12.6 Configuring Lustre with Bonding
13.2 Upgrading Lustre from 1.4.12 to 1.6.4
13.2.2 Supported Upgrade Paths
13.2.4 Upgrading a Single File system
13.2.5 Upgrading Multiple File Systems with a Shared MGS
13.3 Upgrading Lustre from 1.6.3 to 1.6.4
13.4 Downgrading Lustre from 1.6.4 to 1.4.12
13.4.2 Downgrading a File System
14.1 Installing the Lustre SNMP Module
14.2 Building the Lustre SNMP Module
14.3 Using the Lustre SNMP Module
15.1.1 Client File System-level Backups
15.1.2 Performing Device-level Backups
15.1.3 Performing File-level Backups
15.2 Restoring from a File-level Backup
16.2 Running POSIX Tests Against Lustre
16.3 Isolating and Debugging Failures
18.2.2 MDS Failure (and Failover)
Part III Lustre Tuning, Monitoring and Troubleshooting
19.1 Lustre I/O Kit Description and Prerequisites
19.1.2 Prerequisites to Using an I/O Kit
19.4.1 Introduction to LNET Self-Test
19.4.2 LNET Self-Test Concepts
19.4.3 LNET Self-Test Commands
20.1.2 Lustre Timeouts/ Debugging
20.1.3 Adaptive Timeouts in Lustre
20.1.5 Free Space Distribution
20.2.1 Client I/O RPC Stream Tunables
20.2.2 Watching the Client RPC Stream
20.2.3 Client Read-Write Offset Survey
20.2.4 Client Read-Write Extents Survey
20.2.5 Watching the OST Block I/O Stream
20.2.6 Mechanics of Lustre Readahead
20.3.1 RPC Information for Other OBD Devices
21.3 Options to Format MDT and OST Filesystems
21.3.3 Overriding Default Formatting Options
21.5.1 Setting Readahead and MF
21.5.3 Setting Write-Back Cache
21.6 Large-Scale Tuning for Cray XT and Equivalents
22. Lustre Troubleshooting Tips
22.1 Lustre Error Messages and Logs
22.2.2 Write Performance Better Than Read Performance
22.2.3 OST Object is Missing or Damaged
22.2.5 Identifying a Missing OST
22.2.10 Reclaiming Reserved Disk Space
22.2.11 Considerations in Connecting a SAN with Lustre
22.2.12 Handling/Debugging "Bind: Address already in use" Error
22.2.13 Replacing An Existing OST or MDS
22.2.14 Handling/Debugging Error "- 28"
22.2.15 Triggering Watchdog for PID NNN
22.2.16 Handling Timeouts on Initial Lustre Setup
22.2.17 Handling/Debugging "LustreError: xxx went back in time"
22.2.18 Lustre Error: "Slow Start_Page_Write"
22.2.19 Drawbacks in Doing Multi-client O_APPEND Writes
22.2.20 Slowdown Occurs During Lustre Startup
22.2.21 Log Message ‘Out of Memory’ on OST
22.2.22 Number of OSTs Needed for Sustained Throughput
23.1.1 Format of Lustre Debug Messages
23.2 Tools for Lustre Debugging
23.2.1 Debug Daemon Option to lctl
23.2.2 Controlling the Kernel Debug Log
23.2.5 Printing to /var/log/messages
23.2.8 Adding Debugging to the Lustre Source Code
23.3 Using Strace for Troubleshooting
23.4.1 Determine the Lustre UUID of an OST
24.1 Querying Filesystem Space
25.1.2 Disadvantages of Striping
25.2 Displaying Files and Directories with lfs getstripe
25.3 lfs setstripe - Setting File Layouts
25.3.1 Changing Striping for a Subdirectory
25.3.2 Using a Specific Striping Pattern/File Layout for a Single File
25.4.3 Adjusting the Weighting Between Free Space and Location
25.5.1 Making Filesystem Objects Immutable
25.6.1 End-to-End Client Checksums
27.1 Expanding the Filesystem by Adding OSTs
27.2 A Simple Data Migration Script
27.3 Adding Multiple SCSI LUNs on Single HBA
27.4 Failures While Running a Client and an OST on the Same Machine
27.5 Improving Lustre Metadata Performance While Using Large Directories
29. Lustre Programming Interfaces (man2)
30. Setting Lustre Properties (man3)
31. Configuration Files and Module Parameters (man5)
31.2.2 SOCKLND Kernel TCP/IP LND
31.2.8 Portals LND (Catamount)
32. System Configuration Utilities (man8)
32.5 New Utilities in Lustre 1.6
32.5.3 Utilities to Manage Large Clusters
32.5.4 Application Profiling Utilities
32.5.5 More /proc Statistics for Application Profiling
32.5.6 Testing / Debugging Utilities
33.4 Maximum Number of OSTs and MDTs
33.5 Maximum Number of Clients
33.6 Maximum Size of a Filesystem
33.8 Maximum Number of Files or Subdirectories in a Single Directory
33.10 Maximum Length of a Filename and Pathname
33.11 Maximum Number of Open Files for Lustre Filesystems
33.12 OSS RAM Size for a Single OST
| Lustre 1.6 Operations Manual | 820-3681-10 | |
Copyright © 2008, Sun Microsystems, Inc. All Rights Reserved.