RAID5 FAIL => After 3 Drives Go Offline Mysteriously at Once

  • Version: 6.9.15-2 (Shaitan)

    Processor: AMD FX(tm)-8350 Eight-Core Processor

    Kernel: Linux 6.1.0-0.deb11.17-amd64

    I don't care about data. That's not my problem. That makes this matter annoying AF, but not a data loss.

    Here's what (I think) happened. At some point, 3 of 5 RAID5 spindles went offline. (Only two were showing a SMART status) I rebooted, twice, and still only 2 or 5 spindles show up. I shut the system down. Checked that all the spindles were connected to SATA/power. All seemed normal. I powered on the system. BINGO, all 5 spindles are back to SMART green.

    BUT, the software RAID is gone. It's (Device: /dev/md1) not listed under the other RAID (Device: /dev/md0) Of course the files system is "missing". Since RAID5 on device md1 is "gone"

    Because it's not there at all it can't be removed, recovered, or deleted. A new RAID can't be created.

    I know that if I erase each of the spindles, one by one, they will become again available to use them create an new RAID5 but not until erasing each individually.

    Is there another path:?:

    Does this syslog reveal what occurred?

  • KM0201

    Approved the thread.
  • The logs will help you pinpoint when things went wrong and likely show one or more disk not available, md raid not available and filesystem(s) not mounted. a Likewise, if notifications for both filesystem and MD RAID events were configured on your OMV NAS. But they will not show the fundamental cause other than disk hardware failure as may also be reflected in SMART data. Drives can drop out of arrays for a variety of reasons: power glitches, controller issues, backplane issues, cabling issues, etc.

    So what next? There are multiple threads on the subject of MD RAID failures and as often as not, you will need to use CLI commands to help analyse the problem and possibly re-assemble and recover your array.

    The common starting point its to check MD RAID status and the disks that make/made up the array:

    Post the output of:

    cat /proc/mdstat

    blkid | grep raid

    mdadm -E /dev/sd[a-z]

Participate now!

Don’t have an account yet? Register yourself now and be a part of our community!