nerc-project / operations

Issues related to the operation of the NERC OpenShift environment
1 stars 0 forks source link

Track down and disable containers still running from Projects that are archived #522

Closed joachimweyl closed 3 days ago

joachimweyl commented 3 months ago

Motivation

We have a few projects that are archived but the containers are still running. This means we end up with a project in the invoice with usage but no PI associated with it. One of the projects is using at least 2 GPUs.

Completion Criteria

no remove all containers associated with projects that are archived.

Description

List of Projects with no PI

cost last month Project ID
$8259.34 d53dfc39548a48edb024b6e58d8716e6
$29.02 df4b64c4366742ebb8b8f4b9966319f7
$19.34 0b406404755945ecbab2109d32bce661
$19.34 c3d11da447b1473bb77d74cc5b40cdd5
$9.67 05cca114ea014b8ea588016e9d2aff04

Completion dates

Desired - 2024-04-30 Required - TBD

joachimweyl commented 3 months ago

Can this be automated?

joachimweyl commented 1 month ago

@Milstein have these projects been manually shut down?

Milstein commented 1 month ago

Not found Projects:

| 0b406404755945ecbab2109d32bce661 | <Not Found>            |
| c3d11da447b1473bb77d74cc5b40cdd5 | <Not Found>            |

I found these projects detail:

| d53dfc39548a48edb024b6e58d8716e6 | nerc-admin                     |

--> I added this to our non_billed_project list.

| df4b64c4366742ebb8b8f4b9966319f7 | Testing NERC for BU-fff6a31    |

$ ospurge --verbose --purge-project "Testing NERC for BU-fff6a31" --os-cloud=nerc --> purged!

| 05cca114ea014b8ea588016e9d2aff04 | Testing NERC for BU-f778c4d        |

$ ospurge --verbose --purge-project "Testing NERC for BU-f778c4d" --os-cloud=nerc --> purged!


When we change allocation status from Revoked --> Active This creates a new Allocation Project ID and removes the older Project details!

| a7e1529b37464e328c46d84dd9365735 | Testing NERC for BU-d79ec8 |

$ ospurge --verbose --purge-project "Testing NERC for BU-d79ec8" --os-cloud=nerc --> purged!

| 5e47421927c443dabe934f4e6147ce67  | Testing NERC for BU-9719ac  | 

--> Just decommissioned with zero storage quota so will not get charged afterward.

Milstein commented 1 month ago

Manually cleaning up using os-purge except the nerc-admin that is our main internal project.

joachimweyl commented 1 month ago

@Milstein please close this issue once you clear those 4 items.

Milstein commented 1 month ago

@joachimweyl : can you see if in invoicing we can see more details about:

| 0b406404755945ecbab2109d32bce661 | <Not Found>            |
| c3d11da447b1473bb77d74cc5b40cdd5 | <Not Found>            |
joachimweyl commented 1 month ago

661 used 1488 OpenStack CPU SU in March, it was associated with my project Migrating from MOC to NERC-f04a7e9 before it was expired dd5 used 1488 OpenStack CPU SU in only March, not sure what this was.

Milstein commented 1 month ago

For your project "Migrating from MOC to NERC", I can see two allocations:

Migrating from MOC to NERC-fa886c1 | 5587ca2a20c74d219ad20381dfcf5d98
Migrating from MOC to NERC-feeace3 | 11317353d42e41d4be2ee82f5b0038ec

So this is strange behavior, how the Project ID has changed!

joachimweyl commented 1 month ago

What are next steps for closing this out?

Milstein commented 1 month ago

I am adding two more BU projects to this list:

Allocated Project ID    68543e14d86a42e59e6542b8c414f4c5
Allocated Project Name  MSoM Platform-1e370c

Allocated Project ID    3080a54ffbec45659fbc125fccbbac8e
Allocated Project Name  Fitness activity recognition from video-fc77646
joachimweyl commented 1 month ago

@Milstein were you able to clean up the two BU projects?

joachimweyl commented 1 month ago

Blocked until we can confirm they do not show up in the invoice.

joachimweyl commented 4 weeks ago

They showed up in May data but that is not unexpected as they were turned off during May, we will have to wait until June data to be sure they are cleared.

joachimweyl commented 3 days ago

all gone.