Warehouse Errors

This is a permissioning issue. To learn how to set up proper permissions, see the Postgres and Redshift guides.

This error occurs when your cluster is resizing. The warehouse continues on its scheduled run interval. Once the resize completes, Segment loads all data from the failed run.

This is a Redshift 500 - an internal server error. This is usually caused by having too many tables or too many columns. See Column limits for the maximum columns per table. If you're seeing this error, contact the Segment Support team.

This is a networking error that typically arises when Redshift doesn't close properly close the connection or gets rebooted.

If you see this error on consecutive syncs, contact the Segment Support team.

This is a permissioning issue. To learn how to set up proper permissions, see the Postgres and Redshift guides.

This is a credential issue. To fix your credentials, head over to Warehouse > Settings > Connection.

This is a credential issue. To fix your credentials, head over to Warehouse > Settings > Connection.

This is a networking error. The connection times out because Redshift doesn't close properly or gets rebooted.

If you see this error on consecutive syncs, contact the Segment Support team.

This occurs when a Postgres database is incorrectly connected as Redshift. To resolve this, delete the warehouse and reconnect, using the Postgres set up option.

This error is generally a network error when Redshift closes the connection. If the problem persists on multiple runs, contact the Segment Support team.

This is a permissioning issue. To learn how to set up proper permissions, see the Postgres and Redshift guides.

This error is generally a network error when Redshift closes the connection. If the problem persists on multiple runs, contact the Segment Support team.

This error indicates that a column that is attempting to sync has the same title as a reserved keyword in Snowflake. More information regarding Snowflake's reserved keywords can be found in Snowflake's docs.

Most warehouses limit the number of columns that a table can have or that a single load can write. Segment creates a column for each property in a collection, so a collection with many unique or dynamically generated property names can exceed the limit. When a collection exceeds the limit, Segment can't load data into it and the sync reports an error for that collection.

Properties you disable with Selective Sync aren't loaded, so disabling unused properties reduces the number of columns in the collection.

Snowflake doesn't have a hard limit on the number of columns in a table, but it does limit the number of columns a single load can write. Segment loads data with Snowflake's COPY INTO command, which can write at most 4,096 columns in a single query. Segment loads every enabled column in a collection on each sync, whether or not the events in that sync include it, so the number of enabled columns in the collection is what counts. When a collection has more than 4,096 enabled columns, the sync fails with this error:

SQL compilation error: column '$4097' does not exist

Snowflake can also reject new columns with this error:

Operation failed because soft limit on objects of type 'Column' per table was exceeded.

To resolve either error, reduce the number of columns in the collection. See Reduce the number of columns.

To bring a collection under its limit:

• Find properties with dynamically generated names, such as IDs or timestamps used as property keys, and send those values as property values instead.
• Disable properties you don't need in the warehouse with Selective Sync.

If you still see the error after reducing the columns, contact the Segment Support team.