Cosmos DB Mirroring Status Issue - "Running with warnings"

[dmathewg]

Trying to setup mirroring on my Azure Cosmos DB and running into the status issue of "Running with warnings" (Internal system error occurred. ArtifactId: database ID).

I have setup the Azure Cosmos DB (API for NoSQL) with the proper networking, backup and in region East US. Containers are built and have a small amount of data in them. All working great on Azure Portal side.

Set up the Fabric Cosmos DB mirror Azure Cosmos DB v2 connection with proper URI and Key. Connection works and can see tables in SQL analytics endpoint. Replication setup completed but receiving the error noted above on each of my tables. No data is being mirrored as tables are empty.

Microsoft states on their troubleshooting page that "Replication could be working and you're observing a transient error if Azure Cosmos DB is throttling requests from Fabric."

Is there a resolution for this issue or a way to determine the underlying transient error that Microsoft is noting?

[Anonymous]

Hi @dmathewg 

Sorry I don't find any detailed introductions about troubleshooting the status "Running with warnings". Usually we can ignore transient error messages safely. I tried mirroring Azure Cosmos DB Free with a sample database without experiencing any failure. I'm using Fabric in UK South data region. 

Did you try refreshing the Monitor replication pane? Does it still show the same "Running with warnings." status today? 

You said you can see tables in SQL analytics endpoint. Do you mean that you can see data from the SQL analytics endpoint of the mirrored database? Do those Azure Cosmos DB tables have data or are they empty? 

If the mirrored database still doesn't work normally, you may try stop replication and start replication again. Or try creating a new mirrored database. 

Best Regards,
Jing
Community Support Team

[dmathewg]

Thanks for responding Jing.

I had created 2 separate instances of a Azure Cosmos DB accounts per Microsoft setup, Fabric connections using the v2 connector, and Cosmos DB mirrors. I refreshed many times yesterday getting the same result.

This morning, I deleted both Cosmos DB mirrors and recreated just one of them.

When I try to access the mirror monitoring this morning, Fabric is hanging. I am getting the Fabric icon and it won't advance further. This is occurring in two separate workspaces. It is also happening in my SQL mirror that I set up months ago and has been running flawlessly until now.

When I mentioned I could see the tables in the mirror yesterday, I was able to see the tables in the SQL analytics endpoint but the only column that appeared in the table schema was _rid and no data was present in any of the tables.

I had stopped and started replication yesterday many times. I also tried to bump up my Fabric capacity to F4 from F2 to see if that would make a difference and that didn't help either.

Thanks for any help you can provide.

-Darren

[dmathewg]

Just after responding, my workspaces seem to be functioning normally again but the issue with my Cosmos DB mirror still exists. Here are a couple screenshots of the monitoring and SQL analytics endpoint showing the Model Layouts...

Monitoring

SQL analytics endpoint Model layouts

[earnie · Accepted solution]

We released a fix for this edge case four weeks ago, and it has been rolled out to all production regions except West Europe as of October 20th. The West Europe deployment is delayed and the Release Management team is currently working to address the deployment issues.

[jcarville]

I have a trial fabric capacity in North Europe and this issue is still occuring. Is there any more information on it or a possible resolution?

[dmathewg]

Thanks earnie!

It's been a pleasure working with your group in getting this resolved.