Checkliste Systemadministration

Typische Tätigkeiten eines Systemadministrators - etwas Microsoft-lastig (Quelle Sunbelt W2Knews Electronic Newsletter 10.5.2001):

Daily:

Check event log of every server, fix/try to fix as needed.
Creating new directories, shares, and security groups, new accounts, disabling/deleting old accounts, managing account policies.
Make sure backup runs and make sure the restore works as planned.
Plugging Security holes, in both the OS and apps like IIS.
Exchange Management including DL's, users, etc.
Train the training people, helpdesk people, and end users.
Answer all important emails from CFO/CEO/IT-MIS Director.
Glance over T1-hookups, switches, hubs, make sure everything is green.
Check router logs.
Check firewall logs.
Check if Disaster Recovery Systems are still functioning
Various calls to MS Support for things that really aren't your fault.
Check for free space on all servers, for file pollution and quotas.
Ensure that all server services are running.
Ensure that antivirus definitions are up-to-date.
Run defrag and chkdsk on all drives.
Monitor WINS replication.
Monitor directory replication.
Maintain performance baseline data.
Monitor RAM for runaway processes or memory leaks.
Monitor network traffic with sniffer or NETMON to keep performance up.
Keep Service Pack (and/or) hotfixes current as per company policy.
Monitor Web traffic for indications of attacks.
Install software for users
Monitor user email for corporate policy violations.
Check Print Queues.
Keep a log of everything you have fixed or performed maintenance on.
Make sure all apps are shared.
Permissions and filesystem management.
Check for bad system and .ini files on database server (Btrieve).
Make sure load on database server is acceptable and ghosted users are cleared as well as multiple logons.

Nightly:

Backups
(Next Applies to Terminal Server admins only)
Reboot each Citrix server.
Delete all autocreated printers stuck.
Clear out rogue local profiles.
Backups

Weekly:

Clean Servers, check for .tmp files, and other file pollution.
Implement any new policy, permission, logon script, or scheduled script modifications.
Research, Research, Research.
Change any active monitoring & alerting (third party tools) as needed.
Update Website, External and Intranet, process website log reports.
Check PerfMon, NetMon, (or 3rd party tools) for OK baselines.
Reboot Servers if needed.
Keep up-to-date on IT news regarding my networks.
Evaluate software for System Admin purposes.
Try to get some MCSE study time in.
Performance Monitoring/Capacity Planning- Budgeting for the future.
Uptime/Downtime reports.
Auditing network for unauthorized changes, ideally both from the inside but also outside-in.
Plan for W2K migration.

Monthly:

Rebuild Databases as needed.
Gather statistics on Webservers. Send to CEO/CIO/CTO/CFO (Whomever).
Clean exchange mailboxes.
Change Service Account Passwords (not doing this is Russian roulette).
Convincing your boss that most of this stuff _needs_ to be done.
Extended testing backups with restores.
Maintaining applicable Service Level Agreements.
Set System and Application priorities: If more than one thing is broken, what needs to be fixed first.
Managing off-site storage of backup tapes, whether you take them home or a service picks them up.
IT System vulnerability analysis: like "This mail server uses this mail router- what's the impact if one or both are down (if mail server is down mail router will
store inbound mail and may run out of disk space).
Periodically reviewing all of the above, is documentation up to date? Has the Disaster Recovery Plan been updated to reflect changes in the environment?
Periodically reviewing workload. Are some things no longer done?
Periodically review company technical environment. How can it be improved?

Initial or Occasionally:

Disaster Recovery to alternate site, in case of emergency. Configure and maintain DNS - Internal and External, DHCP, WINS, TCP/IP, etc.
Document the full network.
Rebuild corrupt servers.
Test the Restore Procedure.
Reconfigure domain structure.. again.
Get a performance baseline for things like %Processor Time, PageFaults, Disk Queues.
Initial checklist should include status of administrative and service passwords, status of the backups, check out DHCP scope(s), WINS, DNS, remove
unnecessary protocols.

And then of course: drink lots of coffee, post Dilberts all over your cubicle, surf the web, smoke cigars out back, walk around the office looking busy with a concerned look on your face, make personal phone calls, look for better work on company time, download MP3s and other stuff that slows down the T's .. you know .. important stuff like that.