Incompatible state detected .. newBucketInterval, existingBucketInterval


#1

I’m loading data using SourceFile.processAll(‘status==“initial”’, async: true)

As I monitor the invalidation queues: c3Grid(InvalidationQueue.countAll())

I’m seeing that some tasks are being marked as “failed” on the DataLoadQueue row

Using c3QErrs(DataLoadQueue, 100), I see:

Incompatible state detected during indexing for type=... ,pk=... ,newBucketInterval=DAY,existingBucketInterval=MONTH

What does this mean, and what should I do next? How can I fix this?


#2

I have observed this if the data points is loaded before the time series header. One way to solve this would be to serialize the operation. You should also raise a ticket for this to be natively handled in the data load engine


#3

Yeah, serializing won’t work for ongoing data loads. This violates our guiding principle of being insensitive to sequence of data loads. We should file a ticket.