The naming of these systems has not been decided between stars & cluster naming. As these have over a week before the ETA (at the time of order), that can be determined before they arrive.

For now, this task will refer to them as puppetmaster100[12]. If these change to star names, two unused star will need to be picked, please see [[Wikitech:Infrastructure Naming Conventions]]

Before systems are used the below steps need to be figured out. These should be kept in sync with the eqiad onsite task T143219.

• - determine hostnames as cluster names or misc/element names
• - determine partitioning scheme, typically match rhodium (but ensure both eqiad and codfw new systems use same partitioning)

puppetmaster2001

• - receive normally via T143216
• - rack in a free misc systems 1GbE rack a5-codfw
• - setup/test bios/drac/redirection
• - dns for mgmt entries (asset tag + hostname)
• - switch port configuration (private vlan)
• - determine partitioning, this is likely to be an sw raid1 lvm with a single / and 10-15% free for expansion on lvm for emergencies.
• - update install_module
• - install OS (jessie)
• - hand off to @akosiaris or @Joe for service implementation. (don't bother to sign puppet/salt keys, as these will take over as puppetmasters.)

puppetmaster2002

• - receive normally via T143216
• - rack in a free misc systems 1GbE rack d5-codfw
• - setup/test bios/drac/redirection
• - dns for mgmt entries (asset tag + hostname)
• - switch port configuration (private vlan)
• - determine partitioning, this is likely to be an sw raid1 lvm with a single / and 10-15% free for expansion on lvm for emergencies.
• - update install_module
• - install OS (jessie)
• - hand off to @akosiaris or @Joe for service implementation. (don't bother to sign puppet/salt keys, as these will take over as puppetmasters.)