Hello,
TimeXtender 20.10.31.64
ODX: 20.10.31
We are currently running a database on hyperscale with 12vcores and premium memory optimized hardware. ODX is a data lake.
We are running execution packages on 4 threads and we want to reduce our odx transfer times as that is out biggest impact right now.
This is an example of a load.
![](https://uploads-eu-west-1.insided.com/timextender-en/attachment/a8563813-0365-456f-abd3-e116ae18e2b5.png)
I am seeing long ODX transfer times for very small tables that I cant find the cause of or explain.
We have alot of small tables (see example below) that is taking 1-3 minutes, while some tables doing data cleansing on 1million + rows is doing it in the same amount of time.
See this small table that is 400 rows, no transformations and only 10 columns.
![](https://uploads-eu-west-1.insided.com/timextender-en/attachment/3f7a2544-49ec-4d0b-a014-d9ec83d808d9.png)
I can see we are peaking log io at 100% during certain periods.
However, that seems to be during the later parts of the loads and not during odx transfer in the start.
![](https://uploads-eu-west-1.insided.com/timextender-en/attachment/0fb835c6-b59c-4b4c-acfb-f672c843039c.png)
Is there any thing that can cause this or possibility to improve it?
Thank you,
Victor
Best answer by rory.smith
View original