The issue happened again.
I noticed there could be a link with the fact that the computer is up for days and then at next reboot it triggers a reconstruction.
(the 3 disks composing the raid device are healthy)
The computer is very slow during reconstruction.
sudo dmesg | grep md127:
[ 18.394466] md/raid:md127: not clean -- starting background reconstruction
[ 18.394475] md/raid:md127: device sdb1 operational as raid disk 1
[ 18.394476] md/raid:md127: device sdc1 operational as raid disk 2
[ 18.394476] md/raid:md127: device sda1 operational as raid disk 0
[ 18.394697] md/raid:md127: raid level 5 active with 3 out of 3 devices, algorithm 2
[ 18.394765] md127: bitmap file is out of date (10910 < 10911) -- forcing full recovery
[ 18.394772] md127: bitmap file is out of date, doing full recovery
[ 18.479938] md127: detected capacity change from 0 to 1717716385792
[ 18.485209] md127: AHDI p1 p3
[ 18.485234] md127: p1 size 3214817840 extends beyond EOD, truncated
[ 29.228966] md127: AHDI p1 p3
[ 29.228970] md127: p1 size 3214817840 extends beyond EOD, truncated
[ 29.554661] md: resync of RAID array md127
cat /proc/mdstat
Personalities : [raid6] [raid5] [raid4] [linear] [multipath] [raid0] [raid1] [raid10]
md127 : active raid5 sdb1[1] sdc1[3] sda1[0]
1677457408 blocks super 1.2 level 5, 512k chunk, algorithm 2 [3/3] [UUU]
[================>....] resync = 80.0% (671298352/838728704) finish=22.9min speed=121385K/sec
bitmap: 7/7 pages [28KB], 65536KB chunk
Eric