HP Serviceguard Toolkit for NFS manual Editing the File Lock Migration Script nfs.flm

Page 42

Installing and Configuring Serviceguard NFS

Configuring a Serviceguard NFS Package

Editing the File Lock Migration Script (nfs.flm)

The File Lock Migration script, nfs.flm, handles the majority of the work involved in maintaining file lock integrity that follows an HA/NFS failover. The nfs.flm script includes the following configurable parameters:

NFS_FLM_HOLDING_DIR - Name of a unique directory created in one of the shared volumes associated with this package. This directory holds copies of the /var/statmon/sm files for this package. You must create this directory in one of the shared volumes associated with this package so that it can migrate with the package (from the primary server to the adoptive server).

You must dedicate this directory for holding SM entries only. In addition, you must keep it empty. This directory should not have other files or subdirectories when starting the cluster. All files in this directory are deleted after a failover.

An example for this parameter is as follows:

NFS_FLM_HOLDING_DIR="/pkg1a/sm"

PROPAGATE_INTERVAL - Number of seconds between the attempts of the script to copy files from the /var/statmon/sm directory into the holding directory, specified by NFS_FLM_HOLDING_DIR. The default value of this parameter is five seconds.

An example for this parameter is as follows:

PROPAGATE_INTERVAL=5

NOTE

If you enable the File Lock Migration feature, an NFS client (or

 

group of clients) may hit a corner case of requesting a file lock on the

 

HA/NFS server and not receiving a crash recovery notification

 

message when the HA/NFS package migrates to an adoptive node.

 

This occurs only when the NFS client sends its initial lock request to

 

the HA/NFS server and then the HA/NFS package moves to an

 

adoptive node before the FLM script copies the /var/statmon/sm

 

entry for this client to the package holding directory.

 

The probability of hitting this corner-case problem is not very high,

 

because the SM file copy interval is very short (by default, five

 

seconds). The chances of an NFS client (or group of NFS clients)

 

sending its initial lock request (it must be the initial request, since

42

Chapter 2

Image 42
Contents Serviceguard NFS Toolkit A.11.11.04 Administrator’s Guide EditionLegal Notices Contents Index Figures Figures Overview of Serviceguard NFS Overview of Serviceguard NFS Limitations of Serviceguard NFS Overview of the NFS File Lock Migration Feature Overview of the NFS File Lock Migration Feature Supported Configurations Simple Failover to an Idle NFS Server Pkg1 disksFailover from One Active NFS Server to Another Pkg2 DisksHost Configured as Adoptive Node for Multiple PackagesCascading Failover with Three Adoptive Nodes Server-to-Server Cross Mounting NFS NFSSupported Configurations Starting the NFS Services How the Control and Monitor Scripts WorkStarting File Lock Migration Halting the NFS Services Monitoring the NFS Services On the Client Side Installing and Configuring Installing and Configuring Serviceguard NFS Installing Serviceguard NFS ReadmeCmmakepkg -p /opt/cmcluster/nfs/nfs.conf Monitoring NFS/TCP Services with Serviceguard NFS ToolkitMonitoring NFS/TCP Services with Serviceguard NFS Toolkit Before Creating a Serviceguard NFS Package NUMNFSD=10 Before Creating a Serviceguard NFS Package Mount -o nointr relocatableip/usr/src /usr/src Configuring a Serviceguard NFS Package Copying the Template Files Editing the Control Script nfs.cntl IP0=15.13.114.243 SUBNET0=15.13.112.0 Lower Etc/hosts file Configuring a Serviceguard NFS Package Editing the NFS Control Script hanfs.sh Netswitchingenabled Editing the File Lock Migration Script nfs.flm Configuring a Serviceguard NFS Package Editing the NFS Monitor Script nfs.mon Configuring a Serviceguard NFS Package Editing the Package Configuration File nfs.conf Configuring a Serviceguard NFS Package Configuring Server-to-Server Cross-Mounts Optional SNFS0=nfs1/hanfs/nfsu011CNFS0=/nfs/nfsu011 Configuring a Serviceguard NFS Package Run the cluster using the following command cmruncl -v-f Configuring a Serviceguard NFS Package Chapter Sample Configurations Sample Configurations Example One Three-Server Mutual Takeover Basil SageThree-Server Mutual Takeover after One Server Fails Cluster Configuration File for Three-Server Mutual Takeover Package Configuration File for pkg01 NFS Control Scripts for pkg01 Nfs.cntl Control ScriptPackage Configuration File for pkg02 NFS Control Scripts for pkg02 Package Configuration File for pkg03 NFS Control Scripts for pkg03 Example Two One Adoptive Node for Two Packages with File Lock MigrationSage Basil Pkg02 VOLUMEGROUP/dev/nfsu01 VOLUMEGROUP/dev/nfsu02 Package Configuration File for pkg01 NFS Control Scripts for pkg01 Hanfs.sh Control Script NFS File Lock Migration and Monitor Scripts for pkg01 Nfs.flm ScriptPackage Configuration File for pkg02 NFS Control Scripts for pkg02 NFSFLMSCRIPT=$0%/*/nfs2.flm NFS File Lock Migration and Monitor Scripts for pkg02 Example Three Three-Server Cascading FailoverCascading Failover with Three Servers after One Server Fails VOLUMEGROUP/dev/nfsu01 VOLUMEGROUP/dev/nfsu02 Package Configuration File for pkg01 NFS Control Scripts for pkg01 NODENAMEthyme NODENAMEsage NODENAMEbasil IP0=15.13.114.244 SUBNET0=15.13.112.0 Example Four Two Servers with NFS Cross-Mounts Thyme Basil Cluster Configuration File for Two-Server NFS Cross-Mount NODENAMEthyme NODENAMEbasil NFS Control Scripts for pkg01 SNFS0=nfs1/hanfs/nfsu011 CNFS0=/nfs/nfsu011 NODENAMEbasil NODENAMEthyme NFS Control Scripts for pkg02 SNFS0=nfs2/hanfs/nfsu021 CNFS0=/nfs/nfsu021 Index Index Default values
Related manuals
Manual 93 pages 19.37 Kb Manual 28 pages 17.67 Kb