This task will outline the proposed resolution to the issue of restbase1001-1006 being off-spec/4 LFF disk bays, when all other restbase systems (both HP and Dell) are SFF disk bays and can hold twice the disks as the original restbase1001-1006.
The majority of this proposal is an update from IRC conversation between @mark & @RobH on 2016-02-04. This task will be the master tracking task for the ordering of 6 new restbase systems for eqiad, and the reclaiming of the existing restbase1001-1006 into operations spares.
After a discussion with the services team, its been decided that operations will absorb the hp restbase1001-1006 into the spare misc servers pool (as we own them, they are not leased), and we'll generate new quotes/orders for 6 new restbase systems in eqiad that better match the specification of restbase1007-1009 & restbase2001-2006.
We'll be removing the samgsung 1TB SSDs from restbase1001-1006, plus the newly ordered additional SSDs, for installation in the newly ordered restbase systems. The current systems have LFF disk bays, but the sleds are for SFF disks going into LFF bays, so we'll want to swap the NEW orders non samgsung 2.5" SFF disks into the existing old systems and put the existing samgsung SSDs into the new systems. Each new system will need to ship with 5 SFF disks in each system.
As this is the master tracking task for the system swaps, there will be multiple procurement and hardware-requests sub-tasks/blockers linked from this.