Skip to main content

Maybe an unusual question…

Problem → 
We want to upgrade our Software Factory to 2025.2. Nice and all, but we have a Software Factory which is 335 GB. Many tables contain irrelevant data of some years ago (old branches). If I am correct, in 2025.2 data will be cleaned via a scheduled flow of Thinkwise (the scheduled flow existed already but contained a bug which got fixed). Manually cleaning data took me already alot of evenings and is still not done... (I did this for a few weeks… Call me crazy 😉 ). I am a bit worried this schedule will slow our server/environment down because cleaning will not be done in one evening...

Another problem: running tsf_optimize on our Software Factory results into our server running out of resources which also sounds problematic.

Possible solution → 
We wanted to start with a clean Software Factory, and import our model into it. This way we do not have to clean data, and our Software Factory database will be small once again. We DO want to keep important history like code changes and release branches & released model versions. 

While we are at it, starting again with a clean Wagenborg_IAM and IAM_Software_factory might also be wise?

Is this something you would advise? Or is this an approach you would not want us to do? Are we forgetting something? What do you think?

Yesterday I spoke with ​@Ricardo about our SF database which is almost 265gb. Quite large, and also in need of maintenance. I also found some very, very old records… 

Currently on 25.1 and very interested in how this can be optimized/cleaned.

Blommetje

 

Hi Marijn,

We are currently working on a feature for this that will allow developers to clean their Software Factory database of old data. We indeed already have the task to clean-up history data for removed and archived branches. We are expanding this functionality to also be able to clean history data for active branches, taking into account the model versions of the model.

In essence, we want to give developers the ability to clean up branches and remove history data that is no longer relevant for development. For instance, you can clean up all records prior to 2024-01-01 stored in history tables. We are in the midst of working on it. However, I cannot say for sure it will make the 2025.3 release as we need to test this thoroughly.

We could opt to place Retention periods on the Software Factory tables, but the downside is that history data will be removed automatically with no way of preventing that. For most customers, this will not be any issue, but for others it may be. So defining a proper generic retention period for everyone is a tough call to make.

Reply


Terms & ConditionsCookie settings