|
Table Of Contents
Maintaining the System Software
Recovering the System Software
Maintaining the Hard Disk Storage
WCCP Version 1 Sample Configuration
WCCP Version 2 Sample Configuration
Maintaining the System Software
This chapter covers the tasks required for maintaining a Content Engine.
• Upgrading the System Software
• Recovering the System Software
• Maintaining the Hard Disk Storage
Upgrading the System Software
Cisco Cache software 3.x.x system image files exist as binary files only ( .bin suffix). System images are available to download from Cisco Connection Online (CCO). The .bin file contains the full-image software with the graphical user interface (GUI). To upgrade the system software, the Content Engine FTP client software copies the system image file to Content Engine main memory and then writes the image to Flash memory. See the "Recovering the System Software" section for instructions on booting the Content Engine in the event that the software upgrade fails.
Note We recommend that when loading a system image, you redirect traffic around the Content Engine to prevent degrading HTTP service.
The following procedure describes how to upgrade the Content Engine system software and optionally change the boot parameters.
Step 1 Download the .bin system image file to a Windows NT, LINUX, or UNIX host that is running an FTP server.
Step 2 Perform the following substeps to copy the system image file from an FTP server to the Content Engine Flash memory. In this example, the FTP server has the IP address 192.168.88.89 with the system image file (ce590_example.bin) located in the /images directory:
a. At the EXEC command line of your Content Engine, enter the following:
copy ftp flash 192.168.88.89 /images ce590_example.bin
b. If the FTP server username prompt appears, enter your username.
c. If the FTP server password prompt appears, enter your password.
The following messages appear:
Initiating FTP download. . .
Downloaded 10685440 byte image file
A new system image has been downloaded.
You should write it to flash at this time.
Write to flash [yes]: _
To write the new system image to Flash memory, press Enter to accept the default (yes) or enter yes.
The following message appears:
Ok, writing new image to flash . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
d. At the CLI prompt, reboot the Content Engine. Enter:
reload
Recovering the System Software
The Content Engine has a resident rescue system image that is invoked should the image in
Flash memory be corrupted. A corrupted system image can result from a power failure that occurs while a system image is being written to Flash memory. The rescue image can download a system image to the Content Engine main memory and write it to Flash memory.To install a new system image with the rescue image, perform the following steps:
Step 1 Download the .bin system image file to a Windows NT, LINUX, or UNIX host that is running an FTP server.
Step 2 Establish a console connection to the Content Engine and open a terminal session.
Step 3 Reboot the Content Engine by toggling the power on/off switch.
The rescue image dialog menu appears. The following example demonstrates how to interact with the rescue dialog (user input is denoted by entries in bold typeface):
This is the rescue image. The purpose of this software is to let
you download and install a new system image onto your system's
boot flash device. This software has been invoked either manually
(if you entered `***' to the bootloader prompt) or has been
invoked by the bootloader if it discovered that your system image
in flash had been corrupted.
To download an image, this software will request the following
information from you:
- which network interface to use
- IP address and netmask for the selected interface
- default gateway IP address
- server IP address
- which protocol to use to connect to server
- username/password (if applicable)
- path to system image on server
Please enter an interface from the following list:
0: FastEthernet 0/0
1: FastEthernet 0/1
0
Using interface FastEthernet 0/0
Please enter the local IP address to use for this interface:
[Enter IP Address]: 172.16.22.22
Please enter the netmask for this interface:
[Enter Netmask]: 255.255.255.224
Please enter the IP address for the default gateway:
[Enter Gateway IP Address]: 172.19.19.1
Please enter the IP address for the FTP server where you wish
to obtain the new system image:
[Enter Server IP Address]: 172.16.10.10
Please enter your username on the FTP server (or 'anonymous'):
[Enter Username on server (e.g. anonymous)]: anonymous
Please enter the password for username 'anonymous' on FTP server (an email address):
Please enter the directory containing the image file on the FTP server:
[Enter Directory on server (e.g. /)]: /
Please enter the file name of the system image file on the FTP server:
[Enter Filename on server]: dmdsysimg-test.bin
Here is the configuration you have entered:
Current config:
IP Address: 172.16.22.22
Netmask: 255.255.255.224
Gateway Address: 172.19.19.1
Server Address: 172.16.10.10
Username: anonymous
Password: *************
Image directory: /
Image filename: dmdsysimg-test.bin
Attempting download...
Downloaded 10711040 byte image file
A new system image has been downloaded.
You should write it to flash at this time.
Please enter 'yes' below to indicate that this is what you want to do:
[Enter confirmation ('yes' or 'no')]: yes
Ok, writing new image to flash
..................................................................................Finished writing image to flash.
Enter 'reboot' to reboot, or 'again' to download and install a new image:
[Enter reboot confirmation ('reboot' or 'again')]: reboot
Restarting system.
Initializing memory. Please wait.
Step 4 Log on to the Content Engine with your new admin password. Verify that you are running the correct version by entering the show version command.
Username: admin
Password:
Console> enable
Console# show version
Copyright (c) 1986-2000 by Cisco Systems, Inc.
Cisco Content Engine Software Release 3.00 (build eft Sep 21 2000)
Compiled 18:48:10 Sep 21 2000 by (cisco)
System was restarted on Sat Mar 15 22:12:25 2000.
The system has been up for 1 day, 5 hours, 26 minutes, 53 seconds.
You can now use the management interface or the CLI to enable WCCP or proxy mode and start caching.
Recovering a Lost Password
If a user password is forgotten, lost, or misconfigured, perform the following steps to reset the password.
Note There is no way to restore a lost password. You must reset the password to a new one, as described in this procedure.
Step 1 Establish a console connection to the Content Engine and open a terminal session.
Step 2 At the following prompt, press Return:
Cisco CE boot:hit RETURN to set boot flags:0009
Step 3 When prompted to enter bootflags, enter this value:
0x8000
For example:
Available boot flags (enter the sum of the desired flags):
0x4000 - bypass nvram config
0x8000 - disable login security
[CE boot - enter bootflags]:0x8000
You have entered boot flags = 0x8000
Boot with these flags? [yes]:yes
[Display output omitted]
Step 4 When the Content Engine completes the boot sequence, you are prompted to enter the username to access the CLI. Enter the default administrator user name (admin):
Cisco Content Engine Console
Username: admin
Setting the configuration flags field to 0x8000 lets you into the system, bypassing all security. Setting the configuration flags field to 0x4000 lets you bypass the NVRAM configuration.
Step 5 Once you see the CLI prompt, set the password for the user using the user modify username name password new-password command in the global configuration mode.
Step 6 Use the write command to save the configuration change.
Maintaining the Hard Disk Storage
This section describes the procedures for maintaining the disk storage. The principal disk partitions are for the system file system and cache file system (sysfs and cfs, respectively). The sysfs partitions are used principally to store output from logging processes, the cfs partitions store cached network content. Currently, only sysfs and cfs are supported, and only one partition is supported per disk. See the release notes for the specific version of the Cache Software you are running for further information.
The Cisco Content Engine and the Cisco Storage Array may be shipped with unformatted disks, in which case sysfs and cfs disks must be manually partitioned, formatted, and mounted. Use the show disks command to display the complement of disks that can be supported by your Content Engine, as well as the partition and mounting status of any prepared disks.
The following example is derived from a Content Engine 590 connected to a Storage Array (SA-6). The output shows a single sysfs disk mounted on the default volume /local1 and a single unmounted cfs disk. The disks marked "unknown unmounted" are unformatted disks that occupy slots 2-4 of the Storage Array, and those marked "NOT available" indicate that slots 5-7 are empty.
Console# show disks
disk00 (scsi host 0, channel 0, id 0) FS: sysfs mounted at: /local1
disk01 (scsi host 0, channel 0, id 1) FS: sysfs mounted at: /local2
disk02 (scsi host 3, channel 0, id 0) FS: unknown unmounted
disk03 (scsi host 3, channel 0, id 1) FS: unknown unmounted
disk04 (scsi host 3, channel 0, id 2) FS: unknown unmounted
disk05 (scsi host 3, channel 0, id 3) FS: NOT available
disk06 (scsi host 3, channel 0, id 4) FS: NOT available
disk07 (scsi host 3, channel 0, id 5) FS: NOT available
Use the show disk-partitions commands to determine the type of partition and the size of a particular disk.
Console# show disk-partitions disk00
Disk size in blocks: 17921835
Partition 1: SYSFS offset: 16 blocks, size 17921008 blocks
Console# show disk-partitions disk01
Disk size in blocks: 17921835
Partition 1: CFS offset: 16 blocks, size 17921008 blocks
Typical disk configurations for Content Engines are shown in Table 2-1:
Preparing a Sysfs Disk
All Content Engine logging processes store output on sysfs volumes. A sysfs volume must be mounted before these processes can operate. Two sysfs volumes are currently supported by the Cache software, mounted as /local1 and /local2. Each volume makes up an entire disk. The content of sysfs disks can be viewed or manipulated with the commands listed in Table 2-2.
A typical sysfs setup uses either or both of the first two disks (disk00 and disk01) as the sysfs. The first sysfs disk that is partitioned, formatted, and mounted cannot be unmounted. The disk is checked and repaired automatically when it is mounted during the bootup or reload sequence.
Caution Anytime a volume is partitioned or formatted, all the data on the disk is lost.
To prepare a sysfs disk volume, perform the following steps at the EXEC level CLI:
Step 1 Identify the disk to be prepared as a syfs disk with the show disks EXEC command.
In the steps that follow, the disk is disk00.
Step 2 Unmount the intended volume if it is mounted as a cfs volume.
Step 3 Enter the following command:
disk partition disk00 sysfs
Step 4 Enter the following command:
sysfs format disk00
Step 5 Enter the following command:
sysfs mount disk00 local1
The sysfs disk can be mounted at volume name /local1 or /local2
Step 6 Enter the show disks command to verify the disk status.
Preparing a Cfs Disk
Disks formatted with the cache file system (cfs) store cached content from the network. The content on a cfs disk cannot be examined. Commands related to maintaining the cfs are listed in Table 2-3.
Before a disk can be used to store cached content, it must be partitioned, formatted, and mounted with the cache file system. The disk manufacture EXEC command automatically performs these three functions for cfs volumes. Be sure to use the disk manufacture command if the disk was never formatted, or when the last formatting of the disk was other than from the Content Engine.
The disk prepare command unmounts, partitions, and formats the disk, but does not mount the new cfs volume.
Caution Anytime a volume is partitioned or formatted, all the data on the disk is lost.
To prepare a cfs disk volume, perform the following steps at the EXEC level CLI:
Step 1 Identify the disks to be prepared as cfs disks with the show disks EXEC command.
In the steps that follow, the disk is disk07.
Step 2 Enter the following command:
disk manufacture disk07
Disk07 will be unmounted (if necessary), partitioned, formatted, and mounted as a cfs volume.
Step 3 Repeat Step 2 for all disks intended as cfs disks.
Step 4 Enter the show cfs volumes command to verify the disk status.
If all cfs volumes are unmounted, the Content Engine will continue to fulfill client content requests directly with the origin servers, though no content can be cached. To unmount a cfs volume, use the
cfs unmount EXEC command. The cfs unmount command frees the in-memory data structures that map to the physical (disk) cfs partition.The cfs clear command deletes nonbusy objects from the specified cfs volume. A nonbusy object is an object that is not being accessed (read or written). The cfs clear command (without force) deletes all possible objects without generating a broken GIF or HTML message to the client.
The cfs clear force command deletes all objects, busy or nonbusy, and may generate broken GIF or HTML messages for objects that were being read from the disk when the command was executed. If an object is being written to the Content Engine disk when a cfs clear force command is executed, the application stops caching that object but still delivers the object from the Web server to the client.
The cfs reset command unmounts, formats, and mounts a specified volume. Unmounting a volume can result in broken GIF or HTML messages for objects that are being read from the disk (cache hits) when the command is executed. When a cfs volume is reset, all cfs data on that volume is lost.
Note The cfs reset command can be invoked on unmounted volumes.
The cfs format command creates the cache file system internal "dbs" for the cfs partition of the disk if the volume is unmounted. It formats the cfs partition to prepare it for a cfs mount. The cfs mount command creates and maps data structures in memory to the cfs partition.
The cfs sync command synchronizes the cache file system contents from memory to disk. Although synchronization is performed at regular intervals while the Content Engine is running, this command can be used to ensure that all data is written to disk before you reset or turn off the Content Engine. Synchronization can also be done with the cache synchronize command.
Disk Preparation Procedure for the Cisco Storage Array
The Storage Array hard disks may be shipped without formatting or partitions. You must enter the disk manufacture EXEC command for each newly installed hard disk. Use the disk and cfs commands for subsequent maintenance of the disks. It is assumed that all Storage Array disks are to be formatted as cfs volumes.
Perform the following procedure when installing a new hard disk for the Cisco Storage Array:
Step 1 Make sure the disk drives are properly inserted into the Storage Array slots, and that the SCSI cable connectors on the Content Engine and Storage Array are tightened.
Step 2 Power on the Storage Array, and then power on the Content Engine.
Step 3 Determine the disk numbers of the unformatted disks with the the show disks command.
The show disks command displays all unformatted disks as "unknown unmounted." In the following example, a 590 Content Engine is connected to a fully populated SA-6 Storage Array in joined bus mode. Disks 02-07 are not partitioned, formatted, or mounted.
Console# show cfs volumes
disk00 (scsi host 0, channel 0, id 0) FS: sysfs mounted at: /local1
disk01 (scsi host 0, channel 0, id 1) FS: sysfs mounted at: /local2
disk02 (scsi host 3, channel 0, id 0) FS: unknown unmounted
disk03 (scsi host 3, channel 0, id 1) FS: unknown unmounted
disk04 (scsi host 3, channel 0, id 2) FS: unknown unmounted
disk05 (scsi host 3, channel 0, id 3) FS: unknown unmounted
disk06 (scsi host 3, channel 0, id 4) FS: unknown unmounted
disk07 (scsi host 3, channel 0, id 5) FS: unknown unmounted
The leftmost hard disk inserted in an SA-6 Storage Array bus is always disk 02. Counting to the right, the next disk is 03, the next disk is 04, and so on. There can be empty slots between disks on the same bus, but this is not recommended.
In a two-host, split-bus configuration, each bus is counted independently. For example, in a split-bus, fully populated SA-6 Storage Array, bus 0 disk drive numbers are 02, 03, and 04, and bus 1 disk drive numbers are 02, 03, and 04. If the first disk on bus 1 is removed (slot 5 is empty) and the Content Engine is rebooted, bus 0 numbers are still 02, 03, and 04, but bus 1 numbers are 02 and 03. The empty disk slot is skipped, and the count begins with the first detected disk on bus 1.
Step 4 Enter the disk manufacture command for each new hard disk to be installed.
In the following example, the disks with target numbers 02 and 03 are partitioned, formatted, and mounted with the disk manufacture EXEC command.
Console# disk manufacture disk02 Making disk07 a CFS disk and mounting it
Erased all partitions on disk02
Partitioned disk disk02 as CFS disk
Formatted disk to CFS
disk02 mounted as CFS disk
Done disk manufacture successfully
Console# disk manufacture disk03
Making disk03 a CFS disk and mounting it
Erased all partitions on disk03
Partitioned disk disk03 as CFS disk
Formatted disk to CFS
disk03 mounted as CFS disk
Done disk manufacture successfully
Step 5 Enter the show disks command to verify that the disks are mounted.
ContentEngine# show disks
disk00 (scsi host 0, channel 0, id 0) FS: sysfs mounted at: /local1
disk01 (scsi host 0, channel 0, id 0) FS: sysfs mounted at: /local2
disk02 (scsi host 0, channel 0, id 8) FS: cfs mounted
disk03 (scsi host 0, channel 0, id 9) FS: cfs mounted
disk04 (scsi host 0, channel 0, id 10) FS: cfs mounted
disk05 (scsi host 0, channel 0, id 11) FS: cfs mounted
disk06 (scsi host 0, channel 0, id 12) FS: cfs mounted
disk07 (scsi host 0, channel 0, id 13) FS: cfs mounted
Sample Configurations
The following sample configurations are outputs of the show running-config command and can be used as a guide for configuring the Content Engine in WCCP Version 1 and Version 2 environments. If a feature is not enabled, it does not appear in the running configuration display, though settings configured for that feature are displayed.
WCCP Version 1 Sample Configuration
The following is a sample configuration of a Content Engine in a WCCP Version 1 environment:
console# show running-config
Building configuration...
Current configuration:
!
!
logging event-export critical-events alert kernel
logging 10.7.8.9
!
user add admin uid 0 password 1 "ceSzbyeb" capability admin-access
user add britt uid 5001 password 1 "bbdze9eSbS" capability telnet-access
!
!
!
hostname cache1D
!
interface ethernet 0
ip address 10.1.1.34 255.255.255.0
ip broadcast-address 10.1.1.255
exit
!
!
interface ethernet 1
exit
!
ip default-gateway 10.1.1.2
ip name-server 10.2.2.2
ip domain-name cisco.com
cron file /local/etc/crontab
ntp server 10.6.14.7
!
!
http proxy incoming 8080
icp server remote-client 10.4.6.8 no-fetch
url-filter good-sites-allow
wccp home-router 10.1.2.1
wccp version 1
!
transaction-logs destination disk
transaction-logs interval 750
!
trusted-host 10.4.4.4
!
end
cache1D#
WCCP Version 2 Sample Configuration
The following is a sample configuration of a Content Engine in a WCCP Version 2 environment:
console# show running-config
hostname 590-1
!
http proxy incoming 8080
!
!
!
ip domain-name cisco.com
!
exec-timeout 30
!
!
!
interface FastEthernet 0/0
ip address 172.16.16.19 255.255.255.0
exit
interface FastEthernet 0/1
shutdown
exit
!
ip default-gateway 172.16.55.1
!
ip name-server 172.16.2.200
!
!
!
bypass static 10.1.17.1 172.10.7.52
bypass static any-client 172.10.7.52
bypass static 10.1.17.2 any-server
bypass auth-traffic enable
proxy-protocols outgoing-proxy exclude cruzio.com
!
!
!
wccp router-list 1 10.1.1.1
wccp web-cache router-list-num 1
wccp version 2
!
transaction-logs archive interval every-day at 11:26
transaction-logs file-marker
transaction-logs export enable
transaction-logs export ftp-server 1.1.1.1 username passw /ftpdirectory
transaction-logs export ftp-server 2.2.2.2 myname mypass /tmp/logfiles
!
!
username admin password 1 .9ENIWf7GjMm2
username admin privilege 15
!
!
Posted: Fri Jul 1 13:50:34 PDT 2005
All contents are Copyright © 1992--2005 Cisco Systems, Inc. All rights reserved.
Important Notices and Privacy Statement.