Arch Linux

Guidobelix

Member

From: Italy

Registered: 2013-06-21

Posts: 31

Problems with Raid5

Since a month or so I have set-up a Raid5 on mz Archlinux PC with 4x2TB WD disks.
Everything went smooth for the first week, until a few days ago, when I have got an error message while using rsync to back-up the Raid5 to an external USB disk:

rsync: read errors mapping "/share/Anime/Mazinkaiser/Mazinkaiser DVD1/VIDEO_TS/VTS_01_1.VOB": Input/output error (5)
WARNING: Mazinkaiser/Mazinkaiser DVD1/VIDEO_TS/VTS_01_1.VOB failed verification -- update discarded (will try again).
Mazinkaiser/Mazinkaiser DVD1/VIDEO_TS/VTS_01_1.VOB
rsync: read errors mapping "/share/Anime/Mazinkaiser/Mazinkaiser DVD1/VIDEO_TS/VTS_01_1.VOB": Input/output error (5)
ERROR: Mazinkaiser/Mazinkaiser DVD1/VIDEO_TS/VTS_01_1.VOB failed verification -- update discarded.

The strange thing is that mdadm is not reporting any problem:

└──> mdadm --detail /dev/md0
/dev/md0:
        Version : 1.2
  Creation Time : Sat Jan 25 00:09:35 2014
     Raid Level : raid5
     Array Size : 5859840000 (5588.38 GiB 6000.48 GB)
  Used Dev Size : 1953280000 (1862.79 GiB 2000.16 GB)
   Raid Devices : 4  Total Devices : 4
    Persistence : Superblock is persistent

  Intent Bitmap : Internal

    Update Time : Sun Mar  2 09:20:20 2014
          State : active
 Active Devices : 4
Working Devices : 4
 Failed Devices : 0
  Spare Devices : 0

         Layout : left-symmetric
     Chunk Size : 256K

           Name : MyNAS:0  (local to host MyNAS)
           UUID : 82e58d60:774f61d7:7e816c49:de093881
         Events : 396431

    Number   Major   Minor   RaidDevice State
       0       8       17        0      active sync   /dev/sdb1
       1       8       33        1      active sync   /dev/sdc1
       2       8       49        2      active sync   /dev/sdd1
       4       8       65        3      active sync   /dev/sde1

I have checked all disks with smartctl -t long and I couldn't find any problem.
Then I performed a Raid scrubbing del Raid, but also in this case no errors.
Then I have forced a file system check with e2fsck -f and again no error.

But I still get errors if I try to back-up this file, both with rsync and cp.

What can I do more to diagnose the problem? I am not worried about the file (it is just a copy of a CD of mine that I have put on the Raid to be used with Mediatomb), but I am concerned that there is a major problem waiting for me just behind the corner.....

Thanks a lot to you all for your suggestions!

alphaniner

Member

From: Ancapistan

Registered: 2010-07-12

Posts: 2,810

Re: Problems with Raid5

If you have anything irreplacable on there that's not backed up, do so first. Then make sure the rest of the files are readable and intact. You could kill two birds with one stone here and use md5sum hashes or cmp to compare the originals with the backups. With multiple terabytes of data this will take quite a while, though.

---

But whether the Constitution really be one thing, or another, this much is certain - that it has either authorized such a government as we have had, or has been powerless to prevent it. In either case, it is unfit to exist.
-Lysander Spooner

Guidobelix

Member

From: Italy

Registered: 2013-06-21

Posts: 31

Re: Problems with Raid5

Thanks for your suggestion. I've run cmp comparing the file that is giving me the error with its backup, and indeed I get again an Input/Output error.
So it's seems that the file it is really damaged in some way (I can't access it anyway).
But what can I do to understand if the disk is defective or ready to break down? As I was mentioning I have a full backup of the files, but I'd like to "isolate" the problem, fix it and forget about it.
The main reason for me to set-up a Raid5 was to be more robust against disk problems, but it seems I coudn't reach my objective yet....

alphaniner

Member

From: Ancapistan

Registered: 2010-07-12

Posts: 2,810

Re: Problems with Raid5

You've checked the filesystem, the raid, and the drives. I can't think of anything else that could cause this error. So if all other files are readable and intact it's possible this was just a 'one-off' and nothing is wrong overall. You could check your logs and journal from between the time you put the file on the RAID and the time you first encountered the error. Were you ever able to successfully read the file? If not, it's possible the problem was caused when it was first written and may have nothing to do with your RAID at all.

---

But whether the Constitution really be one thing, or another, this much is certain - that it has either authorized such a government as we have had, or has been powerless to prevent it. In either case, it is unfit to exist.
-Lysander Spooner

Guidobelix

Member

From: Italy

Registered: 2013-06-21

Posts: 31

Re: Problems with Raid5

Thanks again.
I have replaced the file with a back-up copy and I'll monitor the situation to see if the problem comes back.
I hope you are right and it was only a one off!

frostschutz

Member

Registered: 2013-11-15

Posts: 1,421

Re: Problems with Raid5

which kernel are you using? 3.13.0 had a raid5 bug.

also, what's in the dmesg / kernel messages log when this error occured?

Guidobelix

Member

From: Italy

Registered: 2013-06-21

Posts: 31

Re: Problems with Raid5

Thanks for the suggestion; I now have 3.13.3, but I don't remember which kernel I had when I have first created that file. Assuming I had 3.13.0, could it have been the cause, even if I upgrated the kernel in the meantime?
I will also look carefully into the dmesg, hunting for clues

Edit: Looking into dmesg I have noticed that indeed there was a filesystem error on the root filesystem: could it have created the problem, even if root fs is on a different disk?

Last edited by Guidobelix (2014-03-09 20:59:34)