[backend] metadata-grpc-service mysql_real_connect failed (SSL enabled)

[andrijaperovic]

Environment

• How did you deploy Kubeflow Pipelines (KFP)? Kubeflow Pipelines Standalone

• KFP version: 1.7.0-alpha.1

• KFP SDK version: 1.6.2

Steps to reproduce

1. Enable SSL on external MySQL configuration (Azure MySQL Server)
2. Restart metadata-grpc-deployment pod
3. Observe error:

   2021-10-08 19:32:42.461171: F ml_metadata/metadata_store/metadata_store_server_main.cc:226] Non-OK-status: status status: Internal: mysql_real_connect failed: errno: 0, error: MetadataStore cannot be created with the given connection config.
   

Expected result

metadata-grpc-deployment should be initialized (not sure if SSL parameter is expected here in deployment config).

Materials and Reference

Running gcr.io/tfx-oss-public/ml_metadata_store_server:1.0.0.

---

Impacted by this bug? Give it a 👍. We prioritise the issues with the most 👍.

[andrijaperovic]

https://github.com/google/ml-metadata/issues/20

[andrijaperovic]

Tried to pass metadata_store_server_config_file command line flag to /bin/metadata_store_server in text protobuf format, but observed a similar error:

      - args:
        - --grpc_port=8080
        - --enable_database_upgrade=true
        - --metadata_store_server_config_file=/config
        command:
        - /bin/metadata_store_server

Config:

connection_config {
      mysql {
        host: "..."
        port: int
        database: "..."
        user: "..."
        password: "..."
        ssl_options {
          verify_server_cert: true
          capath: "/etc/ssl/certs/my-cert.pem"
        }
      }
    }

Error:

2021-10-11 18:04:01.536992: F ml_metadata/metadata_store/metadata_store_server_main.cc:226] Non-OK-status: status status: Internal: mysql_real_connect failed: errno: 0, error: MetadataStore cannot be created with the given connection config.

Is there any example documentation regarding SSL setup for metadata-grpc as it pertains to kubeflow pipelines?

[zijianjoy]

Hello @andrijaperovic , would you like to check the logs in cloudsqlproxy deployment? It should have more information about why the connection has failed.

[andrijaperovic]

@zijianjoy thank you for following up.

I thought cloudsqlproxy is only applicable for gcp based on the kustomize template:

https://github.com/kubeflow/pipelines/blob/74c7773ca40decfd0d4ed40dc93a6af591bbc190/manifests/kustomize/env/gcp/cloudsql-proxy/cloudsql-proxy-deployment.yaml

We are running in an Azure (AKS) environment, so there is no cloudsqlproxy deployment in our KFP installation.

[zijianjoy]

I am not very familiar with SQL connection in Azure, the article I can find is https://docs.microsoft.com/en-us/azure/azure-sql/database/connectivity-architecture.

@berndverst for more help on Azure.

[andrijaperovic]

@zijianjoy since this issue is due to a sub-component shipped from google/ml-metadata I've filed a corresponding ticket there: https://github.com/google/ml-metadata/issues/130

Have reached out to @BrianSong to see whether the google/ml-metadata ticket can be reopened. After my discussion with @berndverst this issue can be addressed in google/ml-metadata only.

[stale[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[ReggieCarey]

I'm experiencing this problem (or one that looks like it). The logs contain repeated

W0213 18:24:37.372198 3635 metadata_store_service_impl.cc:432] Failed to connect to the database: mysql_real_connect failed: errno: , error: W0213 18:24:37.372260 3637 metadata_store_service_impl.cc:432] Failed to connect to the database: mysql_real_connect failed: errno: , error: W0213 18:24:37.372275 3638 metadata_store_service_impl.cc:432] Failed to connect to the database: mysql_real_connect failed: errno: , error: W0213 18:24:37.372458 3639 metadata_store_service_impl.cc:432] Failed to connect to the database: mysql_real_connect failed: errno: , error: W0213 18:24:37.372793 3621 metadata_store_service_impl.cc:432] Failed to connect to the database: mysql_real_connect failed: errno: , error: W0213 18:24:37.465122 3622 metadata_store_service_impl.cc:432] Failed to connect to the database: mysql_real_connect failed: errno: , error:

[ReggieCarey]

This causes Kubeflow Artifacts to fail: W0308 15:29:32.612677 247 metadata_store_service_impl.cc:432] Failed to connect to the database: mysql_real_connect failed: errno: , error: W0308 15:29:32.613674 245 metadata_store_service_impl.cc:432] Failed to connect to the database: mysql_real_connect failed: errno: , error: W0308 15:29:32.614279 233 metadata_store_service_impl.cc:432] Failed to connect to the database: mysql_real_connect failed: errno: , error: W0308 15:29:32.614387 248 metadata_store_service_impl.cc:432] Failed to connect to the database: mysql_real_connect failed: errno: , error: W0308 15:29:32.614480 236 metadata_store_service_impl.cc:432] Failed to connect to the database: mysql_real_connect failed: errno: , error: W0308 15:29:32.616160 240 metadata_store_service_impl.cc:432] Failed to connect to the database: mysql_real_connect failed: errno: , error: W0308 15:29:32.616427 244 metadata_store_service_impl.cc:432] Failed to connect to the database: mysql_real_connect failed: errno: , error: W0308 15:29:32.616662 249 metadata_store_service_impl.cc:432] Failed to connect to the database: mysql_real_connect failed: errno: , error: W0308 15:29:32.618607 201 metadata_store_service_impl.cc:432] Failed to connect to the database: mysql_real_connect failed: errno: , error: W0308 15:29:32.618873 194 metadata_store_service_impl.cc:432] Failed to connect to the database: mysql_real_connect failed: errno: , error: W0308 15:29:32.625411 211 metadata_store_service_impl.cc:432] Failed to connect to the database: mysql_real_connect failed: errno: , error: W0308 15:29:32.625411 222 metadata_store_service_impl.cc:432] Failed to connect to the database: mysql_real_connect failed: errno: , error: W0308 15:29:32.625516 208 metadata_store_service_impl.cc:432] Failed to connect to the database: mysql_real_connect failed: errno: , error: W0308 15:29:32.626050 228 metadata_store_service_impl.cc:432] Failed to connect to the database: mysql_real_connect failed: errno: , error: W0308 15:29:32.626132 183 metadata_store_service_impl.cc:432] Failed to connect to the database: mysql_real_connect failed: errno: , error: W0308 15:29:32.626195 217 metadata_store_service_impl.cc:432] Failed to connect to the database: mysql_real_connect failed: errno: , error: W0308 15:29:32.626196 215 metadata_store_service_impl.cc:432] Failed to connect to the database: mysql_real_connect failed: errno: , error: W0308 15:29:32.626279 207 metadata_store_service_impl.cc:432] Failed to connect to the database: mysql_real_connect failed: errno: , error: W0308 15:29:32.626308 220 metadata_store_service_impl.cc:432] Failed to connect to the database: mysql_real_connect failed: errno: , error: W0308 15:29:32.626356 221 metadata_store_service_impl.cc:432] Failed to connect to the database: mysql_real_connect failed: errno: , error: W0308 15:29:32.630316 246 metadata_store_service_impl.cc:432] Failed to connect to the database: mysql_real_connect failed: errno: , error: W0308 15:29:32.630416 237 metadata_store_service_impl.cc:432] Failed to connect to the database: mysql_real_connect failed: errno: , error: W0308 15:29:32.631172 232 metadata_store_service_impl.cc:432] Failed to connect to the database: mysql_real_connect failed: errno: , error: W0308 15:29:32.631250 238 metadata_store_service_impl.cc:432] Failed to connect to the database: mysql_real_connect failed: errno: , error: W0308 15:29:32.631335 241 metadata_store_service_impl.cc:432] Failed to connect to the database: mysql_real_connect failed: errno: , error: W0308 15:29:32.631448 242 metadata_store_service_impl.cc:432] Failed to connect to the database: mysql_real_connect failed: errno: , error:

[mohamedFaris47]

I have a similar problem with kubeflow. MySQL pod is not running and so grpc and ml-pipeline pods can't connect to the database and are stuck in CrashLoopBackOff. Did you find a solution?

[dashanji]

I have a similar problem with kubeflow. MySQL pod is not running and so grpc and ml-pipeline pods can't connect to the database and are stuck in CrashLoopBackOff. Did you find a solution?

+1

[mohamedFaris47]

For me, when I ran the kubeflow deployment command again, the issue was fixed. Some time mysql pod gets deleted (did not know why) which caused this problem. So running the deployment command again fixed it

[dashanji]

Hi @mohamedFaris47, Thanks for the kind reply. What's the meaning of run the kubeflow command again? BTW, do you refer to the kubeflow installation doc to deploy the KFP standalone instance?

[mohamedFaris47]

I mean running the deployment command mentioned in the Kubeflow documentation here

[dashanji]

I mean running the deployment command mentioned in the Kubeflow documentation here

I see, but the deployment still failed after running the command. The error mysqld: Can't read dir of '/etc/mysql/conf.d/' (OS errno 13 - Permission denied) was encountered in the mysql pod.

[mohamedFaris47]

I don't know the cause of this permission denial unfortunately. That's different from my case, I had only the mysql pod missing so redeploying got it fixed.

[rimolive]

Is this issue still happening in KFP 2.0.5?

[revit13]

I encounter this error on Ubuntu 20.04.6 LTS when deploying standalone kfp 1.8.5 on a Kind cluster. Any help in resolving it is appreciated

[pod/metadata-grpc-deployment-b45564d7d-rqf46/container] WARNING: Logging before InitGoogleLogging() is written to STDERR
[pod/metadata-grpc-deployment-b45564d7d-rqf46/container] F0331 20:05:33.575326     1 metadata_store_server_main.cc:236] Check failed: absl::OkStatus() == status (OK vs. INTERNAL: mysql_real_connect failed: errno: , error:  [mysql-error-info='']) MetadataStore cannot be created with the given connection config.

[github-actions[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[aaj-synth]

not-stale

[rimolive]

@aaj-synth Please let us know if this issue still occurs on KFP 2.2.0

[github-actions[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[github-actions[bot]]

This issue has been automatically closed because it has not had recent activity. Please comment "/reopen" to reopen it.

[922tech]

/reopen

[google-oss-prow[bot]]

@922tech: You can't reopen an issue/PR unless you authored it or you are a collaborator.

In response to this:

/reopen

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.

[HumairAK]

/reopen

[google-oss-prow[bot]]

@HumairAK: Reopened this issue.

In response to this:

/reopen

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.

[github-actions[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[github-actions[bot]]

This issue has been automatically closed because it has not had recent activity. Please comment "/reopen" to reopen it.