Page MenuHomePhabricator

Move cloudsw2-d5-eqiad servers to cloudsw1-d5-eqiad
Closed, ResolvedPublic

Description

Similar to T334641: Move cloudcephosd1021 to cloudsw1-c8-eqiad but with more servers.

cloudsw2-d5-eqiad has been brought up as a stopgap solution as cloudsw1-d5 was becoming full due to servers using two uplinks.

Thanks to T319184: Move WMCS servers to 1 single NIC cloudsw1-d5-eqiad now have 8 free switch ports, 10 more once the last 10 cloudvirt servers will be migrated to 1 NIC.

Those are the cloudsw2 servers:

As it's a total of 10 ports, we will first need T319184 to be completed first. Until then make sure no new servers go to cloudsw2.

Moving them to cloudsw1-d5 will allow us to decom cloudsw2-d5 and thus cleanup the configuration, help to keep a stable infra, free up (U & cabling) space in the rack.

As it's the same vlan on both switches there is no re-numbering needed, the process is straightforward (and they don't need to move physically):

  1. Configure the new switch ports
  2. Move the cables
  3. Un-configure the old switch ports
  4. Update the cables in Netbox

They can also be done one at a time if needed.

Event Timeline

Mentioned in SAL (#wikimedia-cloud) [2023-06-08T11:54:17Z] <dcaro> powering off toolsbeta-test-k8s-etcd-22 (T334644)

Mentioned in SAL (#wikimedia-cloud) [2023-06-08T12:00:03Z] <dcaro> powering off tools-k8s-etcd-18 (T334644)

@Jclark-ctr cloudvirtlocal1001 is ready to be replugged (it's up, let me know if you need it down)

Mentioned in SAL (#wikimedia-cloud-feed) [2023-06-08T12:05:40Z] <wm-bot2> Draining cloudvirt1047.eqiad.wmnet (T334644) - cookbook ran by dcaro@vulcanus

Mentioned in SAL (#wikimedia-cloud-feed) [2023-06-08T12:06:39Z] <wm-bot2> Set cloudvirt cloudvirt1047.eqiad.wmnet maintenance (downtime id: 769349bf-465f-4f0c-a8f3-f2423631ba7e, use this to unset) (T334644) - cookbook ran by dcaro@vulcanus

dcaro updated the task description. (Show Details)

Mentioned in SAL (#wikimedia-cloud-feed) [2023-06-08T12:16:30Z] <wm-bot2> Draining cloudvirt1047.eqiad.wmnet (T334644) - cookbook ran by dcaro@vulcanus

Mentioned in SAL (#wikimedia-cloud-feed) [2023-06-08T12:17:16Z] <wm-bot2> Set cloudvirt cloudvirt1047.eqiad.wmnet maintenance (downtime id: 02920314-1efe-4934-ad81-d2a6cf2e17ab, use this to unset) (T334644) - cookbook ran by dcaro@vulcanus

Mentioned in SAL (#wikimedia-cloud-feed) [2023-06-08T12:17:19Z] <wm-bot2> Drained cloudvirt1047.eqiad.wmnet (T334644) - cookbook ran by dcaro@vulcanus

dcaro updated the task description. (Show Details)

cloudcephosd1023 Relocated to cloudsw1

I see that there are now enough free ports on cloudsw1-d5-eqiad, @Jclark-ctr @dcaro I'm wondering if you could resume the migration ?
I'm happy to help if needed.

Aklapper subscribed.

@Jclark-ctr Removing task assignee as this open task has been assigned for more than two years - See the email sent on 2025-05-22.
Please assign this task to yourself again if you still realistically [plan to] work on this task - it would be welcome!
If this task has been resolved in the meantime, or should not be worked on by anybody ("declined"), please update its task status via "Add Action… 🡒 Change Status".
Also see https://www.mediawiki.org/wiki/Bug_management/Assignee_cleanup for tips how to best manage your individual work in Phabricator. Thanks!

@Aklapper @ayounsi I hadn’t commented earlier because we needed to verify onsite that we still had enough available ports to avoid blocking the current installs when migrating off cloudsw2-d5-eqiad. We still have T378828 in progress with @VRiley-WMF, which will require 4x 10G ports on cloudsw1-d5-eqiad.

Currently, cloudsw2-d5-eqiad has 5 ports that still need to be moved over. From a quick look at Netbox, it appears that cloudsw1-d5-eqiad has 12 open ports available.

All done, thanks a lot!