We have moved from 1.58 to 1.75 just recently on AWS using RedShift DB. In a random pattern, different scheduled jobs where showing "Validation Failure. Input components are not in a valid state" error for "rewrite table" object and "Rename" objects,

Matillion ETL
1

Running manually, the specific object will work, as well as the job, as well as another occasion of the scheduler.

As these are Transformation jobs, no option for "Retry".

2

Hi I have the same issue. We moved to 1.75.2 on AWS using Redshfit DB and scheduled jobs are not in a valid state and the job breaks due to input column not being in a valid state. Anyone knows how to fix this?

3

Hi Michael,

 

Matillion support approached us and suggested add the below in the environment settings of each project:

Advanced Connection Settings:

tcpKeepAlive: true, socketTimeout: 0

 

And also change concurrent connections from 1 to 2 or even 4.

 

Hope it helps you.

4

Hi,

 

thank you very much for your help. I will try that out. What is weird is that when I run the process in the afternoon everything works fine. Under heavy load in the mornings it seems to break, but I guess the settings you meantioned will help.

 

Thanks again.

5

Hi again,

Also in our situation, it also worked as a stand alone, but while scheduled under heavy load, I assume RedShift was too busy to answer any request right away.

6

Hi Michael - Did the update to connection setting resolve the issue ?