Page MenuHomePhabricator

eqiad/codfw: 2x2 VM request for ML-Serve Kubernetes cluster
Closed, ResolvedPublic

Description

Site/Location:EQIAD *and* CODFW
Number of systems: 2x2, (two per location)
Service: ml-serve-ctrl
Networking Requirements: internal
Processor Requirements: 1x
Memory: 2GB
Disks: 20G

Basically these will become the control plane machines for the ml-serve cluster, hosting the k8s control daemons.

Sizing is the same as argon et al, since we have even fewer nodes (but might get more in the future). ML-Team umbrella ticket: T272918

Event Timeline

Created ml-serve-ctrl1001 and ml-serve-ctrl1002 in eqiad, rows B and D.

Created ml-serve-ctrl200[1,2] in codfw, rows C and D (the ones with less VMs)

MAC address for ml-serve-ctrl2001.codfw.wmnet is: aa:00:00:b7:68:43
MAC address for ml-serve-ctrl2002.codfw.wmnet is: aa:00:00:fe:95:00

Change 666931 had a related patch set uploaded (by Elukey; owner: Elukey):
[operations/puppet@production] Add ml-serve-ctrl200[1,2] base config

https://gerrit.wikimedia.org/r/666931

Change 666931 merged by Elukey:
[operations/puppet@production] Add ml-serve-ctrl200[1,2] base config

https://gerrit.wikimedia.org/r/666931

Change 666951 had a related patch set uploaded (by Klausman; owner: Klausman):
[operations/puppet@production] Add ml-serve-ctrl100[1,2] base config

https://gerrit.wikimedia.org/r/666951

Change 666951 merged by Klausman:
[operations/puppet@production] Add ml-serve-ctrl100[1,2] base config

https://gerrit.wikimedia.org/r/666951

jbond triaged this task as Medium priority.Feb 25 2021, 5:38 PM
elukey claimed this task.

VMs created :)