ok i figured it out, my write logs within pool-0 were configured wrong.
they need to be:
+
--sda
--sdb (remote)
+
--sdc
--sdd (remote)

during the problem they were about
+
--sda
--sdb
+
--sdc (remote)
--sdd (remote)

the "bad change" was during replacement of hard drives. Now failover and moving nodes is working correctly.

BTW, this forum should be improved a lot. Dont you trust your own users or why upload of images is disabled/not possible, linking to outside images is made difficult, writing forum posts is only allowed if administrators approve the post?