Fail event on /dev/md1

[Cédric L.]

Hi,

Je receive a mdadm deamon mail with this text:

"A Fail event had been detected on md device /dev/md1."

And in log file from kernel :

WARNING: Kernel Errors Present
Info fld=0x2acc396, Current sde: sense key Recovered Error...: 1 Time(s)
Info fld=0x49c3ed1, Current sde: sense key Recovered Error...: 1 Time(s)
Info fld=0x5925fb1, Current sde: sense key Recovered Error...: 1 Time(s)
Info fld=0x5c82101, Current sde: sense key Recovered Error...: 1 Time(s)
Info fld=0x6967dd9, Current sde: sense key Recovered Error...: 1 Time(s)
Info fld=0xb0e6ec, Current sde: sense key Medium Error...: 1 Time(s)
end_request: I/O error, dev sde, sector...: 1 Time(s)

What's it? Am I losing the hards disks?

Regards,
C.L.

*********
Bonjour,

Je viens de recevoir du deamon mdadm le message suivant :
"A Fail event had been detected on md device /dev/md1."

et régulièrement, j'ai les messages suivants :

WARNING: Kernel Errors Present
Info fld=0x2acc396, Current sde: sense key Recovered Error...: 1 Time(s)
Info fld=0x49c3ed1, Current sde: sense key Recovered Error...: 1 Time(s)
Info fld=0x5925fb1, Current sde: sense key Recovered Error...: 1 Time(s)
Info fld=0x5c82101, Current sde: sense key Recovered Error...: 1 Time(s)
Info fld=0x6967dd9, Current sde: sense key Recovered Error...: 1 Time(s)
Info fld=0xb0e6ec, Current sde: sense key Medium Error...: 1 Time(s)
end_request: I/O error, dev sde, sector...: 1 Time(s)

Qu'est ce que cela signifie? suis en train de perdre les disques?

Cordialement,
C.L.

[Steven E. Protter]

Shalom,

You definitely have a bad sector and should plan on the replacement of the disk that contains /dev/md1

SEP

[Cédric L.]

Thx,

And for sde??? It's the same thing?

C.L

[Matti_Kurkela]

The md devices are "Multiple Disks", i.e. software RAID devices. They act as containers for the physical disks in RAID or multipath configurations.

For example, if your /dev/md1 is configured to RAID1 (=mirroring) mode, you'll find the /dev/md1 device "contains" /dev/sde and at least one other /dev/sd* disk device.

The simplest way to view your /dev/md* configuration is to run "cat /dev/mdstat".

The kernel shows error messages because the actual disk device /dev/sde is failing. The mdadm daemon has noticed the failing /dev/sde belongs to /dev/md1.

In RAID1 or RAID5 configurations, the warning from mdadm is useful because a single failing disk does not cause any immediate problems. But if a second disk from the same RAID set fails, you will lose data: in this case, the message from mdadm means you're no longer protected against another disk failure.

So you should replace the failing /dev/sde and then re-sync the RAID set. If you don't know how to do this, google for Linux Software-RAID HOWTO.

If you have a RAID0 configuration (to increase performance, not reliability), you may have already lost some data. If you don't have your backups up to date, back up anything important on the disks *now*!!!

MK