Home / server / Questions / 966362
In Process
smartenbergen
Asked: 2019-05-09 03:32:26 +0800 CST

mdadm does not show every device as active

  • 772

Recently, I had to resize a linux software RAID array. It has been a little bit complex as I was forced to do many steps to grow the device size while shrinking the number of members from 14 to 6. It nearly took a week. However, everything went fine and the LVM inside has not been harmed. Now the array seems to be fine, but shows only 4/6 devices as active:

[root@kvm24 ~]# cat /proc/mdstat 
Personalities : [raid10] [raid0] 
md3 : active raid10 sdh3[7] sdn3[6] sdl3[10] sda3[17] sdf3[19] sdc3[18]
      5559542784 blocks super 1.2 128K chunks 2 near-copies [6/4] [UUUUUU]

In the details, I can't see a problem:

[root@kvm24 ~]# mdadm --detail /dev/md3 
/dev/md3:
           Version : 1.2
     Creation Time : Wed Nov  1 23:53:09 2017
        Raid Level : raid10
        Array Size : 5559542784 (5301.99 GiB 5692.97 GB)
     Used Dev Size : 1853180928 (1767.33 GiB 1897.66 GB)
      Raid Devices : 6
     Total Devices : 6
       Persistence : Superblock is persistent

       Update Time : Tue May  7 13:28:06 2019
             State : active 
    Active Devices : 6
   Working Devices : 6
    Failed Devices : 0
     Spare Devices : 0

            Layout : near=2
        Chunk Size : 128K

Consistency Policy : unknown

              Name : kvm24:3  (local to host kvm24)
              UUID : 35833398:1c8291c5:8e817efc:6f99d541
            Events : 582653

    Number   Major   Minor   RaidDevice State
      17       8        3        0      active sync set-A   /dev/sda3
      10       8      179        1      active sync set-B   /dev/sdl3
      18       8       35        2      active sync set-A   /dev/sdc3
       7       8      115        3      active sync set-B   /dev/sdh3
      19       8       83        4      active sync set-A   /dev/sdf3
       6       8      211        5      active sync set-B   /dev/sdn3

To be honest, I didn't even realize that, until Check_MK told me CRIT - disk state is [6/4] [UUUUUU] (expected 4 disks to be up).

What might be the problem with the array?

linux

1 Answers

  1. Everything is alright with your RAID, at least according to these logs. Check dmesg for more details on what happend with your RAID earlier. I don't know what is your Check_Ml, but RAID will re-sync from time to time and it's something normal. If you're refering to RAID state, this could be: clean, but also active or active sync. If you see [UUUUUU], then each U here means that each drive is synchronized. You should be worried if one of them will be replaced with _, meaning that one drive is inactive/failed. You can read about it at https://www.kernel.org/doc/html/v4.16/admin-guide/md.html

    • 0