Decommissioning systems

Introduction

This is a basic checklist of things for me to do when I decommission a system.

Procedure

1. Shutdown the machine.
2. Expect complaints from:
   • Check_MK
   • rdw (or other backup software)
   • rocon-check
3. Wait two weeks (and possibly add a wake up call to my calendar).
4. If the system is a VM then:
   1. On each VM server run:

      virsh dumpxml <vm-name> > ~/tmp/<vm-name>.xml
      virsh undefine <vm-name>
      

   2. On each storage server that hosts the VM’s image run:

      drbd-overview | grep <vm-name>
      drbdadm down <drbd-vol-name>
      ...
      rm /etc/drbd.d/<vm-name>.res
      ...
      lvs | grep <vm-name>   #  check not 'a'
      lvremove /dev/vg0/<vm-name>
      ...
      

5. Remove the system from Check_MK.
6. Remove the system from the backup server’s configuration but do not delete the backups themselves yet by running:

   ssh root@<backup-server>
   rdw2 --del-dle nightly <host>-boot
   rdw2 --del-dle nightly <host>-root
   ...
   rdw2 --del-client <host>

7. Remove the system from the rocon server’s /etc/rocon.conf file.
8. Remove the system from DNS.
9. Remove the system from mini-checkmk.
10. Remove the system from cert-checker.
11. Remove the system’s own WordPress page, its link on the Computing page and any other links that a search results in.
12. Wait six weeks (and possibly add a wake up call to my calendar).
13. Remove the system’s backups.

See also

• Computing