Hi - Ive setup a pipeline to inject Postgres / Cloudsql tables using Datastream CDC. This moves the data to GCS using the ‘Datastream to Bigquery’ template. Them using Dataflow and Pubsub to reflect the changes in to Bigquery.
Generally I can see that the full backfills have worked as well as changed data is being captured and replicated in BQ. However this is not consistent as I can see some table changes have not yet been reflected into GCS or BQ since days.
considering that these changes have not reflected in GCS, I suspect the issue to be with Datastream or the inputs coming from the logs of the source DB itself.
The datastream job is running without errors and there is nothing unusual showing in the error logs too.
Any help and guidance will be appreciated.
@s_dmello
You can report this issue to the Datastream team here [ https://issuetracker.google.com/issues ]
Component : Public Trackers > Cloud Platform > Storage and Databases > Datastream
Thanks, I have done so.
Hi, I am facing a similar problem. Everything worked perfectly at setup but now I see missing rows, tables/columns disappearing, etc. Were you able to find a resolution?