Cluster Products

HA Agents for Netscape Servers Installation and Administration Guide

Version 1.0
February 1998
Insight Solutions Inc.
20380 Town center Ln. Suite #215
Cupertino, CA 95014
USA.

Copyright (C) 1998, Insight Solutions, Incorporated All rights reserved.
The computer program listings, specifications and documentation herein are the property of Insight Solutions Incorporated or a third party supplier and shall not be reproduced, copied, disclosed in whole or in part for any reason without the prior express written permission of Insight Solutions Incorporated.

Overview

HA Agent for Netscape Server is designed to work with Sun Cluster software and Netscape servers to
provide high availability data services on a Sun Enterprise Cluster.

High Availability Data Services

The Sun Cluster software provides a framework for supporting high availability (HA) data services. A high availability data service (for example, the Netscape Enterprise Server) is associated with a logical host. The service's executable and data are stored on a multihost disk group attached to a cluster. At any one time, one of the nodes in the cluster runs the service and responds to requests addressed to the logical host. This node is called the current master of the logical host. If the current master fails for any reason, another node can take over as master, minimizing disruption of service.

High availability data services are described in more detail in Chapter 12 of the Sun Cluster 2.1 System Administration Guide.

HA Agent for Netscape Servers

The HA Agent software works with the Sun Cluster high availability (SC HA) framework to prevent interruption of service. The HA Agent software monitors the Netscape server to ensure that it is running properly. If the server stops responding to requests, the HA Agent software restart the server or notifies the SC HA framework to  move it to a different node.

For example, the HA Agent software monitors the Enterprise Server by requesting a specified web page (the test page) at regular intervals. If the server doesn't return the web page in a reasonable amount of time, the HA Agent software logs it as an error. After one or more consecutive errors, the HA Agent software may attempt to restart the server, or migrate it to another node. The administrator can specify the test page, how long to wait for a response, how many faults to accept before attempting to restart or move the server, and so forth.

Installing HA Agent for Netscape Servers

Installation Overview

To install HA Agent, you must have a Sun Enterprise Cluster running the Sun Cluster software. Setting up a high availability data service requires the following steps:

  • Creating a Logical Host for the HA Server
  • Installing the Netscape Server
  • Installing the HA Agent Software
  • Registering the HA Service

You may already have completed some of these steps. For example, if you already have the Netscape Server running on a logical host on your cluster, you should be able to skip to step 3, installing the HA Agent software.

Note: The Sun Cluster HA Internet Pro package provided HA agents for Netscape servers version 2.1 and earlier. If you have this package or any older versions of the Netscape servers installed, you must uninstall them using pkgrm(1M) before installing the new Netscape servers and the Insight Solutions HA Agent packages. The Insight Solutions HA Agent packages require Netscape servers version 3.5 or above in most cases; check the Insight web site at www.insightsol.com for up-to-date compatibility information.

Creating a Logical Host

Use the scconf(1M) utility to set up a logical host. This procedure is covered in section 12.3 of the Sun Cluster 2.1 System Administration Guide.

Installing the Netscape Server Software

To install the server software, use the node that is currently mastering the logical host, and install onto the disk group associated with the logical host. To ensure that the server can access all of its data no matter which physical host is mastering it, you must install all of the server's data on the multihost disk group(s) associated with the logical host. So if you're installing the Netscape Enterprise Server, you must install its HTML files, CGI scripts, and any data that the CGI scripts access on the logical hosts's disk group(s). Otherwise, the server will not be able to migrate properly between cluster nodes.

Suppose you are installing the Netscape Enterprise Server, and you have the following setup:

  • Logical host name: loghost1
  • Mount point for shared disk associated with loghost1: /ns-home
  • Physical host name: node1
  • Domain: mydomain.com

When you are configuring the server, Netscape will ask for the "Server name." The default value will be the physical host name, node1.mydomain.com. Change this to the logical host name, loghost1.mydomain.com. Instead of the default location for the server root (/usr/netscape/suitespot), choose a location on the shared disk (for instance, /ns-home/suitespot).

During the course of the installation, you will create a server instance. Make a note of the directory you specify in the "Server Location" field on the Server Configuration form. You will need to supply this directory when you register the HA service.

Installing the HA Agent Software

You must install the HA Agent software on all the nodes in the cluster. You can install the software from a CD, or download the installation files from the Insight Solutions web site.

Before installing the software, you should have a set of license keys. For each package you're installing, you'll need a unique license key for each node in the cluster. If you don't have a set of license keys, contact Insight Solutions at sales@insightsol.com. You'll need to supply the host ID for each node in the cluster to receive the appropriate license keys.


Downloading the Installation Files

To download the HA Agent installation files, follow these steps:

  • Download the software
  • Download the desired software package from the Insight Solutions web site, www.insightsol.com. The software comes as a compressed tar archive.
  • Extract the installation files
  • Extract the files using uncompress and tar.
  
# uncompress INS_HA_HTTP_1_0.tar.Z 
# tar xf INS_HA_HTTP_1_0.tar
./INSTALL
./InsHAbase
./InsHAhttp
./install.ini

Now proceed with the instructions below for installing and licensing the software.

Preparing to Install from a CD

If you're installing from a CD, follow these steps:

  • Mount the CD
  • Put the CD in the drive. The system should mount the CD automatically.  
  • Change to the installation directory

The CD contains separate installation directories for each HA Agent package. They are:

 
cdrom0/entprise HA Agent for Netscape Enterprise Server
cdrom0/directry HA Agent for Netscape Directory Server
cdrom0/message HA Agent for Netscape Messaging Server
cdrom0/proxy HA Agent for Netscape Proxy Server
cdrom0/admin HA Agent for Netscape Administation Server

So if you want to install the HA Agent package for the Netscape Enterprise Server, and the CD-ROM is mounted at /cdrom, do the following:

  
# cd /cdrom/cdrom0/entprise

Now proceed with the instructions below for installing and licensing the software.

Installing the HA Agent Software

To start installing the software, run the install script, INSTALL. The install script will display a list of packages to install.

  
# ./INSTALL
###################################################################
#                                                                                                                                                  #
#                                                                                                                                                  #
#  Welcome to Insight's world of Highly Available Data Services.                                                        #
#                                                                                                                                                  #
#  Insight Solutions, Inc.                                                                                                                 #
#                                                                                                                                                  #
#  Copyright (C) 1998, Insight Solutions, Incorporated                                                                      #
#  Protected as an unpublished work.                                                                                               #
#  All rights reserved.                                                                                                                     #
#                                                                                                                                                  #
#                                                                                                                                                  #
###################################################################

Insight Solutions Highly Available Data Services Packages

  1  Insight_Administration_Base_Package

  2  Netscape_Enterprise_server_3.5.1

  3  Netscape_Directory_Server_3.0

  4  Netscape_Messaging_Server_3.5

Choose the HA Agent to be installed (default: all) [?,??,q]:

Select the packages you wish to install or press Enter to install all of the packages. (Note that the Insight Administration Base Package is required by the other packages, and will be installed automatically if you select any of the agent packages.) The script now calls pkgadd to install each requested package.
When you install the base package, the script will prompt you for a port number:

  
Base package from Insight Solutions.
(sparc) 1.0

Insight Solutions Inc.

HABase comm server is to be configured on an unused port.
it will listen on this port for requests.
/etc/services and /etc/inetd.conf files will be modified by
this script to make HABase related entries.
Please enter the same port number on all the cluster nodes.

Enter port number [5994]:

This is the port number that the HA Agent software will use to communicate. Note that you must use the same port number on all nodes in the cluster, and this port number must not be in use by any other service.

Next, the script prompts you for a directory to install the software in:

  
Default base directory -  BASEDIR is /opt.
HABase Base package will be installed in {BASEDIR}/HAagents

 Enter BASEDIR [/opt]: 

You can specify an alternate directory to install the software in, or use the default. To avoid confusion, you should use the same directory on all nodes.

  
Using </opt> as the package base directory.

## Processing package information.
## Processing system information.
   3 package pathnames are already properly installed.
## Verifying package dependencies.
## Verifying disk space requirements.
## Checking for conflicts with packages already installed.
## Checking for setuid/setgid programs.
This package contains scripts which will be executed with super-user
permission during the process of installing this package.

 Do you want to continue with the installation of <InsHAbase> [y,n,?]

Enter "y" and press Enter to continue.  

  
Installing Base package from Insight Solutions. as <InsHAbase>

 ## Installing part 1 of 1.

/opt/HAagents/InsHAbase/HA_comms
/opt/HAagents/InsHAbase/HA_send
/opt/HAagents/InsHAbase/haadmin
 [ verifying class <none> ]
## Executing postinstall script.
Modifying /etc/inet/inetd.conf, to add InsHAbase server
Modifying /etc/services
Restarting inetd (234)
Installation of <InsHAbase> was successful.

The procedure is the same for the other packages, except that you only have to specify the port number when installing the Insight Administration Base Package.  

After you've installed all of the packages you want to install, follow the instructions below for licensing the software.  

Licensing the Software After youíve installed the software, you must license it. You should have a list of license keys, one for each node in the cluster. After installing the software on a node, call the install script with the -l flag:
 

 
# ./INSTALL -l

The install script will present you with a menu, allowing you to install license keys for individual packages:

  
#######################################################################
# Welcome to Insight's world of Highly Available Data Services.       #
#  Insight Solutions, Inc.                                            #
#######################################################################

Insight Solutions Highly Available Data Services

Packages

  1  Insight_Administration_Base_Package
  2  Netscape_Enterprise_server_3.5.1
  3  Netscape_Directory_Server_3.0
  4  Netscape_Messaging_Server_3.5

Enter the Data service for which license key is

available. [?,??,q]: 

Select a package to license and press Enter. The install script will prompt you for a license key:

 
Enter the license string, for
Netscape_Enterprise_server_3.5.1 [?]

After you've entered the license key, the install script will ask if you have any more licenses to install:

 
Do you wish to continue to add licenses ...
[y,n,?,q] 

If you have more licenses to install, type "y" and press Enter to install them. Continue until you've installed all the licenses you have for this node.

Registering the HA Service

Once you've installed the HA Agent software, you can use the haadmin command to register the service with the SC HA framework. If you have more than one server instance, you must register each instance separately. You can use haadmin in interactive mode or non-interactive mode. To start haadmin in interactive mode, enter the following command:  

  
# /opt/HAagents/InsHABase/haadmin -r service_name -c service_class \
     -l logical_host -y

Note: If you specified an installation directory other than /opt, you must modify the pathname accordingly.

The service name is a name you choose to represent the service. The identifier can include upper and lower-case letters, numerals, and the underscore characters, and it must start with a letter. So "WWW1" and "http- public" are valid service names, but "1bigWebServer" is not. The logical host name is the name of the logical host associated with the server.

The service class represents the kind of server you're registering (for example, Enterprise Server, Directory Server, or Messaging Server). The following table lists the service classes for the various Netscape servers.  

 
Server Service class
Netscape Enterprise Server nshttp
Netscape Directory Server nsldap
Netscape Messaging Server nsmail
Netscape Proxy Server nsproxy
Netscape Administration Server nsadmin

So, to register an Enterprise Server instance called WWW1, you would use the following command:  

  
# /opt/HAagents/InsHABase/haadmin -r WWW1 -c nshttp -l loghost1 -y

In interactive mode, the haadmin program will prompt you to supply the values of various parameters. For every service, you must supply the APPLICATION_BASE parameter. APPLICATION_BASE specifies the directory holding the start and stop scripts for the service. Other parameters may be required for certain servers. For example, when registering an Enterprise Server instance, you must supply the HTTP_URL parameter. HTTP_URL identifies the test page, which the HA Agent software requests from the server to determine whether the server is running properly. Other parameters are described below in the "Parameters" section.

For example, if you have created an instance of the Enterprise Server named WWW1, and start and stop scripts for the server are located in /ns-home/https-WWW1, then you could use the following parameters:

APPLICATION_BASE=/ns-home/https-WWW1
HTTP_URL=http://loghost1.mydomain.com

If another server instance, WWW2, is running on port number 8080 and its start and stop scripts are in /ns-home/https-WWW2, then you would use the following parameters:

APPLICATION_BASE=/ns-home/https-WWW2
HTTP_URL=http://loghost1.mydomain.com:8080

To register an Enterprise Server instance WWW2, you would use the folowing command:  

  
# /opt/HAagents/InsHABase/haadmin -r WWW2 -c nshttp -l loghost1 -y

When you finish specifying parameters, haadmin will start the service. If you don't want haadmin to start the service automatically, leave off the -y flag.

Registering in Non-Interactive Mode

You can also register the service in non- interactive mode. To do this, you specify parameters using the -p flag to haadmin. You must specify all of the required parameters. For example:
 

  
# /opt/Haagents/InsHABase/haadmin -r WWW2 -c nshttp -l loghost1 \
     -p APPLICATION_BASE=/ns-home/https-WWW2 \
     -p HTTP_URL=http://loghost1.mydomain.com:8080 -y

 Default value will be selected for parameter which are not specified in command.

Service Dependencies

If the service depends on another HA service (for example, ldap1), you can specify this with the -d flag. For example:
 

 
# /opt/Haagents/InsHABase/haadmin -r WWW1 -c nshttp -l loghost1 -d ldap1

In this example WWW1 service is dependeny on ldap1 service. If ldap1 service is associated with logical host other than loghost1, then it is necessary to associate ldap1 service with logical host loghost1, before registering WWW1 service using scconf command.
 

 
# /opt/SUNWcluster/bin/scconf mycluster -s ldap1 loghost1

Using the HA Agent Software

Starting Service

To start the service, call haadmin -y:  

 
# /opt/Haagents/InsHABase/haadmin -y service_name

If you have more than one server instance configured, you can start them all with the following command:
 

 
# /opt/Haagents/InsHABase/haadmin -y -c service_class

Note: Netscape server will be started by HA agent software when service is turned on using haadmin command. If Netscape server has been manuallly started, then it must be stopped before starting the service.

Listing Parameters

To list the parameters associated with a service, call haadmin with the -l flag.  

 
# /opt/Haagents/InsHABase/haadmin -l service_name

Modifying Parameters

To modify one or more parameters, use haadmin -m.  

 
# /opt/Haagents/InsHABase/haadmin -m service_name -p PROBE_INTERVAL=60

Viewing Help

To view help for haadmin or help for error messages, use haadmin -h.  

 
# /opt/Haagents/InsHABase/haadmin -h service_name help_id

For example

 
# /opt/Haagents/InsHABase/haadmin -h WWW1 2463

Stopping Service

To stop the service, call haadmin -n:  

 
# /opt/Haagents/InsHABase/haadmin -n service_name

If you have more than one server instance running, you can stop them all with the following command:  

 
# /opt/Haagents/InsHABase/haadmin -n -c service_class

Unregistering Service

Should you need to unregister a service, call haadmin with the -u flag.  

 
# /opt/Haagents/InsHABase/haadmin -u service_name

Parameters

Common Parameters

The following parameters are used by all services.

 
Parameter Values Default Description
APPLICATION_BASE Valid directory (none) Directory containing start and stop scripts for the server instance. Required parameter.
FM_ENABLE {Y, N} Y Specifies whether fault monitor is enabled. 
PROBE_INTERVAL >0 30 Interval between probes, in seconds. 
PROBE_TIMEOUT >0 30 Number of seconds to wait for a response from server. 
ERROR_THRESHOLD >=1 2 Number of consecutive errors to accept before taking action.
ACTION_ON_ERROR STOP
SWITCH
RESTART
NONE 		RESTART Action to be taken after ERROR_THRESHOLD is reached. 

STOP-stop the server and fault monitor. Server and fault monitor can be restarted manually or during the next cluster reconfiguration.

SWITCH-Attempt to give away logical host, if another node is capable of taking it over. Otherwise, continue as is.

RESTART-Attempt to restart the server on the same node. 

NONE-Take no action (same effect as having the fault monitor disabled).

MAX_RESTART >=0 2 Maximum number of restart attempts to make in a specified interval. If this number is exceeded, the fault monitor will take the action specified by ACTION_ON_MAX_RESTART.
MAX_RESTART_INTERVAL >0 300 If the service needs to be restarted more than MAX_RESTART times in the course of MAX_RESTART_INTERVAL seconds, take the action specified by ACTION_ON_MAX_RESTART. (For example, if all parameters have their default values, the fault monitor will switch the server to another node if it has to be restarted more than twice in 300 seconds.) 
ACTION_ON_MAX_RESTART STOP
RESTART
NONE 		SWITCH Action to take after MAX_RESTART restart attempts.

Enterprise Server Parameters

The following parameter applies only to the Enterprise Server.

 
Parameter Values Default Description
HTTP_URL Valid URL (none) URL for test page. Required parameter. 

Messaging Server Parameters

The following parameters apply only to the Messaging Server.
 

 
Parameter Values Default Description
MTAUID Valid user ID (none) User ID for the MTA user. Required parameter. 
MTAGID Valid group ID (none) Group ID for the MTA user. Required parameter. 
FINGER_DISABLE yes, no yes If yes, the finger daemon is disabled. 
VAR_MAIL_SETUID yes, no yes Should /var/mail have the setgid flag on. 

Note: The MTA user is the UNIX user created to run the messaging server. Check the variable MailUserName in /etc/netscape.mail.conf for the exact name of the MTA user. The user id (UID) and group id (GID) of the MTA user can determined from the /etc/passwd file.

Important: The MTA user must exist on each node of the cluster and its UID and GID must be the same on all nodes. Use the useradd and usermod commands to make this so, if it is not the case already.

Directory Server Parameters

The following parameters apply only to the Directory Server.

 
Parameter Values Default Description
LDAP_PORT Valid port number 389 Port number of the directory server. 
BASE_DN Valid directory node monitor Directory node that can be queried for, without any access control. 

Proxy Server Parameters

The following parameters apply only to the Proxy Server.

 
Parameter Values Default Description
PROXY_PORT Valid port number 8080 Port number of the proxy server. 
SOCKS_PORT Valid port number 1080 Port number of the socket server. 
SOCKS_SERVER_ENABLE {Y, N} Y Specifies whether socket server is enabled. 

Error and Informational Messages

Error Messages

Config file filename does not exist or is unreadable.

The service configuration file can't be read. Check whether the service configuration file exists and has proper permissions.

Could not kill the process associated with pmfadmtag pmfadm_tag

pmfadm(1M) tried to kill the service processes, but failed.

Could not send HUP signal to service. pmfadm -k returned return_value.

Soft restart of the service failed. Soft restart is achieved by sending a HUP signal to the service.

See the pmfadm(1M) manual page for a description of pmfadm return values.

Could not stop the process associated with pmfadmtag pmfadm_tag

Service stop was requested, but the attempt to disable the process monitoring facility failed.

Directory directory not existent or not accessible.

The specified directory doesn't exist. This may be the result of a setup problem. See /etc/netscape.mail.conf and check if the directory specified there is a valid directory.

Error creating link to netscape sendmail

Check directory permissions for /usr/lib and the file permissions for mailserver_install_directory/bin/mail/server/bin/sendmail.

Error linking netscape libraries to shared disk

Check directory permissions for the destination directory and the file permissions for source file(s).

/etc/netscape.mail.conf does not exist.

Either /etc/netscape.mail.conf doesn't exist or does not have read permissions. Please correct this. In case the file is missing, restore from the backup file (/etc/netscape.mail.conf.ha-save) or from another node in the cluster.

Failed to start FM_start_command. pmfadm returned return_value.

Fault monitor start was requested, but pmfadm was unable to start the fault monitor.

Failed to start service_start_command. pmfadm returned return_value.

Service start was requested, but pmfadm was unable to start the service.

FATAL ERROR. CAN NOT PROCEED AHEAD. EXITING...

A fatal error has occurred. A fatal error can occur if:

  1. A mandatory parameter has not been supplied.
  2. An incorrect or invalid parameter value has been specified for a mandatory parameter.

FM_start_command not found

The fault monitor script cannot be found. Reinstall the HA Agent package.

GID of MTA user gid incorrect. Expected GID expected gid.

The Netscape Messaging Server needs an MTA user to exist and the user must have the same GID on each node in the cluster. In this case the expected GID for the MTA user is expected_gid. Please run usermod and make the correction.

Illegal values in netscape.mail.conf

Check parameter settings in /etc/netscape.mail.conf.

Incorrect value for variable_name in config file.

The variable variable_name in the HAconfig.ini file for mail server has an incorrect value.

Internal error: method_name called with incorrect parameters

This is a possible defect. Please contact Insight product support.

Invalid value for parameter: value

The value specified for the parameter is invalid. Refer to the "Parameters" section for a list of valid parameter values. Use haadmin to specify an appropriate value for the parameter.

MTA user error

Either the MTA user does not exist, or its UID/GID is not consistent. This user must be created on each node in the cluster and must have have the same UID/GID on each node. Run useradd or usermod to correct this.

MTA user user_name does not exist on this host.

The Netscape Messaging Server needs a special user to be created. Run useradd to create $MailUserName. IMPORTANT: The UID and GID associated with user_name must be the same for all nodes.

net_diagnose_comm detected errors. initiating switchover of service

The node is not attached to the network, or the network path is broken. Check the network accessibility of the node. Try pinging to or from some other node.

No Match package package_name

The admin file system contains a directory corresponding to the service instance, but the instance does not belong to the package of the start method.

Not recognized line

The service could not interpret the specified line in the configuration file. Refer to the "Parameters" section for a list of valid parameter values. Use haadmin to specify an appropriate value for the parameter.

Package package_name not installed.

The service package was not installed properly. This usually means that the files were copied instead of being installed with pkgadd. Use pkgadd to reinstall the service package.

Service_path exists but is a directory.

The symbolic link representing the service instance does exist. Instead, a directory is present with that name.

Service_path exists but is a regular file.

The symbolic link representing the service instance does exist. Instead, a regular file is present with that name.

Service_path exists but is not linked to package_path

The symbolic link representing the service instance does not point to the package directory, but points to some other file or directory.

Service_start_command not found

The service start program was not found. Check the values of the parameters APPLICATION_BASE and START_PROGRAM. (START_PROGRAM is not a user-configurable parameter, but it appears in the service configuration file, HAconfig.ini.) Make sure that $APPLICATION_BASE/$START_PROGRAM is the start program for the service, and that it has the proper execute permissions.

Service tag pmfadm_tag is not registered. pmfadm -q returned return_value.

Soft restart of the service was called, but the service was not running under pmfadm.

See the pmfadm(1M) manual page for a description of pmfadm return values.

symlink failed. Source file src does not exist

Make sure src exists and has the correct permissions.

TAG: pmfadm_tag was not registered.

Service stop was requested, but the service was not running under pmfadm.

UID of MTA user uid incorrect. Expected UID expected_uid.

The Netscape Messaging Server needs an MTA user to exist and the user must have the same UID on each node in the cluster. In this case the expected UID for the MTA user is expected_uid. Please run usermod and make the correction.

Unable to access netscape.mail.conf

Unable to read netscape.mail.conf. Make sure that it exists and has the proper file permissions. If missing, restore from the backup file netscape.mail.conf.ha-save.

Unable to connect to protocol server.

The daemon for the specified protocol (one of SMTP, IMAP or POP) is not responding. If this situation persists, the fault monitoring agent will attempt to restart or migrate the mailserver.

Unable to remove link link_name

Check the permissions of link_name and its parent directory.

Value not supplied for mandatory parameter VERSION in config file filename

The version information in the configuration file is missing. Use haadmin to reregister the service (unregister, then register the service again).

Value not supplied for mandatory parameter parameter in config file filename

The named parameter is not set in the configuration file. Use haadmin to specify an appropriate value for the parameter.

Variable variable_name not set.

Possible mailserver setup issue. See /etc/netscape.mail.conf and check if the variable is set.

License check failed

The message indicates that the license check has failed for the HA-Agent package.

Start and Stop programs for socket sever missing. Cannot register.

Please check the APPLICATION_BASE directory and make sure that start and stop programs for the socket server are present there.

Informational Messages

Service seems to have been started manually. First stop the service and then bring it under control of HA.

The service start was requested. But the service was already found running, also it was found that it was started extrnally (out side the HA frame work).

Removing fault monitor pid file : FAULT_MONITOR_PID_FILE

Fault monitor pid is stored in a file. It is used to kill the fault monitor. When the fault monitor is killed, the pid file is deleted and the message is given.

Using pid (pid) to kill fault monitor

Fault monitor stop method is called. It is stoping the fault monitor by kill the fault monitor process. The message describes, the fault monitor process id.

Completed first successful probing with service. process_info

When the fault monitor completes, the first successful probing of service it gives out this message. This message optionally may contain some information about the process(es) monitored by the fault monitor.

Fault monitor detected an error regarding secondary service/process(es)

While probing the service, when a fault monitor detects an error, with respect to service (if there are more than one services running as primary / secondry services, then with respect to SECONDRY service ), it gives out this message. It means that there are some problems with the service, and service / some components of service are not running successfully.

Saved pid (pid) of the fault monitor in LOG_DIR/ha_fm.pid

When the fault monitor is started, it saves its pid information in a file and gives out this message.

MAX number of restarts reached, hence disabling fault monitoring for sock server

The socket server was restarted MAX_RESTART number of times, but still it could not be started successfully.

Removing fault monitor pid file : FAULT_MONITOR_PID_FILE

Fault monitor process pid information is kept in a file. This file is to be deleted when the fault monitor stop method is called.

Attempting to start fault monitor for service [service_name] using : FM_start_command

Fault monitor start has been requested. The fm_start method is making an attempt to start the fault monitor.

Attempting to start service [service_name] using : service_start_command

Service start has been requested. The start_net method is making an attempt to start the service using the start command provided for the service.

Creating link: my_base svc_base

A symbolic link pointing to the package directory for the service instance has been created by the start method.

/etc/netscape.mail.conf and path/netscape.mail.conf differ.

/etc/netscape.mail.conf seems to have been modified locally. Please note the difference between /etc/netscape.mail.conf path/netscape.mail.conf. and make the changes, if required, to the netscape.mail.conf on each node and to the shared disk copy.

Fault monitor action : action

The fault monitor has detected ERROR_THRESHOLD number of errors and taken the described action. If you would like to change the fault monitor's behavior, you can change the ERROR_THRESHOLD and ACTION_ON_ERROR parameters using haadmin.

Fault monitor already started: process_IDs

Fault monitor start was requested, but the fault monitor was already running.

Fault monitor detected an error.

The fault monitor detected an error, but no action is being taken at this point.

Fault monitor is disabled, returning from method_name.

The fault monitor is not running.

Fault monitor is enabled.

The fault monitor is enabled and running.

FM_start_command successfully started.

The fault monitor has been started successfully.

Link exists: my_base svc_base

The symbolic link from the service instance to the package of the service already exists.

local_node does not master logical_host ; returning from method_name.

The node issuing the message is not master of the logical host.

local_node masters logical_host.

The node issuing the message is master of the logical host.

Monitoring service: service_name

The fault monitor has been started for the named service.

Restarting the service by first stopping it followed by starting it.

The service is being restarted. To restart a service, the HA framework stops the service and starts it again.

Rotated service_log.

The indicated log file has been rotated.

Sending HUP signal to service.

A soft restart of service has been requested. The fault monitor performs a soft restart by sending a HUP signal to the service.

Service already started: process_ID

Service start was requested, but the service was already running.

Service_start_command successfully started.

The service startup script has been called successfully.

Successfully stopped the service using pmfadm_stop_command

Service stop was requested and service has been shut down gracefully, using the stop command associated with the service.

Successfully stopped the service with pmfadmtag pmfadm_tag

Service stop was requested. The service has been shut down ungracefully by having pmfadm kill it.

Switching the service to another node.

Switchover of service has been requested. The node is giving up mastery of the logical host to another cluster node.

Total [error_threshold] errors detected.

The fault monitor has detected ERROR_THRESHOLD errors, but ACTION_ON_ERROR is set to NONE, so no action was taken.

Unable to connect to protocol_name server.

The daemon for the specified protocol (SMTP, IMAP or POP) is not responding. If this situation persists, the fault monitoring agent would will attempt to failover the mailserver.

 
Copyrigh