[bitnami/spring-cloud-dataflow] libsnappyjava.so not found in prometheus-rsocket-proxy

[vai-thnas]

Name and Version

bitnami/spring-cloud-dataflow 34.0.0

What architecture are you using?

amd64

What steps will reproduce the bug?

1. On an AWS EKS cluster
2. Running with Kubernetes v1.30, AL2 nodes 5.10.230-223.885.amzn2.x86_64
3. Deploy with

helm upgrade "dataflow" "artifactory/spring-cloud-dataflow" \
  -f "helm/values.yaml" \
  --install \
  --atomic \
  --namespace dataflow \
  --create-namespace \
  --version "34.0.0" \
  --wait \
  --timeout=10m0s

4. Functionally everything seems to work and the metrics can be scraped, but the prometheus-rsocket-proxy logs are filled with errors

Are you using any custom parameters or values?

global:
  imageRegistry: "$DOCKER_IMAGE_REGISTRY"
  imagePullSecrets:
    - "$DOCKER_IMAGE_PULL_SECRET_NAME"
server:
  replicaCount: 1
  configuration:
    batchEnabled: true
    streamingEnabled: false
mariadb:
  enabled: false
skipper:
  enabled: false
rabbitmq:
  enabled: false
externalDatabase:
  dataflow:
    url: "$POSTGRES_URL"
    user: "$POSTGRES_USER"
    password: "$POSTGRES_PASSWORD"
  hibernateDialect: org.hibernate.dialect.PostgreSQLDialect
metrics:
  enabled: true
  serviceMonitor:
    enabled: true
  readinessProbe:
    initialDelaySeconds: 60
  livenessProbe:
    initialDelaySeconds: 60

What is the expected behavior?

No errors in the logs

What do you see instead?

In the bitnami/prometheus-rsocket-proxy:1.5.3-debian-12-r31 container's logs

1. After container starts, an error logs complains that /tmp/snappy-1.1.10-b15b886e-73d2-47ba-ba4e-cc9fe5941d72-libsnappyjava.so cannot be found
2. Every 30 seconds, 2 stack traces (one at WARN one at ERROR level) are complaining that org.xerial.snappy.Snappy cannot be initialized

Logs dump : prometheus-proxy.log

Additional information

tried overriding prometheus-rsocket-proxy to latest version 1.5.3-debian-12-r34 but same errors

[dgomezleon]

Hi @vai-thnas ,

Could you please confirm if you face the same issue using the latest version 34.1.1 and the values below? (the only change is the externaldatabase):

server:
  replicaCount: 1
  configuration:
    batchEnabled: true
    streamingEnabled: false
mariadb:
  enabled: false
postgresql:
  enabled: true
skipper:
  enabled: false
rabbitmq:
  enabled: false
metrics:
  enabled: true
  serviceMonitor:
    enabled: true
  readinessProbe:
    initialDelaySeconds: 60
  livenessProbe:
    initialDelaySeconds: 60

If so, please confirm if it happens just after starting the chart.

[vai-thnas]

Hi @dgomezleon and thanks for your support,

• first I upgraded to 34.1.1 while keeping the above configuration, the same errors appear
• then I used the config you provided, and this time we have a Flyway initialization error on the SCDF server side: server.log
• lastly I tried to switch to mariadb as I suspected embedded postgresql might not be supported (at least I couldn't find the "postgres" properties group in the README), and the same errors could be seen in the logs, even after a manual restart of the prometheus-proxy pod

[dgomezleon]

Hi @vai-thnas

Thanks for the feedback. I have just created an internal task to solve this issue.

[github-actions]

This Issue has been automatically marked as "stale" because it has not had recent activity (for 15 days). It will be closed if no further activity occurs. Thanks for the feedback.

[vai-thnas]

hi @dgomezleon, is it possible to bind this issue's lifecycle to the internal task you created so that this one will be kept open until the other is resolved ?