• To_Do

To Do

Routines

• update this page

• check voicemail from home every 2 hrs - 587-6999 2958# XXX *

Today

• call kurt re apt

• call dr re anger management class

• automate failover
  • work on manage_ha

• fix dev2 out-of-space

Soon

• read time management book
• figure out good ways to use phone to organize shit
• projects

  • automate failover (see AutoFailover)

  • nagiosify everything
  • add quota checks to nagios
  • make sure nagios checks include the age of the check data
• fixes
  • switch to newer computer in office
  • make sure account creator scripts are committing usermap
  • make rsync.command shut up
  • don't rsync live logs, only compressed ones.
  • make cfrun delete bad sig files

Everything Else

• remove ftp from edison
• rsync
  • log files - only sync old compressed log files
• nagios - no more email alerts!
  • make way for nagios to check if accounts thingy and sysadmind are up on non-live servers - accept non-ssl ? make nagios accept bad SSL name/cert thingies?
  • add awstats checker
  • do I check the web user lists and wireless user lists? I should... They have time stamps
  • add some sort of UPS status to nagios
• cfrun
  • add httpd and mysqld startup checker to cfrun
  • install git on every box
  • check root's crontab and make sure backups are made
• addwebaccnt needs to send email to address not web accnt name when they are different
• make whoslive keep a log and make a way to check that log if arthur is down
• no servers should rely on DHCP! (if they boot before the DHCP servers, I can be SOL)
• make quarterly web user emails
• clean office
  • keep desk clean
  • wipe disks on unused machines
• backups

• documentation - see Documentation Requirements

  • check existing docs
• port infoserver to windows and make agassi provide goodies on infoserver
• make all appropriate filesystems mount "noexec,nodev,nosuid" where appropriate (it's in the standard)
• shibboleth
  • is it dead?
  • should we drop incommon?
• bzr
  • bzrbatch make alert if locked up
  • look for bzr'ed dirs that aren't in bzrbatch cronjob list
• switches should be attached to UPS
• make machines with oimap/stunnel authentication test and restart stunnel
• dhcpd-master
  • print who made changes and what the changes were in emails
• http-watcher
  • use cfrun
  • add 1 sec pause between launch and check
  • make it list running http processes and netstat -an output when restarting
  • SCCC::Util
  • make ps_to_test not die when #items not 11
  • make ps_to_test use wantarray
• cfrun - I think this is all done...
  • installer
  • automatically give my gpg key ultimate trust
  • put _cfrun's home in /home/system
  • dependencies
    • cfrun depends on cronadd.pl
    • my perl mods depend on Module::Build rpm = perl-Module-Build
    • something depends on YAML rpm = perl-YAML

    • AppConfig

    • BSD::Resource
    • Readonly

Documentation Requirements

• Backups
• Documenting the process for restorations
• Documenting what you do for mission critical processes (Shelly has done this for keycards—a good example).
• Listing the locations of needed hardware and software
• Listing a person and/or vendor to contact for expertise beyond that of our remaining network staff - with you under the truck. This might be an expensive resource, but at that time we won’t care.

Long Term Goals

1. Move away from daily status mail toward logging and nagios
2. Documentation

CategoryCategory