Performing Routine Maintenance

Introduction

This document describes DIGITAL VNswitch module maintenance procedures that you may need to perform periodically, and those that you should perform regularly.

Accessing and Exiting the Boot Config Prompt

You must access the Boot config prompt (Boot config>) to configure and manage many of the parameters discussed in this document. The prompt is accessed from the Config prompt (Config>).

Accessing the Boot Config Prompt

To access the Boot config prompt, perform the following steps:

1

At the Config prompt, enter boot.

2

Press Return. The following message is displayed, followed by the Boot config prompt:

TFTP Boot/dump configuration
Boot config>

Exiting the Boot Config Prompt

You exit the Boot config prompt to return to the Config prompt.

Example
To return to the Config prompt (Config>) from the Boot config> prompt, enter exit and then press Return.

 

Restarting the Switch

Certain switch configuration tasks require that you restart the switch for new configuration settings to take effect. The switch is also started when power is applied.

Restarting the switch causes reinitialization of the switch software, using the same executable and configuration it is currently running. The executable (also known as the boot or image file) and configuration parameters are stored in the switch's nonvolatile flash memory. Restarting the switch also clears information stored in volatile memory, including all bridge table entries not saved to NVRAM, and drops any packets in the bridge. Recovery of the lost packets is the responsibility of the sending and receiving nodes.

Note:  Restarting a switch from a remote console terminates the console's Telnet session.

For information on restarting the switch, refer to Resetting Configuration Parameters.

Note: You can also restart the switch from the logon menu.

Upgrading and Reinstalling Switch Software

An upgrade replaces the current version of software with a newer version. A reinstallation reloads the switch with a copy of the current version of the software. You may need to reinstall the software if, for example, you have a problem upgrading to a newer version and need to reinstall the earlier version until the problem is resolved.

The steps you complete to perform an upgrade or reinstallation are the same. Therefore, for the purpose of this discussion, the term installation is used to mean either an upgrade or a reinstallation.

For information on upgrading or reinstalling switch software, refer to Upgrading Device Firmware.

LED Lighting Sequence During Load and Reload

The module's LEDs provide information regarding the progress of an installation or upgrade. If the TFTP file transfer is successful, a CRC of the image received at the module is performed. Table 1 describes the LED lighting sequences and the status of the upgrade they reflect.

Table 1: LED Lighting Sequence During Load/Reload

LED Lighting Sequence

Upgrade Status

VNbus Traffic LED (upper): green

VNbus Status (lower): yellow

CRC is in progress.

VNbus Traffic LED (upper): yellow

VNbus Status (lower): green

CRC detects an error.

VNbus Traffic LED (upper): green

VNbus Status (lower): green

The image is being written to flash, following a successful CRC.

VNbus Traffic LED (upper) and VNbus Status LED (lower) both flash alternating green and yellow for about 10 seconds, and then remain solidly lit (either green or yellow) for another 10 seconds

The process of writing the image to flash is complete.

Caution:  Do not power cycle the module while the image is writing to flash. The module may become disabled.

Updating the Software Version Number

You should update the version number in configuration memory for newly upgraded software. The basic steps you perform to update the version number is shown in the following steps. However, refer to the instructions provided in the release notes sent with the software upgrade for detailed instructions.

To update the version number in configuration memory for an upgrade, perform the following steps:

Step Action
1 At the Main prompt (Main>), type Config and press Return.

1

At the Config> prompt, enter update version-of-SRAM.

2

Press Return. The version number is updated in configuration memory, and the Config> prompt is displayed.

Backing Up and Restoring the Switch

You can back up the switch's configuration settings. Switch configuration settings that are stored in NVRAM can be restored manually.

Automatic Image Recovery

The VNswitch 900 family of modules automatically tries to reload the switch's image if the currently installed image is corrupted. Although the current image is stored in NVRAM and does not need to be reloaded in the event of a power outage, the image can become corrupted due to an unusual event such as a power surge. The automatic download of a new image only occurs over the LDM port.

Location of Source Software

The source software used for automatic recovery must be stored on a network host or server. The location can be the same as that used for normal software upgrades or installations described in the Upgrading and Reinstalling Switch Software section. If the switch software image is corrupted, the switch must look to an external device to determine the location of the software it must reload. Such a device is referred to as a BootP server. In this discussion, the VNswitch module is referred to as a BootP client.

BootP Server and Clients

The BootP server contains a file that lists all the BootP clients for which the server is responsible, including the clients' IP addresses, and the locations and names of their boot files. This list of clients and boot file location information is maintained by the network administrator. If a switch's image is corrupted, the switch (BootP client) broadcasts a request to the BootP server. The request, in the form of a UDP packet, includes the client's MAC address. When the server receives the request, it looks up the client's address in its database of client information. If it locates the client's address in the database, the BootP server responds to the client, providing it with information about the location of the software it is to install. The switch then initiates a TFTP request for a download of the software from the boot server on which the software resides.

Note: The BootP server can use any BootP software available on a variety of operating systems.

DIGITAL Online Services

VNswitch software is also available through Digital's Bulletin Board System and Network Products Business Web pages. Refer to the following online services for additional information:

WWW
The DIGITAL Network Products Home Page on the World Wide Web is located at the following addresses:

North America

http://www.networks.digital.com

Europe

http://www.networks.europe.digital.com

Asia Pacific

http://www.networks.digital.com.au

Configuring Automatic Image Recovery

To configure your system for automatic recovery of the switch's executable software, you must configure the BootP server. You do not need to configure the BootP client (the VNswitch module). The image is automatically reloaded through the LDM port on the switch.

The network node you use as a BootP server can use any operating system that supports the BootP utility. You must properly configure the server to include a BootP database that contains a list of all the BootP clients (switches) for which the server is responsible, including the clients' IP addresses, and the locations and names of their boot files. Refer to the operating system documentation provided by the system vendor for configuration instructions.

Backing Up Configuration Settings

Default configuration settings and most of the settings you configure are stored in a configuration database in the VNswitch module's NVRAM. Although the configuration database does not need to be reloaded if there is a power outage, the database can become corrupted due to an unusual event such as a power surge. You can make a copy of the configuration database and back it up on a remote server or host. Should the switch's configuration database be corrupted, you can then copy the backup to the switch, minimizing the need to reconfigure custom settings you previously entered.

For information on backing up configuration settings, refer to Saving and Restoring Configuration Data.

Before You Begin

You must also configure the switch's TCP/IP Host Services to include one IP address for the module, before attempting backup or restoral. Because the configured IP address is lost if NVRAM is corrupted, the address must be reconfigured before restoral. You can also reconfigure the switch's host name (optional). Refer to the Configuring IP Host Services for information about configuring TCP/IP Host Services.

Restoring a Configuration Database

The configuration database can be restored to a VNswitch module of the same type (900EE, 900EX, and so on) from which it was backed up. The restoral process includes verification of a "magic number" that ensures the configuration database is being restored on a switch of the proper type. The switch's host name is also checked to make sure the configuration is restored on the correct switch.

For information on restoring configuration settings, refer to Saving and Restoring Configuration Data.

Caution: Do not power off or reset the switch while restoring the configuration database. NVRAM may become corrupted, requiring you to reinstall the module's configuration database.

Note: You cannot transfer a configuration database using the remote system's version of TFTP-client.

Checking Available RAM

The VNswitch module's volatile memory is used to store information such as bridge tables and various buffers. A predefined amount of volatile memory is allocated from the switch's random access memory (RAM) at startup. If more volatile memory is required, additional RAM is variably allocated to volatile memory, as needed.

Refer to Monitoring Memory for more information about monitoring switch memory.

Capturing Restart or Crash Messages and Diagnostic Data

The VNswitch module logs messages and diagnostic data whenever the module is restarted or experiences a fatal error. You can display these messages to help troubleshoot problems and can download detailed diagnostic data for further problem analysis.

Displaying and Managing Restart or Crash Error Messages

The VNswitch module generates informational messages whenever it is restarted and a problem is encountered, and whenever a fatal error such as a bug halt occurs.

The restart messages, generated by diagnostic routines during power-up, are recorded in a diagnostic log. For information on displaying diagnostic log messages, refer to Monitoring Diagnostic Log Entries.

Messages related to fatal errors (crashes) are recorded in a crash log. For information on displaying crash log messages, refer to Managing Crash Log Entries.

Both error logs are stored in NVRAM and, therefore, the information they contain remains intact after a restart or power outage.

You display the messages in these logs to determine the results of diagnostics and to detect crashes that might occur when you are not present. (The occurrence of a crash may not be obvious because the switch typically restarts automatically.) However, a detailed interpretation of the messages is possible only with the assistance of DIGITAL service representatives, who can assist troubleshooting problems you encounter. Each log retains up to 254 messages. If the maximum of 254 messages is reached, each new message overwrites (wraps) the oldest existing message. Refer to the Determining When a Crash Occurred for detailed information about how to read the crash log to determine when a crash occurs.

You can also delete messages from the logs if you want to make detection of new messages easier.

Determining When a Crash Occurs

You can find out when a crash occurs by determining when a message is logged in relation to other messages, and by inserting a date/time stamp or similar marker at specific points in the crash log.

Message Sequence
Two numbers are automatically assigned to each message, a crash log entry number and a historical message number. The crash log entry number is displayed on the line preceding a message. It is used to identify messages you want to delete. The historical message number is used to determine when a specific message is logged in relation to other messages. The historical message number is displayed on the same line as the message. Refer to Figure 1 for an example.

The maximum number of crash log entries that can be recorded is 254. If this number is exceeded, the next message logged overwrites the oldest crash message starting at crash log entry number 1. However, when the historical message number reaches 254, it continues to increment by 1 to 255, 256, and so on as shown in Figure 1, up to and including 9999. It is then reset to 1. Therefore, the crash message having the highest historical message number is the most recent message logged. Using Figure 1 as an example, crash log entry number 5 with a historical entry number of 259 is a more recent entry than crash log entry number of 6 with a historical entry number of 6.

The crash log is cleared and the crash log entry number is reset to 1, if the total message record size is exceeded and a new message is logged. Because the length of each message is different, the speed with which the record becomes full varies, depending on the total number of characters of all messages logged.

If the time has been set using either the Config>time set or Config>time host commands, then a timestamp is saved with each crash message. If not, the uptime of the switch is saved.

Figure 1: Sample Crash Log

CRASH LOG ENTRY NUMBER 5 | Message 259 | Recorded (Uptime): 0:01:48 (T2.0-022) Illegal Instruction at PC 0xA

CRASH LOG ENTRY NUMBER 6 | Message 6 | Recorded: 1/5/1993 17:27:16 (T2.0-023) User induced bughlt via DVT

Crash Log Markers
You can insert a marker in the crash log so that you can identify when messages that follow a marker occur.

For information on entering a marker in the crash log, refer to Adding Message to Crash Log.

Markers appear as a message in the crash log, and are assigned a crash log entry number and historical message number. Refer to crash log entry number 5 in Figure 1 for an example.

Deleting All Diagnostic or Crash Log Messages

You can delete (clear) all messages from either the diagnostic or crash log. Refer to Monitoring Diagnostic Log Entries or Managing Crash Log Entries, respectively.

Deleting a Single Diagnostic or Crash Log Message

For information on deleting a message from the diagnostic log, refer to Deleting a Single Diagnostic Log Message.

Refer to Managing Crash Log Entries for information on deleting a single message from the crash log.

 

Downloading Diagnostic Data for Problem Analysis

You can capture a "snapshot" of a failed module's status at the time of failure.This information can be used as a diagnostic tool by Digital services representatives, should you require assistance.

Capture the diagnostic data by configuring your switch to automatically download (dump) the information to one or more host systems when the module is reset due to a hardware or software failure. The data is also downloaded when the Reset/Dump button located on the module's front panel is pressed.

This section discusses the following topics:

Configuring and Enabling Dump Files

You can configure up to eight remote locations to which diagnostic information is downloaded. It is recommended that you configure at least one location. Configuring multiple locations helps ensure that, in the event a download to one location is not successful, the download to another location is likely to succeed.

The VNswitch module dumps files over the Load/Dump/Management (LDM) port only. The LDM port is located on the front panel of the module.

To configure the location of the host or server that is to receive a dump file, perform the following tasks:

Task

Description

1

Configure an IP address for the LDM port and enable TCP/IP host services.

2

Create a dump file on the TFTP server.

3

Specify the name of the dump file, and the location of the server to which the file is to be downloaded.

4

Enable dumping on the switch.

5

Retain multiple dumps at a single location (enable unique-naming). This task is optional.

6

If routing is enabled, use the add address command to configure an IP address for the LDM port.

7

Perform a test dump.

Configuring an IP Address for the LDM Port
To be able to download information, the module must be configured with an in-band IP address and TCP/IP Host Services enabled.

If an IP address is not assigned to the LDM port, the LDM port defaults to the IP-HST address that has been configured for the module. You can specify a different IP address for the LDM port by using the add address command at the Boot config prompt.

Note: If routing is enabled, you must use the add address command at the Boot Config prompt to configure an IP address for the LDM port.

Refer to the Configuring IP Host Services for information on enabling TCP/IP Host Services.

Creating a Dump File on the TFTP Server
On most TFTP servers, you must create a file on the TFTP server with appropriate network access before a dump operation can be completed successfully.

Specifying the File Name and Server Location
To specify the name of the dump file and the location of the server to which the file is to be downloaded, perform the following steps:

Step Action

1

At the Boot config prompt, enter add dump-entry.

2

Press Return. The following message is displayed:

remote host [0.0.0.0]?

3

Enter the IP address of the remote host to which the dump file is downloaded. The default address is 0.0.0.0.

4

Press Return. The following message is displayed:

via gateway (0.0.0.0 if none) [0.0.0.0]?

5

Enter the IP address of the first hop router, if any. The default address is 0.0.0.0. The default should be used if there is no first hop router.

6

Press Return. The following message is displayed:

timeout in seconds [10]?

7

Enter the desired TFTP timeout value. TFTP is the protocol the switch uses to download the dump file. If the TFTP request is lost, TFTP retransmits the request several times. The timeout value affects the amount of time that passes before the next retry. The default timeout value of 10 seconds is recommended. A value greater than 10 seconds should be used if the path over which the dump is to occur is typically slow.

8

Press Return. The following message is displayed:

file name [ ]?

9

Enter the file name to be assigned to the dump file.

Example: file name [ ]?/usr/tftp/switch11.dmp

10

Press Return. The specified values are set and the Boot config prompt is displayed.

11

Repeat steps 1 through 10 for each remote location you want to configure. You can configure a maximum of eight locations.

Enabling and Disabling Dumps
The switch can dump diagnostic data to the remote system location you specify only after you enable dumping. Disabled is the default setting. Dumping remains disabled until you enable it, and remains enabled until you disable it. (Refer to the Displaying All Boot Config Settings for information about how to determine whether dumping is enabled on a switch.)

Enabling Dumping
To enable dumping of diagnostic data to the configured remote system location, perform the following steps:

Step Action

1

At the Boot config prompt, enter enable dumping.

2

Press Return. Dumping is enabled and the Boot config prompt is displayed.

Disabling Dumping
To disable dumping of diagnostic data to the configured remote system location, perform the following steps:

Step Action

1

At the Boot config prompt, enter disable dumping.

2

Press Return. Dumping is disabled and the Boot config prompt is displayed.

Retaining Multiple Dumps at a Single Location
You have the option of configuring the switch to enable multiple downloads to the same path on a server without overwriting earlier dumps. You do so by configuring the switch to assign a unique name to each dump file that is downloaded. If a unique name is not assigned to each file, then new dumps overwrite the previous file downloaded to the same location.

Note:  Because most TFTP servers require that a dump file be created before the dump can occur, the enable unique-naming command may not be appropriate or useful.

If retaining multiple dumps (unique naming) is enabled, the switch appends a random suffix of one to five hexadecimal characters to the base file name you specify according to instructions in the Specifying the File Name and Server Location section. Disabled is the default setting. Retaining multiple dumps remains disabled until you enable it, and remains enabled until you again disable it. (Refer to the Displaying All Boot Config Settings section for information about how to determine whether unique naming is enabled on a switch.)

Enabling Multiple Dumps
To retain multiple dumps of diagnostic data to the same location on a server, perform the following steps:

Step Action

1

At the Boot config prompt, enter enable unique-naming.

2

Press Return. Retaining multiple dump files is enabled and the Boot config prompt is displayed.

Disabling Multiple Dumps
To disable the retention of multiple dump files downloaded to the same location, perform the following steps:

Step Action

1

At the Boot config prompt, enter disable unique-naming.

2

Press Return. Retaining multiple dump files downloaded to the same location is disabled and the Boot config prompt is displayed.

Performing a Test Dump
DIGITAL recommends that you perform a test dump of newly configured dump locations. Doing so helps verify that the diagnostic data is properly downloaded in the event of a hardware or software failure, or when the Reset/Dump button is pressed.

Note:  For the dump command to work, dumping must first be enabled at the Boot config> prompt.

To test the newly configured dump location, perform the following steps:

Step Action

1

At the Main prompt (Main>), enter dump.

2

Press Return. The following message is displayed:

Dumping will invoke a RESTART.
Are you sure you want to dump memory? (Yes or [No])

3

If you want to continue the download and initiate a restart, enter Yes.

If you do not want to continue the download, enter No.

4

Press Return.

If you entered Yes, the contents of switch memory are downloaded to the remote host and file name you specified. The switch is restarted and the VNswitch module Installation Menu  is displayed.

If you entered No, the Main prompt (Main>) is displayed.

Performing a Test Dump Using the Reset/Dump Button
The Reset/Dump button is located on the module's front panel. When you press this button, the contents of the switch memory is downloaded to the host and filename you specified. The switch is restarted and the VNswitch module Installation Menu is displayed.

Note: If you press the Reset/Dump button without configuring and enabling dump file(s), the VNswitch module is retarted without downloading any diagnostic data.

Displaying Dump Status Information

You can display the status of the most recent dump attempted for a switch. The following information is displayed for each configured dump location:

Information Displayed

Description

Status

Possible outcomes include Successful, Failed, and Not Attempted.

Pathname

The path and file name on the remote server for which the download is configured.

IP Address

The IP address of the remote system to which the dump is downloaded is displayed. If a first hop router is used, the IP address of the first hop router is also listed.

Note: A maximum of eight remote dump locations can be configured for each switch.

To display the status and location of the most recent dump, perform the following steps:

Step Action

1

At the Monitor prompt (Monitor>), enter dump information.

2

Press Return. The status and location information about the most recent dump is displayed.

Example
Monitor>dump information

1:Dump Failed to "/usr/router1.dmp" on 1.2.3.4 error = TFTP protocol error

2:Dump Not attempted to "/usr/tftp/router1.dmp" on 10.23.2.5

3:Dump Successful to "/usr/tftp/router1.dmp" on 1.2.3.4

Viewing Dump File Locations

You can view the locations to which diagnostic data (dump files) are downloaded. To view a list of the file names to be assigned to dump files, and the location of the server on which the files are to reside, perform the following steps:

Step Action

1

At the Boot config prompt, enter list dump-entries.

2

Press Return. A reference number is displayed for each configured interface, followed by a colon. The reference number is then followed by the location to which each dump file is to be downloaded, including the path and name of the file on the remote system, the IP address of the remote host to which the download occurs, the IP address of the first hop router (if any), and the TFTP retransmission timer value.

Example
Boot config>list dump-entries

Dump to:

1: "/usr/local/sw1.dmp" on 1.2.3.4 via 0.0.0.0 for 10 secs

2: "/usr/tftp/sw1.dmp" on 13.12.2.3 via 1.1.2.7 for 10 secs

Modifying Dump File Locations

Modifying the remote locations to which diagnostic data is downloaded includes the ability to change and delete the currently configured location.

Changing the File Name and Server Location
To change the name assigned to a dump file or the location of the server to which the download is to occur, perform the following steps:

Step Action

1

At the Boot config prompt, enter change dump-entry.

2

Press Return. The following message is displayed:

Change which entry [1]?

3

Enter the reference number associated with the entry you want to change. The default is 1. The reference number is the first number, followed by a colon, in each item listed when you display information about dump file locations using the list dump-entries command. Refer to the Viewing Dump File Locations section for additional information.

4

Press Return. The following message is displayed:

remote host [18.123.0.16]?

5

Enter the new IP address of the remote host to which the dump file is downloaded. The previously configured address is the default.

6

Press Return. The following message is displayed:

via gateway (0.0.0.0 if none) [0.0.0.0]?

7

Enter the IP address of the first hop router, if any. The default address is 0.0.0.0. The default should be used if there is no first hop router.

8

Press Return. The following message is displayed:

timeout in seconds [3]?

9

Enter the desired TFTP timeout value. TFTP is the protocol the switch uses to download the dump file. If the TFTP request is lost, TFTP retransmits the request several times. The timeout value affects the amount of time that passes before the next retry. The default timeout value of 10 seconds is recommended. A value greater than 10 seconds should be used if the path over which the dump is to occur is typically slow.

10

Press Return. The following message is displayed:

File name [user/lib/gw/gwimage.ldb]?

11

Enter the path and file name for the location on the remote server to which the dump file is downloaded.

12

Press Return. The specified values are set and the Boot config prompt is displayed.

Deleting a File Name and Server Location
To delete the name to be assigned to the dump file and the location of the server to which the file is downloaded, perform the following steps:

Step Action

1

At the Boot config prompt, enter delete dump-entry.

2

Press Return. The following message is displayed:

Delete which entry [1]?

3

Enter the reference number associated with the entry you want to delete. The default is 1. The reference number is the first number, followed by a colon, in each item listed when you display information about the configured location for the dump file using the list dump-entries command. Refer to the Viewing Dump File Locations section for additional information.

4

Press Return. The specified location is no longer configured for use during downloads.

5

To verify that the entry is deleted, use the list dump-entries command. Refer to the Viewing Dump File Locations section for additional information.

Displaying All Boot Config Settings

You can view a report that shows all switch settings configured using the Boot config process, including the following information:

  • IP address and subnet mask used during reload and dump operations

  • Location and file name of boot files on remote systems

  • Whether dumping is enabled

  • Whether retaining multiple dumps at a single location (unique-naming) is enabled

  • Locations and file names on remote systems to which diagnostic data is dumped

Refer to the Configuring Automatic Image Recovery, Downloading Diagnostic Data for Problem Analysis, and Retaining Multiple Dumps at a Single Location sections for detailed information about dumps, BootP software loads, and enabling multiple dumps, respectively.

To display all switch settings that are configured from the Boot config prompt, perform the following steps:

Step Action

1

At the Boot config prompt, enter list all.

2

Press Return. All switch settings configured from the Boot config prompt are displayed.

Example
Boot config>list all

Interface addresses:

1: 192.9.1.1 on interface 0, mask 255.255.255.252

2: 192.9.223.39 on interface 5, mask 255.255.255.0

Boot files:

1: "/usr/bt/inst.ldb" on 192.9.1.2 through 0.0.0.0 for 3 secs

2: "/usr/bt/in.ldb" on 192.9.2.2 through 192.9.1.4 for 3 secs

Dumping disabled

Unique-naming disabled

Dump to:

1: "/usr/local/sw1.dmp" on 1.2.3.4 via 0.0.0.0 for 10 secs

2: "/usr/tftp/sw1.dmp" on 13.12.2.3 via 1.1.2.7 for 10 secs