The Server Fault Wiki of recommended practices

So I've noticed that there are several recommendations on basic practices on Server Fault, but there doesn't seem to be a cohesive view as to how those recommendations would all fit together. So I thought I would lump these together as a kind of mental exercise to see what the "ServerFault Community IT Department" would look like if it were implemented. This would give a few things: it would make a reasonable wiki (in the true wiki spirit of many contributions), it would provide several links to well-vetted practices, and it would be kind of fun to see what the amalgamation would look like. And who knows, it may even point out some interesting issues between different forms of "best practices", although I would be stunned if there was a conflict hidden in there someplace...

Add your favorites from Server Fault as answers, and I'll re-edit this section with the results. Here's a few catagories to collect different ideas together.

Hardware Configuration(s)

• Server room configuration.
• Server room temperature
• Firmware Updates and Scheduling

Storage Configuration(s)

• Selecting a NAS box
• Linux: Dealing with /tmp
• Linux: Install apps in /var or /opt?

Network Configuration(s)

• checking DNS health and compliance

Security Practice(s)

• Password (General) Best Practices
• Password sharing methods
• Windows Update
• Updating Windows Servers that are hosts for VMs

Network Service(s)

User Service(s)

• User Naming & Deletion

Upgrade Process(es)

Disaster Recovery

• Checking Backups
• Documenting an outage for a post-mortem review

Last Edit: 2010-02-17

From serverfault Avery Payne

• Under security Practices: Best Practice and Solutions for Sharing Passwords

  From Avery Payne

• You could start with consolidating other questions on the site: The Perfect Server Room

  Avery Payne : That's certainly the idea. :) But it's a large task to do this, and I want people to nominate what they think is important. A quick glance at the question section shows there are several thousand questions, but only a few are best-practice related. I figured, this would make a handy place to aggregate those answers together, instead of having to search for each one. So, I'm willing to add answers and maintain this for a bit, if people are willing to chip in a little bit.
  From Joe

• How about adding a link to Documenting an Outage for a Post-Mortem Review as well?

  Avery Payne : done. Thanks for the link!
  From Kamil Kisiel