Mira has had a failed nic, and as the nic is onboard to the mainboard, it is a failing mainboard. The system is out of warranty, and has been replaced fully by naos in T162900.
---
This checklist, copied from server lifecycle page (also edited to add some deployment server specific items)
[x] - all system services confirmed offline from production use
[x] - set all icinga checks to maint mode/disabled while reclaim/decommmission takes place.
[x] - remove system from all lvs/pybal active configuration - https://gerrit.wikimedia.org/r/#/c/353116/
[x] - any service group puppet/heira/dsh config removed - https://gerrit.wikimedia.org/r/#/c/353116/
[x] - remove site.pp (replace with role::spare if system isn't shut down immediately during this process.) - system will be SHUTDOWN immediately @robh submitting the patchset) - https://gerrit.wikimedia.org/r/#/c/353116/
[x] - dba task T164968 created for mira's grant removal for being deployment system
[] - deployment task created for rename of deployment-mira.deployment-prep.eqiad.wmflabs
START NON-INTERRUPPTABLE STEPS
[x] - disable puppet on host
[x] - remove all remaining puppet references (include role::spare) - https://gerrit.wikimedia.org/r/#/c/353116/
[x] - power down host
[x] - disable switch port
[x] - switch port assignment noted on this task (for later removal) asw-b-codfw:ge-5/0/13
[x] - remove production dns entries https://gerrit.wikimedia.org/r/#/c/353131/
[x] - puppet node clean, puppet node deactivate, salt key removed
END NON-INTERRUPPTABLE STEPS
[] - system disks wiped (by onsite)
[] - system unracked and decommissioned (by onsite), update racktables with result
[] - switch port configration removed from switch once system is unracked.
[] - mgmt dns entries removed.