Encountered exception from SPICEClient: table id 61d2f4fe-8879-4836-aad8-70f8b06d4707 does not exist for customer

I encounter this error when I import a bundle into one database, but not in another. The database structure should be the same.

{
“Status”: 200,
“JobStatus”: “FAILED”,
“Errors”: [
{
“Arn”: “arn:aws:quicksight:eu-west-2:************:dataset/78897c71-ea2a-4859-a562-ceb03a819d9b”,
“Type”: “com.amazonaws.services.quicksight.model.InternalFailureException”,
“Message”: “Encountered exception from SPICEClient: table id 61d2f4fe-8879-4836-aad8-70f8b06d4707 does not exist for customer ************”
}
],

The next time I try, I get this message which is the same but a different table GUID:

{
“Status”: 200,
“JobStatus”: “FAILED_ROLLBACK_COMPLETED”,
“Errors”: [
{
“Arn”: “arn:aws:quicksight:eu-west-2:************:dataset/78897c71-ea2a-4859-a562-ceb03a819d9b”,
“Type”: “com.amazonaws.services.quicksight.model.InternalFailureException”,
“Message”: “Encountered exception from SPICEClient: table id 25620533-d875-466e-8280-e9f3576f9eed does not exist for customer ************”
}
],

In order to not to have to worry about Credentials, my bundle does not include the datasource. That exists in QS already. I simply change the datasource to point to one database or the other.

The GUID being complained about does not exist in any JSON file in the bundle.
If I import the bundle with the datasource pointing at the GOOD database, it all refreshes perfectly.

If I then edit the datasource to point to the BAD database, once again I can refresh everything without error.

Before I do the test imports, I remove all the datasets, dashboards and analyses. The environments is empty.

Anyone got any ideas?
Thanks

Hello @Jeremy_Likezero

What are the two data sources that you’re using, is it the same database just a different connection or different services?

It’s the same connection. I just change the definition of the datasource in the GUI from one database to another. The tables are the same in each database.

Hey @Jeremy_Likezero

My guess is that the GUID you’re seeing has more to do with SPICE logging something for the table. Then when you try to recreate the dataset from a new database it can’t recognize it.

I know that you said they are the same table in each database but is there any chance that columns are named differently, data types are different, or the schema is different in anyway?

Thanks Duncan. I had a 75 minute session with AWS support on Friday, checking that I hadn’t done something stoopid. Apparently the SPICE team have accepted they have an issue and will update when they have a solution. I’ll keep the community posted.

1 Like

16 days after the support call, no feedback of any description from AWS development. I get the feeling this is a “mature” product for AWS and they have no serious development team working on QS any more.


This is the offending dataset. It is 3 wide by 3 deep. That should work, shouldn’t it?

Hello @Jeremy_Likezero

That’s definitely frustrating. I wouldn’t bet that the dimensions of your joins are going to cause a problem.

The only workaround I can think of is to try achieving this using custom SQL and direct query and see if it works. Basically just trying to get around SPICE.

Hey @Jeremy_Likezero

Were you able to try using direct query and custom SQL instead of SPICE?

Hi, no. The idea here is to try to reduce the impact on the database, not increase it. I am thinking that the additional cost of a second, replicated RDS database to handle reporting might justify the additional cost of using Qlik Sense or similar that offers a relational model and GROUPing in the reporting engine!

Hey @Jeremy_Likezero

Have you heard an update from AWS regarding your support ticket and phone call?

If there’s no news I can archive this topic tagged as a bug in SPICE for the time being.

Its acknowledged as a bug with no timescale to resolve.

1 Like