NEC ExpressCluster SRE Version 3.
NEC ExpressCluster SRE 3.1 Installation and Configuration Guide (FT Linux edition) Document Number EC-SRE-001-LN, Version 1.2, April 2005 Part Number 456-01701-000 Copyright © 2005 NEC Solutions America, Inc. Information in this document is subject to change without notice. No part of this document may be reproduced or transmitted in any form or by any means, electronic or mechanical, for any purpose, without the express written permission of NEC Solutions America, Inc.
Contents About this guide ________________________________________________________ 5 Using this guide ______________________________________________________ 5 Where to go for more information ______________________________________ 6 Chapter 1: Server Agent Installation _______________________________________ 7 Prerequisites ________________________________________________________ Hardware__________________________________________________________ Minimum Hard disk and Memory Capacity ________________________
Windows Installation ________________________________________________ 25 Setting Java Security________________________________________________ 26 Linux Installation ___________________________________________________ 29 Setting Java Security________________________________________________ 29 Chapter 6: Management Client Security Configuration _______________________ 31 Overview __________________________________________________________ 31 Configuration Steps _________________________________________________
About this guide NEC ExpressCluster SRE 3.1 is an integrated software solution designed for monitoring and automatically restarting processes and daemons on a single server. Using this guide This document includes step-by-step instructions to help you install NEC ExpressCluster SRE 3.1 and configure it to monitor processes and daemons on the NEC-FT Linux Operating System.
• • • Configuration (Chapter 2) Management (Chapter 3, Chapter 4, Chapter 5 and Chapter 6) Customization (Chapter 7) For installation, configuration and customization to monitor other general processes please read the following chapters in sequence: • Installation (Chapter 1) • Configuration (Chapter 2) • Management (Chapter 3, Chapter 4, Chapter 5 and Chapter 6) • Customization (Chapter 8) Where to go for more information Please refer to additional documentation under the “documentation” directory on th
Chapter 1: Server Agent Installation Please note that only Server Agent Installation (this chapter) should be performed at the factory since configuration and customization are dependent on customer site-specific environment. Prerequisites Hardware Server Model The NEC Express5800/320Lb server Minimum Hard disk and Memory Capacity Hard Disk: 77MB, Memory: 512 MB Software Operating System NEC FT Linux Version 2.1.0.
The output will be: ecsre-svr-3.
Chapter 2: Server Agent Configuration This chapter describes the one-time initial configuration procedure for NEC ExpressCluster SRE 3.1 in the final deployment environment. Computer Name Configuration If the computer name (as shown by running the “hostname” command) is changed from the factory setting (“localhost”) then you MUST perform the following steps: Step 1 Open the file /opt/nec/clusterpro/etc/clp.
Step 2 Locate the FT Linux ExpressCluster SRE License Key at the back of the server along the edge as indicated in following figure: Key: A1234567-B1234567-C1234567-D1234567 S/N: AA000000 FT Linux ExpressCluster SRE Lic.
Selection of product division 1. Product 2. Trial Select product division [ 1 or 2 ] ... 1 Step 5 Enter the number of license as 1: Enter the number of license [ 1 to 99 (default:2) ] ... 1 Step 6 Enter the serial number similar to the license as in the sample below: Enter serial number [Ex. XX000000] ... ... AA000000 Step 7 Enter the license key similar to the license as in the sample below: Enter license key [Ex. XXXXXXXX- XXXXXXXX- XXXXXXXX- XXXXXXXX] ] ...
Starting ExpressCluster SRE Step 1: With the login as root, type in the following at the command prompt: /opt/nec/clusterpro/management/sre-register This command registers the daemons of ExpressCluster SRE with “chkconfig” so that the daemons startup automatically on all subsequent system reboots. Step 2: Run the following: /opt/nec/clusterpro/management/sre-start This command starts the ExpressCluster SRE daemons.
Chapter 3: Management Client Installation Overview The Management Client is a Java applet that allows you to monitor the status of the cluster, start and stop the groups and shutdown and reboot the server. Server Prerequisites Network On the FT Server, either the 100 base or the 1000 base pair of Network Interface Cards need to be teamed to form one Virtual Interface. An IP address needs to be assigned to this Virtual Interface.
If you do not have JRE installed, please refer to Appendix C: Installing JRE for installation instructions. If you already have JRE installed, please refer to Appendix D: Modifying JRE for modifying your existing installation on Windows. The Management Configuration Client (explained in Chapter 5) requires the Support for Additional Languages for JRE, as it uses shift-JIS character encoding on the Windows platform.
Chapter 4: Basic Operations The following sections describe how to use the management client to perform basic ExpressCluster SRE monitoring and control operations. Please note that screenshots in the example below have one resource “ping” (ping 127.0.0.1) and a monitoring resource “ping monitor” which monitors “ping”. Also, the host name of the server is SRE-SRV and the group name is SRE-GRP Prerequisites 1.
This will open up the user interface similar to the one below: 16
Exiting the Management Client Step 1 Click File -> Close in Internet Explorer Browser as below: 17
Viewing Group Status Step 1 Use the left mouse button and click on the group name SRE-GRP as below: The status for the group, server and ping resource are shown Online.
Viewing Monitored Resource Status Step 1 Use the left mouse button and click on the resource name “ping” as below: The status of the “ping” resource (/bin/ping 127.0.0.1) is shown Online.
Viewing Monitor Resource Status Step 1 Use the left mouse button and click on the monitor resource name “ping monitor” as below: The status of the “ping monitor” resource is shown Online.
Stopping a Group Step 1 Right click on the group SRE-GRP and select stop as below: Step 2 Click OK on the following screen: 21
The group will be stopped and all resources are offline as below: 22
Shutting down a Server Step 1 Right click on the server SRE-SRV and select Shutdown: Step 2: Click OK on the following screen: The server will shutdown completely.
Restarting a Server Step 1 Right click on the server SRE-SRV and select Reboot: Step 2: Click OK on the following screen: The server will be restarted.
Chapter 5: Management Configuration Client Installation Overview The Management Configuration Client is a Java Applet used for configuring scripts to monitor processes on the server Prerequisites Please refer to the Server Prerequisites and Standalone Computer Prerequisites sections of Chapter 3, since prerequisites for the Management Configuration Client are the same as the Management Client and can be installed on the same machine.
Step 2 Click OK on the following screen: Step 3 Click Close on the following screen: This completes the installation of the Management Configuration Client. Setting Java Security Pre-requisites: Ensure that you exit out from all Browsers before you execute the steps below: Step 1 Open Windows Explorer and browse to the following location: C:\Documents and Settings\ Here is the name of the user logged in.
Step 2 Check to see if .java.policy file is present. If not present, please execute Step 3 and skip Step 4. If present skip Step 3, and execute Step 4. Step 3 (Execute this step if not present) Copy the .java.policy file present in C:\Program Files\nec\clptrek\etc\ to the following directory: C:\Documents and Settings\ Step 4 (Execute this step if present) Open the .java.
Step 5 Open Internet Explorer, type in C:\Program Files\nec\clptrek\clptrek.html and confirm that the Script Customization Tool is installed successfully.
Linux Installation Step 1 Insert the installation CD into the CD ROM drive Step 2 Login as “root” with the password set for the “root” user. Step 3 Type in the following to mount the CD as read only: mount –o ro /dev/cdrom /mnt Step 4 Type in the following: rpm –i /mnt/manager/ecsre-cfg-3.1-2.i386.rpm Step 5 You could verify that the rpm is installed successfully by the following command: rpm –q ecsre-cfg-3.1-2 The output will be: ecsre-cfg-3.
Step 2 (Execute if not present) Copy the .java.policy file present in /opt/nec/clptrek/etc into your home directory. Step 3 (Execute if present) Open the .java.policy file in the “vi” editor and add the following lines to the end of the file: grant codeBase "file:/opt/nec/clptrek/clptrek.jar" { permission java.security.AllPermission; }; Note: The path specified above needs to be delimited by “/” not “\”. Step 4 Open the browser, type in /opt/nec/clptrek/clptrek.
Chapter 6: Management Client Security Configuration Overview Management Client Security can be configured by restricting server access either to clients with specific IP addresses or to all clients within a network segment. Once the security is configured, only clients with the specific IP addresses or clients within the specified network addresses can use the Management Client user interface to operate the server.
Open the browser, in the address bar type in /opt/nec/clptrek/clptrek.html.
Step 2 Click File -> Open the configuration file-> Change the cluster configuration and navigate to C:\Program Files\nec\clptrek\etc\ and open clp.
Click File -> Open the configuration file-> Change the cluster configuration and navigate to /opt/nec/clptrek/etc/ and open clp.
Step 3 The following screen will be displayed: 35
Step 4 Right click on the cluster SRE-VCC and select the Property popup menu item as below: 36
Step 5 Select the WebManager tab, check the Connection is restricted by Client IP Address check box and click the Add button: 37
Step 6 In the edit box specify either of the following: 1. A specific client IP address which would be the only client that would have access to the server (e.g. 10.123.187.3) OR 2. A network addresses which would enable access to all the clients within the specified network segment (e.g.10.123.187.
Step 7 The Operation check box is selected by default as below: If selected, clients accessing the server can view the status and also perform basic operations like shutdown, reboot, stop group, start group etc.
A sample screenshot for a client with IP address 10.123.187.
If de-selected, clients accessing the server can only view the status and would not be able to perform basic operations like shutdown, reboot stop group, start group etc. A sample screenshot for a client with IP address 10.123.187.3 with the Operation checkbox un-checked is as below: Check or Uncheck the Operation check box according to your requirement, click Apply and OK once done.
Step 8 Select File -> Save the configuration file: 42
Step 9 Select clp.
Select clp.
Step 11 Note: You do not need to reboot the server at this time. Please ignore this dialog below and click OK: This completes the security configuration for management clients. Configuration Distribution Pre-requisites Please ensure that the ftp daemon is started on the FT-Linux server. Please refer to Appendix E: Starting the ftp daemon to start the wu-ftpd daemon on the server. Distribution Steps Note: Steps 1 through 3 needs to be executed just one time.
Step 4 From the client computer start the command prompt (Start->Run->cmd) and type in the following: cd C:\Program Files\nec\clptrek\etc\ From the client computer type in: cd /opt/nec/clptrek/etc/ Step 5 Connect to the server using ftp with the user name as “sreadmin” and password “sreadmin” e.g. ftp 10.123.187.1 Note: If you encounter the error: “ftp: connect : Connection refused” you missed the pre-requisite stated above.
Note: You do not need to reboot the server at this time The group will be started and the cluster will be resumed. Step 9 Execute the following command: /opt/nec/clusterpro/management/sre-mgr-restart The webalert daemon and the webmanager server will be restarted. Step 10 The security settings for the management clients will now be set and only clients permitted to access the server can view/operate the server using the Management Client user interface.
Chapter 7: Customization for Sendmail Prerequisites 1. The sendmail daemon needs to be stopped. Please verify its status using: service sendmail status The output will be similar to: sendmail (pid 8366) is running... If sendmail is running, please stop sendmail using the command: service sendmail stop 2. The sendmail daemon should not startup automatically when the Operating System boots up.
Step 4 You can verify the status of the cluster using clpstat –s The output will be similar to: ====================== CLUSTER STATUS =========================== Cluster : SRE-VCC *localhost ....... : Online Standalone Server lanhb1 : Normal LAN Heartbeat SRE-GRP .........
Chapter 8: General Customization Overview Processes to be monitored may be classified into two categories: 1. Processes that remain attached to a terminal when started (e.g. “ping”). These processes do not fork child processes. 2. Processes that do not remain attached to a terminal when started (e.g. “squid”). These processes fork other child processes. The customization steps explained are common for both the categories above, unless stated otherwise. The customization involves two major steps: 1.
Customization Steps Creating the target resource Step1 Open Internet Explorer, type in C:\Program Files\nec\clptrek\clptrek.html in the address bar.
Open the browser, in the address bar type in /opt/nec/clptrek/clptrek.html.
Step 2 Click File -> Open the configuration file-> Change the cluster configuration and navigate to C:\Program Files\nec\clptrek\etc\ and open clp.
Click File -> Open the configuration file-> Change the cluster configuration and navigate to /opt/nec/clptrek/etc/ and open clp.
Step 3 The following screen will be displayed: 55
Step 4 Right click on SRE-GRP in the tree view and click Add resource 56
Step 5 Select execute resource from the list box as below: 57
Step 6 Change the name to squid and type in a comment as appropriate for your process: 58
Step 7 Select the User Application radio button and click Edit as below: 59
Step 8 In the Start edit box, enter the absolute path of the process (and start parameters if any) you would like to start. In the Stop edit box, enter the absolute path of this process (and its stop parameters if any). Click OK.
On the FT server, you could create a script that takes in command line arguments start and stop. This script should also have the logic of starting, stopping and monitoring the process by based on its Process ID (“pid”), and should terminate if the process terminates. You could refer to this sample script (/opt/nec/clusterpro/monitors/squidexec) customized for the squid process: #!/bin/bash case "$1" in start) /etc/init.
In the Start edit box, enter the absolute path of this script with its start parameters In the Stop edit box, enter the absolute path of this script with its stop parameters Click OK.
Step 9 Click the Tuning button. This takes you to the Parameter tab. Asynchronous means ExpressCluster SRE will not wait for the completion of the start script execution. Always select Asynchronous here.
Step 10 (Optional) In the Maintenance tab, you could specify the absolute path to a log file where standard output and standard error are redirected to. If this is left blank, standard output and standard error redirect to /dev/null. Click OK once you are done.
Step 11 Click Next > on the following screen: 65
Step 12 Activity Retry Threshold indicates the number of times ExpressCluster SRE will try to re-start the process, incase the process fails to startup the first time. When this threshold is crossed, the Final Action will be executed and the server would be rebooted. Set “3” here. Set the Failover Threshold to “0”.
Step 13 Uncheck Default dependence is followed and click Complete 67
Creating the monitoring resource Step 1 Right Click on Monitors, and select Add monitor resource: 68
Step 2: Select pid monitor from the list box shown: 69
Step 3 Change the Name to squid monitor and add a suitable comment and click Next >: 70
Step 4 Interval indicates the monitoring interval in seconds. Set this to 2. Timeout indicates the timeout value within which the status of the process should be normal. If this exceeded, ExpressCluster SRE will consider it abnormal. Set this to 60. Retry Count indicates the number of times abnormality should be detected consecutively, before the status of the resource is considered abnormal. Set this to 0. Start Monitor Wait Time is the monitoring interval time in seconds.
Step 5 Select squid and click Ok.
Step 6 Click Next > 73
Step 7 Recover Object is the object you want to recover.
Step 8 Select the squid resource here to recover squid. This means that ExpressCluster SRE will try to restart squid as part of recovery. Click OK.
Step 9 Reactivation Threshold indicates the maximum number of times the recovery action will be executed. Set this to 3. Set the Failover Threshold to 0 as this is not applicable here for a single server. If Reactivation Threshold is crossed, the Final Action is executed. Select Stop Cluster Daemon And OS Reboot for the Final Action and click Complete.
Step 10 Select File -> Save the configuration file: 77
Step 11 Select clp.
Select clp.
Step 13 Note: You do not need to reboot the server at this time. Please ignore this dialog below and click OK: This completes the creation and configuration of the required resources. Configuration Distribution Pre-requisites: 1. Ensure that the process you are trying to monitor is not already running. If it is running, please stop it. Also for processes that startup when the OS reboot, please disable the automatic startup of the process at boot time using the command: chkconfig --del e.
When prompted for the new password type in “sreadmin” Step 3 Create a directory called scripts in the /home/sreadmin folder using the following commands: cd /home/sreadmin mkdir scripts Step 4 From the client computer start the command prompt (Start->Run->cmd) and type in the following: cd C:\Program Files\nec\clptrek\etc\ From the client computer type in: cd /opt/nec/clptrek/etc/ Step 5 Connect to the server using ftp with the user name as “sreadmin” and password “sreadmin” e
… … Need to shutdown system and reboot please shutdown system after push. (hit return) : file delivery to server 127.0.0.1 success. … … Note: You do not need to reboot the server at this time Step 9 You can verify the status of the cluster using clpstat –s Step 10 You can verify that the squid daemon has started using the command: service squid status The output will be similar to: squid (pid 30750 30748) is running... If squid is terminated, it is automatically restarted.
Appendix A: Changing the Computer Name Please execute Step 1 before changing the computer name: Step 1 Run the following command to suspend the cluster: /opt/nec/clusterpro/management/sre-suspend Please go ahead and change the computer name now. After changing the computer name, please execute Steps 2 through 6. Step 2 Open the file /opt/nec/clusterpro/etc/clp.conf using the “vi” editor Step 3: Look for the following line and change the computer name to the new name (say “XXX”).
Appendix B: Assigning an IP address The following steps team the 10/100 (100 Base) Ethernet adapters and assign an IP address to it. Note: 1. If you want to team the 1000 Base Ethernet adapters and assign an IP address to it replace “7” by “4” for all steps below. 2. It is mandatory for ExpressCluster SRE that this IP address is static. 3. Steps below assign the following static IP address as an example: IP: 10.123.187.1 Mask: 255.255.255.
Step 4 Use the following command to check the status of LAN card: vndctl status 7 “ha0” indicates the interface name (ha0) which is used to set dual configuration to the two LAN cards and then virtually treats them as one LAN card. Step 5 Use the following command to set the IP address: vndctl config 7 Step 6 The screen will be switched to the entry screen. (netconfig command is being executed.) The message “Would you like to set up networking?” will appear. Click “Yes” to continue.
Set the IP address, Subnet Mask, Default Gateway IP and the Primary Name Server as per your Network environment. Step 7 Use the following command to activate the dual LAN card configuration of the slot 7: vndctl up 7 Step 8 Check the status using: vndctl status Dual LAN card configuration has completed when the screen displays the status as “UP”, and the link is “LINK” (with LAN cable connected) as shown above.
Appendix C: Installing JRE Windows You can install it from the EC SRE – FT Linux CDROM from the following directory: \tools Step 1 Double click the j2re-1_4_2_07-windows-i586-p.
Step 3 Select Custom and click Next 88
Step 4 In addition to the Java 2 Runtime Environment, also select the Support for Additional Languages as shown below and click Next > Note: The Support for Additional Languages is required because the Customization Tool uses shift-JIS character encoding on the Windows Platform.
Step 5 Select Microsoft Internet Explorer and click Next > 90
Step 6 Click Finish on the following screen: This completes the installation of JRE on Windows.
Linux Step 1 Login is “root” with the set password for the “root” user Type in the following: cd / Step 2 Insert the installation CD in the CDROM drive and mount the CD as read only using: mount –o ro /dev/cdrom /mnt Step 3 Type in the following: /mnt/tools/j2re-1_4_2_07-linux-i586-rpm.bin Step 4 The script displays a binary license agreement, which you are asked to agree to before installation can proceed. Once you have agreed to the license, the install script creates the file j2re-1_4_2_07-linux-i586.
Appendix D: Modifying JRE The following section walks you through the steps to modify the existing installation of JRE on the Windows Platform. These steps install the Support for Additional Languages as the Client Configuration Tool uses shift-JIS character encoding on the Windows Platform. Step 1 Double click the j2re-1_4_2_07-windows-i586-p.
Step 3 Select Modify and click Next > 94
Step 4 In addition to the Java 2 Runtime Environment, also select the Support for Additional Languages as shown below and click Next > 95
Step 5 Click Finish on the following screen: 96
Appendix E: Starting the ftp daemon The following section walks you through the steps to start the wu-ftp daemon on the FT-Linux server Step 1 On the FT-Linux server, with the login as root, open the wu-ftpd file in the /etc/xinetd.d folder using the command: vi /etc/xinetd.d/wu-ftpd Step 2 Change the default setting from disable = yes to disable = no as below: # default: on # description: The wu-ftpd FTP server serves FTP connections.
