Page MenuHomePhabricator

reclaim and return all cisco servers
Closed, ResolvedPublic


This task will track all cisco systems currently held in eqiad and codfw. Once they are returned to spares, the SSDS will need to be pulled from ALL systems with SSDs.

We cannot return the SSDs, as we cannot be certain the data was securely erased.

Any HDD systems will need to be wiped.

Once all systems are listed off and ready for return, we'll need to determine where to send these back to. Perhaps finance or legal has copies of the original donation deal/contract? That contract should also include all serials, or atleast the confirmed quantity.

All Cisco USC 250 C1 servers in racktables.

All the servers will need to be listed on this google sheet.

The google sheet (as of 2016-07-19) is missing a few servers at each site. (2 at eqiad, 2 at codfw). Racktables only shows 38 and we should have a total of 40 systems. Sub-tasks have be created for each on-site to audit the cisco listings for their site and correct/update/add as needed.

Once we have all of the Ciscos listed on the tracking sheet (and not in use), we can start the process of getting rid of them.

Event Timeline

Restricted Application added a subscriber: Aklapper. · View Herald Transcript

So in looking up old emails, I found that it is indeed 40 systems:

Both ship sets shipped Oct 6th.
40 cartons.
SO# 50808791 ss2 and ss3
Item - R250-2480805W = 40 servers.
Weight = approx 60 lbs per box
Fed Ex Tracking numbers:
2311010251 (1)
2311010262 (1)

The labsdb1001-1003 systems are still in active use. The current plan is to migrate off of them, onto new labsdb machines. Unfortunately, the timeline for this is in (5-6) months, due to the work involved and the personnel bandwidth.

I'd recommend we audit and get all the other cisco systems accounted for, and return them in advance of the last three.

@Papaul confirms he has 18 on site (and will work via the sub-task to audit them).

@Cmjohnson confirms he has 22 onsite (and will work via the sub-task to audit them).

RobH added a subscriber: jcrespo.

labsdb1002 can be removed, too- it is essentially dead (although maybe the disks are ours?).

The last two systems are labsdb100[13]. Current roadmap off T142807 is to have them decommissioned in January 2018.

This is provisionally a goal, and one that should be doable early next quarter.

RobH removed RobH as the assignee of this task.Dec 21 2017, 4:50 PM
Ottomata triaged this task as Medium priority.Jan 16 2018, 7:23 PM

This task is now pending the last steps of the decommissioning process for labsdb1001 and labsdb1003 (T184832).
I have tried to update the spreadsheet on the task description but it I have no rights.

All CISCO servers are off the rack and fully decommissioned. I do not have any clue on what to do next with them but would love to get them out of the storage area. @RobH can you help with this pleas.e

RobH added a subscriber: wiki_willy.

@wiki_willy has actually been working on this for weeks now, so assigning this to him.

codfw Cisco servers were returned last quarter via Cisco's takeback program:

We'll need to hold off on returning the remaining ones at eqiad until Covid19 restrictions are lifted by Equinix.

Pickup date for remaining Cisco servers at eqiad has been set for June 16. @Jclark-ctr to work with Equinix in prepping for the pickup date. Thanks, Willy

Remaining Cisco servers picked up by Cisco today. Thanks @Jclark-ctr !