[MCOL-1711] Poor performance of the adapter in case of bulk data loading Created: 2018-09-11  Updated: 2023-10-26  Resolved: 2018-09-12

Status: Closed
Project: MariaDB ColumnStore
Component/s: None
Affects Version/s: 1.1.6
Fix Version/s: Icebox

Type: Bug Priority: Major
Reporter: Elena Kotsinova (Inactive) Assignee: markus makela
Resolution: Not a Bug Votes: 0
Labels: None


 Description   

Configuration:
TX -> MaxScale -> ColumnStore

Use case:
Data was loaded from file with load command in TX.

Result:
CDC adapter starts inserting data row by row.
After 134970 records have been loaded in TX, they are inserted into CS with speed of half second per record.
This is around 18 hours inserting saga.

So slow speed is not acceptable for daily operations in this configuration.
Even these are regular inserts the adapter performs with the same average speed of half second per record.



 Comments   
Comment by markus makela [ 2018-09-11 ]

What was the exact command that was used? It's possible that it used the default of one row per bulk insert.

Comment by Elena Kotsinova (Inactive) [ 2018-09-12 ]

Yes, you are correct. I missed the -r option.

Generated at Thu Feb 08 02:30:48 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.