Am 14.05.19 um 20:53 schrieb Eric Valette: > On 14/05/2019 20:38, Reindl Harald wrote: >> >> >> Am 14.05.19 um 20:33 schrieb Eric Valette: >>> Fine. Again where is it documented? The documentation the contrary. So >>> go and fix the doc instead of ranting again end user. >> >> you better cool down given that *i am* an enduser and when you think you >> reach anything by piss off ousers hwich show you how your setup should >> look like you are likely wrong > I only say two things: > > 1) you were not helping me on my important problem which was > restoring the array and recovering my data, just blaming a config file > content fine, if it's no help for you get as much as possible informations as possible.... > 2) explaining here by blaming someone who carefully followed the > wiki,faq and mdadm readme how RAID should be created and mdmadm config > file should be written is probably not efficient. Update the docs luke, frankly when you are not interested in how setups look which had never problems with dying disks to improve your setup it#s fine for me > BTW it appears, device naming was not the real problem in the end rather > that if a device disappears after a reboot, you have nothing special to > do except reassembling the array and force rebuilding the needed spare > if you have sufficient spare. I just expected it to be automatic. and that's what i don't understand and likely is because of a non perfect and not well tested setup, mine boots as long as there are two disks holding both mirrors all the time didn't you test how your setup behaves when you plug a drive and how to restore it *before* put data on it - that may sound arrogant but any redundant setup or backup is worthless until it#s tested against the expected points of failure
