Your Linux proxy server is usually very reliable, but may occasionally require a little bit of maintenance (this may follow unforeseen power problems)

1. Clearing the Cache

To clear your proxy server's cache,

log in to the proxy server as root

type ./clearcache (press enter)

Clearcache is an automated prompt that will shut down the running squid program, clear and reinitialize the cache, then restart the squid program. The process may take up to five minutes to complete, after which you should be able to browse for Internet resources again. Once complete, log out of the proxy server.

2. Clearing your Log Files

If the box appears to function correctly but it will not respond to web requests, the log files may have grown too large. Ordinarily they are automatically rotated every Sunday, but this may not happen if it is interrupted by a power failure.

See item 3 to determine whether your squid is running, If the last line returned suggests "squid dead" you will need to manually rotate the logs.

log in to the proxy server as root

type:

./rotate_squid_logs (press enter)

service squid start (press enter)

The process should only take a few seconds, after which you should be able to browse for Internet resources again.

3. Testing whether squid is running

log in to the proxy server as root

type service squid status (press enter)

4. File System Checks

This occasionally affects older systems, usually after a power failure.

If the proxy server won't start at all and cannot even be "pinged", you may have a disk problem that needs resolving. Connect a monitor and keyboard to the system to determine what is displayed on the screen. You may see

Disk inconsistency, enter root password or Control D to restart:

Each disk partition should be checked. Firstly, determine your disk type: most systems have ide disks, but a small number have scsi based systems so check whether yours is one of these. If so, replace the "hda" in the example below with sda (or sometimes sdb or sdd), there is sometimes an element of trial and error to locate the drives.

Type in the root password

type fsck /dev/hda1 (press enter)

Press y to accept any messages about inconsistencies

When complete,

type fsck /dev/hda3 (press enter)

Again, press y where prompted (hold down the y key to store some up)

type fsck /dev/hda2 (press enter)

Again, press y where prompted

It is likely that one of the drives reported an error and did not carry out the file system check (fsck). This is to be expected, only two of the three partitions need checking, but this varies from system to system. If none were able to be checked, you may not have located the disks. Try instead: /dev/hdb1, /dev/sda1 or call for advice.

Once your file system has been checked, press Control D to restart the server.

Once the server has restarted you should be able to browse for Internet resources again.

5. Obtaining missing script files

Some systems may not have been updated with copies of the clearcache or rotate_squid_logs scripts. These can be obtained from a couple of "script repositories" on the network:

log in to the proxy server as root

type:

scp col@172.18.200.9:/home/col/repository/clearcache ./

scp col@172.18.200.9:/home/col/repository/rotate* ./

Following each command, you may be asked to accept the system's public key, and will be asked for the col user password.

The scripts will be copied to the CSD on the local system (usually /root/).

6. Building with a floppy Kickstart disk

This is our linux G3 proxy server building information for our tech support staff. It's here only for want of a better location.

Boot from CD1

At boot prompt:

• Insert Prefect floppy
• type:   linux ks=floppy initrd=initrd.img text ramdisk_size=8192

If you need to booting from a USB floppy, type:

linux ks=hd:sda:/ks.cfg initrd=initrd.img text ramdisk_size=8192

Add "dd" to command line if you have a drivers diskette.