HHMI hub #3136
HHMI hub #3136
Conversation
|
Merging this PR will trigger the following deployment actions. Support and Staging deployments
Production deployments
|
There was a problem hiding this comment.
This LGTM and consider this approved if you provide oauth 15.1 functional config now that the oauth16 upgrade was reverted. Work towards oauth16 can be found in #3144
| cpu: | ||
| guarantee: 0.5 | ||
| # We're on n1-highmem-16 machines | ||
| limit: 14 |
There was a problem hiding this comment.
This probably needs a G after
There was a problem hiding this comment.
@yuvipanda, I think it's ok because this is the cpu limit? I just double-checked and the memory guarantee and limit have the G.
Co-authored-by: Erik Sundell <[email protected]>
Co-authored-by: Erik Sundell <[email protected]>
GeorgianaElena
force-pushed
the
hhmi-hub
branch
from
1e0f051
to
f6cb3ce
Compare
|
Thanks @consideRatio and @yuvipanda for the review. I've now also added (and deployed) a prod hub too. Also updated the auth config for oauthenticator < 16
Per Erik's comment above, I will merge this now. |
|
🎉🎉🎉🎉 Monitor the deployment of the hubs here 👉 https://github.com/2i2c-org/infrastructure/actions/runs/6222063303 |
| @@ -8,7 +8,7 @@ | |||
| prefix = "{{ cluster_name }}" | |||
| project_id = "{{ project_id }}" | |||
|
|
|||
| zone = "{{ cluster_region }}" | |||
| zone = "{{ cluster_region }}-b" | |||
There was a problem hiding this comment.
Quick question in a high-level post review (so I might be totally wrong 😉): do we always want to deploy to the -b zone? Should not get the zone from foo_cluster.tfvars file instead of "building it" from the region?
cc @GeorgianaElena
There was a problem hiding this comment.
Should not get the zone from foo_cluster.tfvars file instead of "building it" from the region?
@damianavila, so this template is actually used by the deployer to generate a foo_cluster.tfvars file. This initial file generated by the deployer is more of a starting point with some "default" basics that are meant to be edited to fulfill all of the needs for the cluster.
do we always want to deploy to the -b zone?
Not necessarily, but it is my understanding that this is the go-to zone for us-west2 (which I believe is the default region if none is requested) if we want to be able to have GPU machines (which for a daskhub is probable) available and a vaster CPU types machines, including AMDs.
There was a problem hiding this comment.
@damianavila, so this template is
I knew I was missing something (well-placed "so I might be totally wrong 😉" message).
Thanks for the clarification!!
Not necessarily,
Yep, it makes sense now that I actually realize what I am seeing 😉
Again, thanks for your patience and feedback.
|
I believe this issue can be closed since the hub has been deployed. |
For #3080
This currently only has a staging hub, deployed at https://staging.hhmi.2i2c.cloud/.
Need to add a prod hub for it