Unable to access iceberg data created through AWS glue and stored in s3

Dremio Cloud
1

Hi, I’ve had this error that’ been bogging me for some time! Would appreciate any help / guidance.

error I get:
Something went wrong. Please check the log file for details, see https://docs.dremio.com/advanced-administration/log-files.html.Show more

on server.log:
nable to get table: java.lang.NullPointerException
at com.amazonaws.glue.catalog.metastore.GlueMetastoreClientDelegate.getTable(GlueMetastoreClientDelegate.java:435)
at com.amazonaws.glue.catalog.metastore.AWSCatalogMetastoreClient.getTable(AWSCatalogMetastoreClient.java:1023)
at sun.reflect.GeneratedMethodAccessor490.invoke(Unknown Source)
…

2023-08-31 10:45:36,463 [qtp449668621-1416] WARN c.d.e.store.hive.HiveStoragePlugin - Plugin ‘titanic_s3’, database ‘titanic_iceberg’, table ‘funds’, DatasetHandle empty, table not found.

context:
We have an iceberg Datalake stored on S3 and being written / read from Glue. I am exploring the possibility to read and more importantly write on the tables created with Glue through Dremio.
on the dremio GUI, I connected to the dataset through Glue Catalog (dremio is deployed on the same account as the glue context).
Then I clicked on the table I want to interact with in dremio. And I get the error mentioned above.

Looking forward to any reply!
Thanks.

2

@anechi Happy to Help ! WOndering if you would be able to upload the below 2

  • Job profile of the failed job
  • server.log from the coordinator when this error happened

Thanks
Bali

4

Hi Balaji,

  1. I am not sure how to generate the job profile, the interface is not loading because of the error I reported.
  2. I will join only 10 relevant lines of the server.log because I keep getting blocked by the moderator for trying to copy paste more.

Something went wrong. Please check the log file for details, see https://docs.dremio.com/advanced-administration/log-files.html.Show more

image
image1809×894 37.2 KB

server.log:

2023-09-01 14:23:56,834 [qtp449668621-1705] WARN  c.d.e.store.hive.HiveStoragePlugin - Plugin 'titanic_s3', database 'titanic_iceberg', table 'funds', DatasetHandleempty, table not found.
2023-09-01 14:27:27,666 [scheduler-16] INFO  c.d.e.w.p.ActiveQueryListService - Starting activeQueryListTask on this coordinator.
2023-09-01 14:27:27,667 [FABRIC-rpc-event-queue] INFO  com.dremio.sabot.exec.MaestroProxy - There are no active queries on this executor. So nothing to reconcile.

2023-09-01 14:23:56,834 [qtp449668621-1705] INFO  c.d.exec.store.hive.HiveClientImpl - Failure while trying to read table 'funds' from db 'titanic_iceberg'
org.apache.hadoop.hive.metastore.api.MetaException: Unable to get table: java.lang.NullPointerException

Thanks, Aziz.

5

I am attaching a simple screen recording of how to download job profile

download-profile.mov.zip (3.2 MB)

6

Thank you for the small demio. Here’s the profile.

5178a8b0-1e43-4e9f-a1ea-dc2ac1419c39.zip (7.1 KB)

7

@anechi I see you are on 20.x which was one of the earlier versions to support Iceberg and by default is false. I see you have made it truem Are you able to upgrade to a higher version like 23 or 24 and retry the query. Even if it fails again but since the feature is on by default and code has changed, it is easier to troubleshoot

8

Issue was solved by upgrading to dremio 24.0
Thanks a lot @balaji.ramaswamy !

9

@anechi Thanks for the update, glad you are unblocked now

1 Like