as soon i do the command all led flash
Hmm, to bad - would have been nice - I'm a bit wondering though, the cage is connected via three SATA ports, so it should be distinguishable, but well seems not to be the case.
just thinking how do this people which have a lots of disk to manage?
A lot use some sort of inventory system, especially if there is a higher count of other servers and even other HW (Desktops in offices, switches, ...) involved.
That really can start out as a excel sheet, or a internal wiki. Some monitoring applications also have a inventory feature (e.g., Zabbix has one, not the most straight forward but once one is into it it's OK, IMO) there are also independent server/hardware inventory systems, may be worth it only once one has to much HW to sanely keep an overview with simpler methods.
on the Node where the Disk are, a fild "Bemerkung" to write in - which SATA place it is and would merge to the Serial Number!
In general I agree with you here, they main reason that does not yet exists is rather a technical one, i.e., we do not save state of any disks - we always just scan what's there. So we would need to add some disk-to-notes mapping somewhere. A single metadata file per cluster with a disks with the ID or the UUID as index and then some metadata like a comment could work here.
when i put the disk back it will not join automaticly the zpool?
If the server was offline during this then it should join just fine (somewhat obviously as ZFS did not noticed that it was away), else you may not yet installed the
zfs-zed package which handles those events automatically, it should get pulled in on update since PVE 6.1 though.
Additionally, you can always try to bring up a device online manually by doing:
