!caaCTdhYlLGpbJVvly:matrix.org

kubeflow-pipelines

741 Members
2 Servers

Load older messages


SenderMessageTime
10 May 2022
@_slack_kubeflow_U02TJGKTTRN:matrix.orgMicah Smith I don't follow this completely, but you can create a third component that sits between the two and is just responsible for reshaping the param 21:23:20
11 May 2022
@_slack_kubeflow_U02JYD39G57:matrix.orgZoltán R. Jánki changed their display name from _slack_kubeflow_U02JYD39G57 to Zoltán R. Jánki.11:24:06
@_slack_kubeflow_U02JYD39G57:matrix.orgZoltán R. Jánki set a profile picture.11:24:08
@_slack_kubeflow_U02JYD39G57:matrix.orgZoltán R. Jánkiimage.png
Download image.png
11:24:10
@_slack_kubeflow_U02JYD39G57:matrix.orgZoltán R. Jánki Hello Everyone, We would like to build a hybrid cluster using WSL. Native is not an option for us because the NFS storage driver is not supported. In WSL br_netfilter module does not exist, but we cannot activate the VXLAN. However, the Kubeflow can communicate with the pods, the pods cannot communicate with each other (e.g.: they cannot push artifacts, logs, etc.). This issue stops the pipeline at the beginning. Here, you can find the logs too. Do you have any suggestion how to solve the problem mentioned above? Thank you in advance! Best regards, Zoltán J. 11:24:10
@_slack_kubeflow_U030GQ7G4RG:matrix.orgmax cantor the problem is with how the input args are themselves parsed unfortunately, otherwise that would make a lot of sense 😕. Anyway, we're using a different approach now where we write the output to a mounted nfs volume, and still haven't quite figured out the best way to handle some inline arguments we want to allow for, but that's more to do with our own setup and not a kfp issue. Thanks though! 13:12:29
12 May 2022
@_slack_kubeflow_U03FJHZLY7J:matrix.org_slack_kubeflow_U03FJHZLY7J joined the room.21:39:42
13 May 2022
@_slack_kubeflow_U03EVG1CG2K:matrix.org_slack_kubeflow_U03EVG1CG2K joined the room.02:47:44
14 May 2022
@_slack_kubeflow_U02AYBVSLSK:matrix.orgAlexandre Brown Hello, is it possible to use KFServing component (https://github.com/kubeflow/pipelines/blob/master/components/kubeflow/kfserving/component.yaml) For KFP SDK v2 compatible mode? Getting an error when applying it. kfserving_component = load_component_from_url('https://raw.githubusercontent.com/kubeflow/pipelines/master/components/kubeflow/kfserving/component.yaml') kfserving_component( action='apply', inferenceservice_yaml=get_inference_service_definition_task.output ) ERROR : File /opt/conda/envs/notebookenv/lib/python3.9/site-packages/kfp/compiler/v2_compat.py:170, in update_op(op, pipeline_name, pipeline_root, launcher_image) 161 runtime_info["inputArtifacts"][artifact_name] = artifact_info 163 for parameter, spec in sorted( 164 component_spec.output_definitions.parameters.items()): 165 parameter_info = { 166 "type": 167 pipeline_spec_pb2.PrimitiveType.PrimitiveTypeEnum.Name(spec.type 168 ), 169 "path": --> 170 op.file_outputs[parameter], 171 } 172 runtime_info["outputParameters"][parameter] = parameter_info 174 for artifact_name, spec in sorted( 175 component_spec.output_definitions.artifacts.items()): 176 # TODO: Assert instance_schema. KeyError: 'InferenceService Status' Running Kubeflow 1.4.1 + KFServing 0.6.1 (the version delivered with Kubeflow) 12:09:25
@_slack_kubeflow_UERAV8XJ8:matrix.orgshotarok Hello, I have a question about a backup for the reoccurring jobs. I've been taking a backup on the database, but I'm curious about the other possible approaches. How do you persistent the reoccurring jobs settings? Thanks! 12:29:53
@_slack_kubeflow_U03EDPHJPFZ:matrix.orgAvi Klaiman joined the room.20:01:54
15 May 2022
@_slack_kubeflow_U03DGNK5EBT:matrix.orgAmanjeet Sahu joined the room.08:38:47
16 May 2022
@_slack_kubeflow_U03FWQYAPR7:matrix.orgSteven Tobias joined the room.02:23:53
@_slack_kubeflow_U03FWQYAPR7:matrix.orgSteven Tobias changed their display name from _slack_kubeflow_U03FWQYAPR7 to Steven Tobias.02:46:48
@_slack_kubeflow_U03FWQYAPR7:matrix.orgSteven Tobias Any ideas? 02:46:50
@_slack_kubeflow_U03FWQYAPR7:matrix.orgSteven Tobias set a profile picture.02:46:52
@_slack_kubeflow_U03FWQYAPR7:matrix.orgSteven Tobias I deployed from the manifests example, but can't seem to get auth working for the life of me. Showing this when trying to list experiments from kfp pipelines Internal error: Unauthenticated: Request header error: there is no user identity header when I click "experiments" in UI I get RBAC Access denied 02:46:52
@_slack_kubeflow_U03FWQYAPR7:matrix.orgSteven Tobias Restarted EVERYTHING and it worked 😓 03:11:34
@_slack_kubeflow_UL38EU0BG:matrix.orgThomas Ounnas https://tenor.com/bj95q.gif 16:39:00
@_slack_kubeflow_U03GCSE8U48:matrix.org_slack_kubeflow_U03GCSE8U48 joined the room.20:44:07
17 May 2022
@_slack_kubeflow_U02P9TMR92N:matrix.orgAmit Gelber Serge Tarkovski did it work? 07:54:53
@_slack_kubeflow_U02P9TMR92N:matrix.orgAmit Gelber Can I use artifact functionaliy in custom component? like write HTML? 07:56:02
@_slack_kubeflow_U02AYBVSLSK:matrix.orgAlexandre Brown What does "Artifact functionality" mean? 10:20:37
@_slack_kubeflow_U02P9TMR92N:matrix.orgAmit Gelber https://github.com/kubeflow/pipelines/issues/6116 10:21:57
@_slack_kubeflow_U02P9TMR92N:matrix.orgAmit Gelber works! 11:39:20
@_slack_kubeflow_U03DU40KGS3:matrix.orgHarcharan Singh joined the room.15:35:38
@_slack_kubeflow_U02Q0DARM8B:matrix.org_slack_kubeflow_U02Q0DARM8B joined the room.15:46:16
@_slack_kubeflow_U02AYBVSLSK:matrix.orgAlexandre BrownRedacted or Malformed Event16:42:54
@_slack_kubeflow_U02AX5UF7L0:matrix.orgClark Updike Anyone know how to get around this error when I click Experiments in the UI? "error":"Invalid input error: Invalid resource references for experiment. ListExperiment requires filtering by namespace." 18:37:57
@_slack_kubeflow_U02AX5UF7L0:matrix.orgClark Updike This is in Multi-User mode with Kubeflow 1.2 18:38:57

Show newer messages


Back to Room ListRoom Version: 6