sccl:usage

Usage
- Start
- Commands
- Examples

Usage

Start

Translation

Das Starten des Clusters. Für die Clusterkommunikation wird auf allen Clusterknoten ein unix2webd-Prozess gestartet, zu beachten ist ggf eine Freischaltung des verwendeten Ports in der jeweiligen Firewall der Clusterknoten.

SLES

Das SCript-CLuster ist so konfiguriert, dass es beim nächsten Reboot des Rechners automatisch gestartet wird. Die Konfiguration des SCript-CLusters (siehe voriges Kapitel) muss vor dem Start des Clusters erfolgen.

Manuelles Starten über:

# /etc/init.d/sccl_cluster reload
Config File: /etc/unix2web/sccl.conf

Config File: /etc/unix2web/sccl.conf
/usr/local/bin/unix2webd -D -p 3579 -h /usr/local/sccl/bin  -r "SCCL Administration" -E -t 300 -U -sp "/usr/local/sccl/bin/u2w_pwd.dat" -or " 10.1.3.113 10.1.3.113"

Dies muss auf jedem Clusterknoten erfolgen. Die Konfigurationsdateien sollten dafür auf allen Clusterknoten gleich sein.

HP-UX 11.23

Manuelles Starten erfolgt über

# /sbin/init.d/sccl_cluster reload
Config File: /etc/unix2web/sccl.conf

Config File: /etc/unix2web/sccl.conf
/usr/local/bin/unix2webd -D -p 3579 -h /usr/local/sccl/bin  -r "SCCL Administration" -E -t 300 -U -sp "/usr/local/sccl/bin/u2w_pwd.dat" -or " 10.1.111.11 10.1.111.12 10.0.11.11 10.0.11.12"

Dies muss auf jedem Rechnerknoten erfolgen. Die Konfigurationsdateien sollten für alle Rechnerknoten gleich sein.

Commands

sccl_disable

sccl_disable_node

sccl_dist_config

sccl_dist_config [-m] [-x <excludenode>] [-n <onlynode>] [files ...]

Distribute the cluster-relvant files to all other nodes. Normally, you do not specify any files explicitly.

sccl_dist_config

Mainly the files:

etc/sccl/sccl.conf
etc/sccl/adminpwd.dat
etc/sccl/userpwd.dat
etc/sccl/packages.conf
etc/sccl/resources.conf
etc/sccl/scripts/

These files will overwrite any existing files on the destination nodes:!:

Special action: define <other_node> also as master.

sccl_dist_config -m -n <other_node>

Remember: only a master node can add more nodes

sccl_enable

sccl_enable_node

sccl_get_aktnode

Find the active cluster node(s) for a cluster package.

$ sccl_get_aktnode <package>
{ [<node1> [<node2>]...] | 0 }

All names of all cluster nodes are listet, on which the <package> is active. If the <package> is not active on any cluster node, the output is:

Example:

$ sccl_get_aktnode package3
clnode1 clnode2

The package package3 is active on the cluster nodes clnode1 and clnode2.

sccl_list_cluster

sccl_list_cluster_nodes

sccl_list_nodes

sccl_list_pakete_on_node

sccl_restart

sccl_show_cluster

sccl_start

sccl_start_node

sccl_stop

sccl_stop_node

sccl_test_res

more to come

Examples

Example 1

This example script-cluster configuration describes a productive environment of:

two servers, host001 and host002
- one application server
- one database server
each server has a package
- 'tomcat' for the application server
- 'database' for the database server

This is a very simple example and does not contain any failover functionality.

The following configuration files only show the significantly changed parts or parameters.

/etc/sccl/sccl.conf

# Clustername
CLUSTER=SAMPLE01
 
# Cluster nodes
NODES="host001 host002"

The package 'tomcat' can only run on host 'host001'. The package 'database' can only run on host 'host002'.

/etc/sccl/packages.conf

# <package>  [-]  <host1> [<host2> [<host3>]...]
 
# Application
tomcat host001
 
# Database
database host002

/etc/sccl/resources.conf

# <package>  <package_name>  <resource1> [<resource2> [<resource3]...]
 
database My_Database   PRG:db_startstop   RST:tomcat:CLUSTER
 
tomcat Tomcat_Application_Server   CPKG:database:WAIT:3   PRG:tomcat_startstop

Alternatively you can use: