We’re have some performance issues with the server (getting clock drifts and errors writing to the internal database).
The number of threads writing to config.idb-journal file is also very high. Does anyone know what can cause this? Nobody is currently editing on that project, so I’d guess the traffic should be rather low.
Apparently it started working by itself.
I guess the VM Host just had issues with other VMs which caused slow writes to the disk.
I’m still wondering what’s getten written to the internal DB though. I though it was mostly configuration (windows, tags, …) so didn’t have a lot of write traffic when in production.
A while ago i had massive problems with the internal DB performance. It turned out that every value change of an expression or query tag is written to the internal DB. Im my case it was a simple query tag with a rather large result dataset that caused an overload.
Well, all tags. That's how they retain their values through a shutdown or crash. Any large chunks of data that don't need to be retained through shutdown should be handled entirely in scripting and/or client tags/session properties.
That begs the question what the 'Store values between edit and restarts' checkbox in the tag config is for 
I was under the (wrong) impression that values are not stored when that box is not checked.
