search

MicrosoftDocs / azure-docs

Open source documentation of Microsoft Azure
https://docs.microsoft.com/azure
Creative Commons Attribution 4.0 International
10.29k stars 21.47k forks source link

Bug: #82035

Closed mysoultan closed 2 years ago

mysoultan commented 3 years ago

The repo container three version.

in version v0.1.1 gives Error: unable to build kubernetes objects from release manifest: unable to recognize "": no matches for kind "Deployment" in version "apps/v1beta2".. I have checked the templates, and found that that the apiversion stil referring to apiVersion: apps/v1beta2.. where the K8s version is 1.20

in version 0.3.3.. it gives the following error Error: unable to build kubernetes objects from release manifest: unable to recognize "": no matches for kind "PodDisruptionBudget" in version "policy/v1"

Document Details

⚠ Do not edit this section. It is required for docs.microsoft.com ➟ GitHub issue linking.

YutongTie-MSFT commented 3 years ago

Thanks for the feedback! We are currently investigating and will update you shortly.

YutongTie-MSFT commented 3 years ago

@mysoultan Thanks for the feedback again. Assign to author for update.

@aahill Could you please check on this issue and update the document? Thanks.

aahill commented 2 years ago

assign:eric-urban

eric-urban commented 2 years ago

Sorry for the delay. At this point there is no clear doc issue so if its still an issue please reach out for help via one of the options here: Create an Azure support request

eric-urban commented 2 years ago

please-close

Bunoviske commented 1 year ago

Fixed the issue by manually changing the apiVersion of each yaml file (Deployment, PodDisruptionBudget and HorizontalPodAutoscaler). For all files, I just removed the "beta2" or "beta1" part.

First I pulled the Helm Chart locally: helm pull microsoft/cognitive-services-speech-onpremise --untar

Previously I had the file textToSpeech-autoscaler.yaml :

{{- if and .Values.enabled .Values.service.autoScaler.enabled }}
apiVersion: autoscaling/v2beta2
kind: HorizontalPodAutoscaler

...

Now I changed to:

{{- if and .Values.enabled .Values.service.autoScaler.enabled }}
apiVersion: autoscaling/v2
kind: HorizontalPodAutoscaler

...

Then I run the Helm Chart locally: helm install ./cognitive-services-speech-onpremise-0.1.1/ --name onprem-speech --values myvalues.yaml

As fix suggestion, I recommend updating the apiVersion of each of these files so new users avoid having this problem. Besides that, the textToSpeech configuration is outdated, pointing to the deprecated Microsoft Registry: https://mcr.microsoft.com/en-us/product/azure-cognitive-services/speechservices/text-to-speech/about

The new containers for textToSpeech are now here (Neural Text to Speech): https://mcr.microsoft.com/en-us/product/azure-cognitive-services/speechservices/neural-text-to-speech/about.

Considering this change, I also think the container resources request is outdated, so the Helm Chart should be updated, especifically the file charts/textToSpeech/templates/_image.tpl. For the new Neural Text to Speech, it is recommended having at least 12 gb RAM (https://learn.microsoft.com/en-us/azure/ai-services/speech-service/speech-container-howto#container-requirements-and-recommendations), but in the current configuration, it says the minimum is 2 gb:

Calculate resources for container
*/}}
{{- define "tts.image.resources" -}}
{{- $_ := .  }}
{{- template "tts.CPU.calculate" $_ -}}
resources:
  requests:
    cpu: {{ $_.millicores }}m
    memory: "2Gi"
  limits:
    cpu: {{ $_.millicoresLimit }}m
    memory: "3Gi"
{{- end -}}

Another question would be: the CPU resources calculation still applies or there was some change with the new Neural Text to Speech?

Note: already opened an issue to update text to speech configuration: #116709

Thanks!