Rapid Database Growth - Rancher HA

I have Rancher setup in HA mode, as per guidelines.
Everything working well, but seeing database growth, more specifically the process_execution table growing by at least 1GB per day.

Changed My Loadbalancer to the Rancher server from being full round robin, to keeping one Rancher server in backup mode, thinking this may reduce the database growth, but alas.

I have been able to truncate the table now several times, without any obvious harm, but its a concern at the rapid rate of growth.

Any idea’s as to why, and suggestions to reign or “corral” this in?

Yep, this is a known issue. Honestly just been lazy about this one. We can get this fixed in the next release. The process_instance and process_execution table should be cleaned up because they can grow at a dramatic rate. If those tables are growing it is because Rancher is actively orchestrating things. It has nothing to do with the load on the UI or API, it all about how many things are changing. For most users they don’t change enough to notice it grow quite a bit. I wonder if rancher is processing things you aren’t aware of. You can get an idea of active tasks running by doing

select * from process_instance where end_time is null

We will be building out more administrative functions in Rancher that will give you more visibility to what is going on inside of Rancher.

Thanks for the response.

Well I gather Rancher is doing a lot, since I have currently about 25 Dev Rancher Hosts across 10 Dev environments
All relatively active being worked on moving current applications across to Docker platform, not including, Redis,Zookeeper and Mesos cluster automation being designed.

All the entries I inspected seemed legit, but yes, alarming rate of growth and only getting larger and faster the more server and projects being added.

Indication of time frame looking at for more admin features?

BTW…thanks for awesome product, its will be a central feature for us moving forward.

Would it be safe to delete the rows where end_time is not null?

@willseward yes, that is safe for the process_instance and process_execution table.