Upgrade from Sonacube 9.4 to 9.9 takes too long because live_measured table size is too large

Make sure to tell us:

  • What version are you upgrading from?
    from 9.4 to 9.9

  • System information (Operating system, Java version, Database provider/version)

  • What’s the issue you’re facing?
    Database migration has not shut down since the 9.4.0 → 9.9.2 server upgrade, so I checked and found that there was too much data in the live_measures table.
    Currently, the live_measures table size is 106G, and there are more than 200 million rows. I need a guide on how to reduce this and reduce upgrade time.

Hi!

How much is “too long”? 106 Gb is not too much. It should be done in less that one day.

It seems you are opening several posts for the same question:

I suggest you to close one and keep all the info in the same post.

Best regards.

1 Like

System Information:
linux/amd64, doker container image sonarqube:9.4.0-enterprise, AWS rds postgresql 11.9

We don’t know exactly how long it takes to migrate at the request of the customer. But we used the term “long time.”
On the customer’s side, the migration didn’t actually stop, it continued.
Upgrade uses docker image to upgrade sonarqube:10.2.1-enterprise from sonarqube:9.2-enterprise.

When I checked the table data, it seems that the data remains in live_measures even though I deleted the project. Please tell me how to check if it is deleted together when deleting the project.

I also want to ask you a question about housekeeping. (I will share the settings as an attachment)

  1. The value “Delete all analysts after” is default 5 years. If I adjust the value short, will live_measures data be deleted together?
  2. If it is correct to delete, I wonder how the dbCleaner operation is triggered when this value is adjusted. I wonder if the cleansing operation is carried out as soon as the value is changed or if it can be triggered manually.
  3. Since the data size is large, I need to check it because I think it will affect if the cleansing is carried out during the operating hours.

Hi, Antonio Calero,
I checked the data using the db query statement.
SELECT * FROM sonarschema.projects;
SELECT distinct uuid, project_uuid from live_measures;

As a result, we also checked the data of the deleted project. Is it OK to delete the data of the deleted project?
In the future, when deleting the project, we will have to take measures to delete the live_measures as well. Do you happen to have any related settings? I want to get an answer.

And DB cleaner said that it will be triggered on analysis, but the project has already been erased. Please guide me on how to delete it. Is it okay to erase it manually?
At worst, I think I have to erase the data on the table myself, but when I manually erase the data, it’s hard to see if there are any other errors. Is there a guide in that area?