Change fallback OVMS URL to point to our built image #768
Conversation
Xaenalt
force-pushed
the
ovms-gpu
branch
3 times, most recently
from
11be14c
to
99ee11c
Compare
Xaenalt
changed the title
|
I'm also not completely convinced we need this per se too, we may want the fallback image to point elsewhere, but I'll let the reviewers decide on that |
|
@Xaenalt We need to upstream this repo and host these images in quay.io/opendatahub instead of pulling from a downstream image repository for a build system that is already in use upstream |
|
I don't have write access to quay.io/opendatahub to make that change sadly. Also, if there's a robot account, we could configure automatic pushes to these places from the openshift-ci builds too |
|
/retest /approve |
|
@Xaenalt: The following test failed, say
Full PR test history. Your PR dashboard. Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. I understand the commands that are listed here. |
|
The fail does not look due to the openvino. |
|
Let me rebase it onto main, that might help |
|
I didn't investigate the root cause but it looks like the operator was not installed successfully OR the odh-core |
|
@Xaenalt A rebase should kick off a new test run |
Image includes Nvidia GPU inferencing support
|
/lgtm |
|
/approve |
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: Jooho, LaVLaS The full list of commands accepted by this bot can be found here. The pull request process is described here
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
| @@ -3,6 +3,6 @@ monitoring-namespace= | |||
| odh-mm-rest-proxy=quay.io/opendatahub/rest-proxy:v0.11.0-alpha | |||
| odh-modelmesh-runtime-adapter=quay.io/opendatahub/modelmesh-runtime-adapter:v0.11.0-alpha | |||
| odh-modelmesh=quay.io/opendatahub/modelmesh:v0.11.0-alpha | |||
| odh-openvino=quay.io/opendatahub/openvino_model_server:2022.3-gpu | |||
| odh-openvino=quay.io/opendatahub/openvino_model_server@sha256:20dbfbaf53d1afbd47c612d953984238cb0e207972ed544a5ea662c2404f276d | |||
This comment was marked as resolved.
This comment was marked as resolved.
Sorry, something went wrong.
There was a problem hiding this comment.
params.env file is for optional input from kfdef file. Openvino is not changed frequently so it is ok to use sha256 directly. however, for the others, we are thinking to use the doc to support a disconnected environment. In other word, the image information in this file is just for default value but we can change it with sha256 with kfdef to support a disconnected environment.
| newName: openvino/model_server | ||
| newTag: "2022.3" | ||
| newName: quay.io/opendatahub/openvino_model_server | ||
| newTag: "2022.3-release" |
This comment was marked as resolved.
This comment was marked as resolved.
Sorry, something went wrong.
There was a problem hiding this comment.
this information has nothing to do with a disconnected environment. The mirror command line is not referring this information.
There was a problem hiding this comment.
but the ImageContentSourcePolicy generated by oc mirror, or, in my case generated manually, needs sha256 type digests, not tags, in manifests. Note my content "restricted network environment installs (on-prem) with mirrored repos and ImageContentSourcePolicy"
| @@ -26,8 +26,8 @@ images: | |||
| newTag: 0.5.2 | |||
|
|
|||
| - name: ovms-1 | |||
There was a problem hiding this comment.
that name is nowhere in the image field of https://github.com/opendatahub-io/odh-manifests/blob/master/model-mesh/runtimes/ovms-1.x.yaml#L37 instead, there is a kustomize parameter value in there from https://github.com/opendatahub-io/odh-manifests/blob/master/model-mesh/base/params.env#L6 Same goes for the other three ServingRunetime yamls referenced ...
Image includes Nvidia GPU inferencing support