在 2024/1/25 22:49, Gandalf Corvotempesta 写道:
Ok, i think i did a mess.
I have a 3 way mirror with 3 drives on it
All drives are working, not failed.
I would like to replace one of these drive (because it's very old)
with a new drive.
I would like to to the add and replace in the same phase, without
reducing the redundancy level
I did this multiple times before, during the sync, the array is a 4 way mirror
BUT this time i think i did an issue, i've just run this:
$ mdadm /dev/md0 --add /dev/sdd1 --replace /dev/sda1 --with /dev/sdd1
It is right.
sda was marked failed immediately, and sdd1 is marked as active sync.
Less than 1 second.
Is this ok or i've just added a blank drive and forced the removal of
a working one ?
Honestly, and this is something that comes to mind right now, the
partition is very very small, like 64MB so 1 second sync (on a SSD)
could be ok.
but:
1. which is the right command to run to add-sync-remove a not-failed disk > 2. how can I ensure that sdd is working as expected and data were
synced properly?
you can check dmesg. The following means the beginning and ending of
recovery:
md: recovery of RAID array md0
md: md0: recovery done.
Also, you can get more information with 'mdadm -D /dev/md0'
syncing:
Number Major Minor RaidDevice State
0 8 0 0 active sync /dev/sda
3 8 48 0 spare rebuilding /dev/sdd
1 8 16 1 active sync /dev/sdb
2 8 32 2 active sync /dev/sdc
sync success:
Number Major Minor RaidDevice State
3 8 48 0 active sync /dev/sdd
1 8 16 1 active sync /dev/sdb
2 8 32 2 active sync /dev/sdc
0 8 0 - faulty /dev/sda
--
Thanks,
Nan
