Your observations are interesting

The manufacturer of the SATA RAID engines which you mention, appear to have a German language-only website, so I'm not sure how widely they are selling/supporting their products outside of the German-speaking countries... I don't have time to try translating all their website & PDF docs (my German language skills are rusty

), but I've got a few comments below, based on my experience of similar products.
falther wrote:
According to manufacturer - and my usual own experience - the second drive is created on the fly - even while using the drive as Windows system drive.
Agreed - since it's an external RAID engine, then it should not have any OS-dependency. There are also some HBAs that can be configured to mirror disks in the background, again with no OS interaction (LSI 1068 chipset with certain f/w types, for example).
falther wrote:
The customers drive was Windows 7 - and no way - changed every cable, changed adapters and brands - to get the drive working on the adapters.
This seems to be a bug in the RAID product since you (and the RAID engine manufacturer) have no ability to change Windows 7 and whatever commands it sends to disks - and you know that the customer's disk (without the RAID engine) works fine with whatever SATA command sequences are sent by Win 7. The RAID engine should be able to cope with any series of SATA commands - if it can't, then IMHO that's a bug.
falther wrote:
But: When I started the direct connected hdd - and connected the adapter with previously made clones - I found the partition in the windows drive management - but "not online". I choose the hdd in this windows feature and change to online - and was able to access the drive.
I wonder if that behaviour means that the mirroring process did not finish? Any diagnosis will be much easier if the RAID product you have has some kind of out-of-band access (e.g. serial port, as on RAID products / RAID HBAs that I've worked on), to monitor its progress/status etc., otherwise you can only guess at what is happening inside the RAID engine (or else try to infer more about what is happening, by using a SATA protocol analyser to watch the I/Os being sent to the "new" drive, and ideally, what commands are being received from the host at the same time.)
Personally, I hope this product does have some kind of monitoring/management utility which can run on the OS, for their RAID engine to report issues e.g. SMART failure of a disk, bad blocks, other mirroring failure, and not just an LED on the board, which could be inside the PC - otherwise one disk could be dropped from the mirror (loss of redundancy protection), without the customer/sys admin being aware of that critical situation.
falther wrote:
Question:
[...] What could cause this (strange) different behaviour ?
As I mention above, IMHO this seems to be a bug in the RAID engine product, since your tests appear to show that it cannot cope with the commands or command
sequence sent by Win 7 - but see below for another possibility.
If I was in your situation using Win 7, I would ask the RAID engine manufacturer if this is a known problem - if yes, then that is your answer.

; if no, then perhaps there is something
different about your config (e.g. background process sending "unusual" command to the disk - SMART monitor perhaps?), which would allow normal Win 7 to work OK during manufacturer testing, but where something specific to
your config exposes a bug in the product which the manufacturer hasn't detected.
My thought is: Why is the RAID engine manufacturer
not getting
lots of support calls, if their product is fundamentally incomptible with Win 7 SATA command / command sequences? What is different about your config/testing which would allow the product to work OK for other Win 7 customers?
I'll be interested if you get to the bottom of the issue.
