Thanks for your suggestions @dpfaffenbauer
6 hours have been the estimated time, in the end it took 25 hrs.
What was most interesting about it is that the import bacame very slow in the end. For example the last 5 % took about 5 hrs. (While the import was running the calculation has always been correct 50 % = 215k articles imported and so on). I think the Bottleneck has been the SSD Controller because the partition got only 5GB left and as i do remember the most ssds are going to be kinda slow when they run out of space.
Aren’t the imports parallized ? I was sure that the process-manager does parallizing because all my six cores were acting at about 30 - 50 % as the import has been active which was different to the pimcore build-in import where only one core was active.
Nevertheless I think I am not capable of writing the routines to write those routines
But maybe someone has some experience with different hardware stacks and could tell me how, for example, a ryzen 9 3950X + SSD Raid system would impact the performance.
To be honest atm I don’t even know if i do really need the performance. But before using something in productive setups I want to have the best performance and stability possible as also a tested backup setup especially because it’s causing a ton of problems if you have to change something fundamental in a running environment…
Btw.: The performance after the import has finished is really impressive. Filtering 10k articles of a specific manufacturer is done in like real-time.