This chapter covers the FailSafe Express Setup Manager used with the Origin200 FailSafe network server (OFNS) system. The topics necessary for use of the express screens and troubleshooting errors are covered in the following sections:
The following information presumes that you have a console terminal connected to tty_1 on one of the two Origin200 host systems. When you bring up the system:
Login as the superuser (root).
Change the directory to /var/ha/ofns.
You may wish to confirm this by entering the pwd command, which lists your working directory.
Enter the following at the root prompt:
# ha_express_config -setup |
You should see the express setup manager screen appear.
This screen provides requirements for the system and other helpful information to get you started. An example of the screen is shown in the next section, “Express Setup Manager Screen (#1).”
A highly available IP alias is an IP address to which clients can communicate even when software and hardware upgrades or crashes occur on an OFNS system. A hostname IP address stays with a particular host system (Origin200), while a highly available IP address name can move from host to host within a cluster.
When you enter the express setup manager screen, you see information similar to this:
EXPRESS SETUP MANAGER FOR FAILSAFE 1.2
Express Setup Manager for FailSafe allows you to set up a FailSafe cluster for Highly Available Web and NFS in dual-active or active-standby configurations. For other configurations, you must use the FailSafe GUI Configurator fsconf(1M).
It is STRONGLY recommended you use `-setup' option when configuring your cluster for the first time.
To properly configure the cluster, contact your network administrator to get the following information:
Hostnames for the two machines.
Appropriate netmask for the network to which these machines will be connected.
Two highly available IP alias names
Actual numeric addresses (in `.' format) for hostnames and highly available IP aliases (Only required if addresses cannot be resolved from /etc/hosts or another name service).
Defaults are provided for most of the input values, but they may not be appropriate for your environment. For example, ip-addresses will be specific to your network.
Do you want to continue [Y]?
If you have all the information required by the setup manager screen, press Enter and go on to the next screen. Screen two tests the hardware configuration of the Origin200 FailSafe network server.
The second screen provides a list of hardware tests that the express setup manager runs and displays on the console monitor. The following example shows a normal system with no problems. OFNS is an acronym for Origin200 FailSafe network server. The term hans is an acronym for highly available network server.
HARDWARE CONFIGURATION TEST FOR OFNS
---------------------------------------
Checking for volumes ....PASSED
Testing Private Network Interface ....PASSED
Testing Private Network Interface ....PASSED
Testing Serial Connection to hans2 ....PASSED
Testing Serial Connection from hans2 ....PASSED
Hardware Diagnostics Succeeded
Go on to Network Configuration [Y]?
|
If the OFNS system fails any part of the hardware configuration test, an error message appears. The problem must be resolved before you can continue with the configuration setup using screen three (the network configurator screen). See “Troubleshooting Tips and Error Messages” for a list of error messages and possible solutions.
If you believe you may have a loose or improper cable connection, see Chapter 4, “Cabling and Configuration” for information on proper connections.
The network configurator screen prompts you for the machine name for your first server. If you want to assign a machine name other than the default hans1 name, you should enter it. The example below shows the use of ofns1 as an assigned machine name.
NETWORK CONFIGURATOR FOR OFNS
Enter the hostname for this server [hans1]? ofns1
You have entered : ofns1 : Is This Correct ? [Y]
Enter the hostname for the other server [hans2]? ofns2
You have entered : ofns2 : Is This Correct ? [Y]
Enter the domain name []? sgi.com
You have entered : sgi.com : Is This Correct ? [Y]
Enter netmask [0xffffff00]?
You have entered : 0xffffff00 : Is This Correct ? [Y]
Changing hostname of hans1 to ofns1 .... Done
Changing hostname of hans2 to ofns2 .... Done
..................................
It is necessary to reboot hans2 for these changes to be in effect
Reboot hans2 [Y]?
Rebooting hans2 ....
You have changed hostname of hans2 to ofns2
Applying these changes now ....
Waiting for hans2 to reboot ....priv-hans2: Connection timed out
rsh: connection failed
Testing Public Network Interface ....PASSED
Network Configuration Succeeded
Go on to FailSafe Configuration [Y]?
If you are satisfied with the information on hostnames, you are ready for the fourth screen, the FailSafe express configurator selection. Note that the messages
.priv-hans2: Connection timed out rsh: connection failed |
are benign messages and do not indicate an error.
This screen confirms the specific options you want for your OFNS system and asks you to choose the primary server for the Web or NFS applications.
EXPRESS CONFIGURATOR FOR FAILSAFE
A Highly Available IP alias is a floating IP address to which
clients can communicate while software and hardware upgrades
and/or crashes occur on a OFNS server.
This value should be a name string and *NOT* in `.' format
For example, it can be “HA-Web” but not like 192.0.2.15
Enter Highly Available IP Alias for ofns1 []? ha-serv1
Enter Highly Available IP Alias for ofns2 []? ha-serv2
Do you want to configure ofns1 as Web server [Y]?
Do you want configure NFS on ofns1 [Y]?
Do you want to configure ofns2 as Web server [Y]?
Do you want configure NFS on ofns2 [Y]?
You have the opportunity to change these values
Press Enter to continue ....
If you press enter, you move to a screen that shows a summary of all choices made on this screen.
Configuration Selected for OFNS
Name of node1 : ofns1
Name of node2 : ofns2
IP Alias for ofns1 : ha-serv1
IP Alias for ofns2 : ha-serv2
ofns1 is the Primary Server for the following applications :
Web
NFS
|
Backup server is ofns2
ofns2 is the Primary Server for the following applications :
Web
NFS
|
Backup server is ofns1
Is this configuration Correct [N]? y
.........................
FailSafe Configuration completed successfully
If you answer no to the correct configuration question above, your inputs to the screen are erased and you are returned to the express configurator selection screen (#4).
You can run the FailSafe Express interface screens anytime you need to reconfigure some functional aspect of the Origin200 OFNS system. You must be at the location /var/ha/ofns, and you must have superuser (root) access to both host systems.
Enter the command
# ha_express_config |
You see a screen that looks similar to the following:
EXPRESS SETUP MANAGER FOR FAILSAFE 1.2 Main Menu 1. Configure Network 2. Verify Hardware 3. Create Configuration File 4. Exit Enter Your Choice [1]? |
This screen allows you to directly access and modify information without stepping through each screen process in a linear fashion. This method of configuration is recommended only after you have completed the initial setup of the OFNS rack system.
This section offers some general tips and troubleshooting information along with error messages you may encounter using the configuration screens.
The Express Setup Manager screens used with the OFNS rack system do not configure mail on the machines in the cluster. Therefore, the mail notification feature of FailSafe will not be available.
Users wishing to have their Web sites remain highly available must be careful to put the Web documents in the appropriate filesystem. Consider the following example:
Configuration Selected for OFNS
Name of node1 : ofns1
Name of node2 : ofns2
IP Alias for ofns1 : ha-serv1
IP Alias for ofns2 : ha-serv2
ha-serv1 is the Primary Server for the following applications :
|
Web
NFS
|
Backup server is ha-serv2
ha-serv2 is the Primary Server for the following applications :
|
Web
NFS
|
Backup server is ha-serv1
|
The highly available network filesystem for ha-serv1 is /fs1. The clients of this filesystem must access it as ha-serv1:/fs1. The primary document root for the Web server on ha-serv1 is located at /fs1/www.
Similarly, the highly available network file system for ha-serv2 is /fs2. The clients of this filesystem must access it as ha-serv2:/fs2. The primary document root for the Web server is /fs2/www.
If the cluster is configured so that one node is not actively serving any application, then filesystems /fs1 and /fs2 will both be configured as highly available NFS.
Note that the node from which ha_xpress_config is run will always serve /fs1, while the other node will serve /fs2.
If you believe there is a problem with a public Ethernet connection, you can check the LED activity on the PCI Ethernet board.
There are two LEDs on the RJ45 connector:
the packet detect LED (yellow)
the transmit LED (green)
Figure 2-1 shows the locations of these LEDs.
There are four LEDs on the I/O plate that indicate the status of the board and the network. Table 2-1 summarizes the LEDs and their meanings.
Table 2-1. Fast Ethernet PCI Board and Network Status LEDs
LED Number | Color | Meaning |
|---|---|---|
1 | Yellow | When lit, the network is in Fast (100Mbps) mode |
2 | Green | When lit, indicates that the board is receiving power |
3 | Yellow | When lit, indicates that the network is in Full Duplex mode |
4 | Green | When lit, indicates that Link is established between the board and other network equipment |
Figure 2-2 shows the locations of the LEDs.
If one of the Origin200 system disks fails in the OFNS rack system, you must replace it using the procedures outlined in this section. If the disk has been backed up regularly, you may restore its content from backup tapes. If the disk has not been backed up, use the procedures in this section to restore OFNS functionality.
Substituting a standard Origin200 system disk does not restore OFNS functionality.
A system disk may need to be replaced due to either mechanical failure or file corruption. Note that the procedures listed restore only the FailSafe configurations done by ha_xpress_config.
Obtain the latest WebForce system disk and install it in the Origin200. Use the instructions in the Origin200 and Origin200 GIGAchannel Owner's Guide (007-3708-xxx) if you need help.
Connect a serial console terminal to the Origin200 you are restoring and power it on.
Install the following subsystems:
nfs subsystem – from the NFS CD supplied with the OFNS package
IRIS FailSafe 1.2 – from the FailSafe 1.2 CD
IRIS FailSafe NFS 1.2 recovery kit – from the IRIS FailSafe NFS 1.2 CD
IRIS FailSafe WEB 1.2 recovery kit – from the IRIS FailSafe Web 1.2 CD
IRIS FailSafe 1.2 OFNS install utilities – from the FailSafe 1.2 CD
the most recent NFS rollup patch – contact your service provider
the most recent IRIS FailSafe rollup patch – contact your service provider
![]() | Note: See the IRIX Admin: Software Installation and Licensing manual (007-1364-xxx) for help with command line software installation. For help with the graphical Software Manager tool, see the Personal System Administration Guide (007-1366-xxx). |
Run the restore_config command using the following steps:
# cd /var/ha/ofns
# restore_config <good_node>
Where <good_node> is the name of the other properly functioning Origin200 in the OFNS rack system.
After completing the restoration steps the OFNS system is ready to serve highly available applications as before. If you wish to change any OFNS configuration, you must have the console terminal connected to the Origin200 that was originally used to configure the OFNS system (called hans1 or ofns1 in the examples). You cannot use the secondary Origin200 (called hans2 or ofns2 in the examples) as the express configurator.
This section lists some error messages you might encounter while working with the express configuration screens.
If you receive the error message
Error : Configuration DB file ./sections does not exist
this indicates the accidental deletion of the ./sections file. This is a database file used by ha_xpress_config to create the FailSafe configuration file. This file must be present in /var/ha/ofns.
If this file has been accidentally deleted, please reinstall the OFNS software from the IRIS FailSafe 1.2 CD.
I don't know how to proceed .... Mfg config has been changed before running this script. Set node names to “hans1” and “hans2” and private network addressess to priv-hans1 and priv-hans2 correspondingly Or try running from the other node. |
This error will appear in the Network configurator screen, if either one of the following is true:
This is the first time ha_xpress_config is being run. But the hostnames were changed at the site before running ha_xpress_config.
This is not the first time this script is run, but now the script is not being run from the machine that was first used to configure the OFNS cluster.
If number 1 is true then you must change hostnames and network addresses as instructed in chapter 3 of the release notes for OFNS.
![]() | Note: The following error message refers to option [2]; the correct option is [1]. |
Hostname is other than hans1 or hans2 Select option [2] from Main Menu to configure hostname and networks. |
This message can appear if the user selects option 3 from the main menu when the command ha_express_config is run for the first time and the menu order is not followed.
The OFNS rack system is pre-configured with certain default values. As a consistency check, the command expects this configuration to be present when ha_express_config is run for the first time. This error is reported when the machine names have been changed before running the script. That is, the default hostnames are other than hans1 and hans2.
Go to the main menu and select item 3 to configure hostname and networks information.
Additional information regarding resetting names and network addresses is available in Chapter 3 of the release notes for the Origin FailSafe network server subsystem (ha_ofns).
WARNING: Your FailSafe configuration file (ha.conf) has been modified by other means since you last ran the Express Setup Manager. If you continue to use the Express Setup Manager, changes you made using other methods will be lost. Do you want to continue ? [N] |
This message appears if the IRIS FailSafe configuration file has been changed by means other than ha_xpress_config. If you continue, these changes will be lost.
IP alias for <node2> cannot be the same as <ipAlias1> |
The two nodes must have different IP aliases. This is a FailSafe requirement.
Check the log file cfgverify.out for errors. |
This message is displayed if one or more ip-aliases are not present in the /etc/hosts file. These entries may not be present in the /etc/hosts file because the user did not want the hosts file to be edited by ha_xpress_config. This error also occurs if any configuration was changed by any means other than ha_xpress_config.