Administration of a System
The administrator of a server system has numerous tasks to perform. They can be divided into groups according to how often they need carried out. The following lists of tasks ranges from, those that must be performed more often than once a day, to those that need to be performed less often than once a month. The frequency that the administrator may have to do some of the tasks in the following list may vary depending upon the size and complexity of the system.
Remember: If you are responsible for multiple sites these tasks should be carried out at ALL sites.
- As needed:
- Record all system modifications and events in a log
The first page of the log should contain a list of all devices configured on the system.
Make a note of all terminal devices e.g. tty000 and if virtual screens are configured on that terminal (Use tty from the $ prompt).
Where the printers are located, what are they called and where are they connected (Use lpstat -v) .
Is there a modem, ups or cash drawer installed.
As an example:
Device Location Connected Comments Main screen Office tty01 Terminal Cons-1 tty000 Wyse 55 mono Virtual Cons-1 tty0001 Terminal Reception 192.168.9.11 Axel PC Terminal Partners 192.168.9.22 PC Label printer SA Disp. tty008 Printer label1 POS printer Reception on Axel 9.11 Printer pos Cashdrawer Reception Prt POS Attached to POS UPS tty2a UPS monitor Modem tty1A The log book should also include:
- Backup's date and time, did the backup work?, How much data
was transferred, Backup type Incremental, Master or Specific
files.
- Any changes to terminals and printers.
- Any System panic's e.g. Kernel warning: No File.
- When upgrades to either the operating system or application
software was carried out. What has changed and any problems
encountered during the upgrade.
- Application system errors, what was happening on the system?
What screen was the user on?
- Hardware problems, date engineers called, what was the problem?
- Backup's date and time, did the backup work?, How much data
was transferred, Backup type Incremental, Master or Specific
files.
- Be on call for panics, crashes, power spikes, user questions
You are the one point of contact for queries, how do I do
this?, why is this not working?. If you have a power cut, you
have to make sure the system comes back up correctly.
- Maintain security of hardware, software, data file access.
- Make sure that new staff know how to work the terminals, they have been allocated the correct access levels and passwords.
- Record all system modifications and events in a log
- Daily Tasks
- Backups
Check the overnight backup worked (if configured) otherwise ensure a backup is carried out. Swap over tapes.
- Mail
Look at root and UUCPs mail to see if the system has mailed any error/warning messages
- Communications
If Cross updating is installed, use xmenu to check the accessibility of the remote sites and if there is any queued work.
- Devices
Check for unattended login sessions, these can leave records locked, affect the backup and swallow up system resources. Use the 'w' (or who -Hu) command and look at the idle time.
Look for run-away processes. These are jobs that have got out of hand and are consuming resources. Use ps -ef and look at the time column.
Check the printers are running and there are no unexpected jobs queued . Use lpstat -t
- Disk Usage
Check disk space use df -v , this will show a screen similar to:
Mount Filesystem blocks used free %used / /dev/root 1025996 487564 538432 47% /u /dev/u 716800 93830 622970 13% /home /dev/home 1025022 52796 972226 5% /odata/dev/odata 1025022 311342 713680 30%you should monitor the % used column, it should increase gradually, any BIG jump should be investigated.
- Backups
- Quarterly
- Full system backup
Make a full backup of the system not just the Premvet data.
- Root/Boot Floppies
Regenerate your root/boot floppies and check the operation of them.
- UPS
Test the UPS to make sure it is connected and will switch over in case of power failure. The easiest way to test this is shutdown the system and at the 'Safe to power off' point switch OFF the mains to the ups the system should keep running.
- Passwords
Change the root password and review the access levels currently being used.
- Full system backup
- As required
- Hardware
Perform hardware maintenance, clean floppy and tape drives, make sure printers ribbons are ok, no labels jammed etc.
- Add Users
If new members of staff are employed, create a login, allocate a password and show them the basics of the system.
- Remove Users
If any staff have left, retire their account, check and remove their home directory, check and remove their mail files. If the user had access to the root password - change it.
- Security
Use the 'last' command to check for unauthorised login's especially if a dial-in modem is installed.
- Cross Updating
Review the times cross updating takes place and adjust if
needed. Monitor the traffic (uulog / uutraf / tua) usage
for peak time traffic.
- Hardware
Review current hardware usage and requirements.
- Manuals
Read a chapter or two of the Unix manuals to refresh your memory, find out what else the system can do.
- Hardware
What do you need to know.
In order to administer the system you should be familiar with:
- Logging in/out as root
- Directories and the directory structure.
- Adding/Removing users
- Backup's and transferring files to/from tape or disk
- Moving around the system areas
- How to look, list and edit files
- Starting up/Closing down the system
- Root/Boot floppy generation.
Commands you should be familiar with:
cat, more, vi, cd, l, tar