Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[THN]

Hello,

by accident I notice in the status tab that "HARD DRIVE INFO" reports 2 drives in RAID 1 with Sync Time Remaining: Completed, but further down I get only the Right HDD reported (SAMSUNG HD204UI) and no left HDD. 

Is there a reason for worries?

I have fun_plug installed, so how would I make the left drive appear, or else further diagnose and repair any possible issues? 

Thanks a lot in advance for your help.

THN

[fordem]

Yes - it is cause for concern.

As far as I know fun_plug is not supported by DLink so any questions concerning it's use should be raised with the developer of the fun_plug you are using.

[THN]

Without considering fun_plug, how would I further analyze and repair the situation?

THN

[fordem]

It's really not that easy - considering that with a fun_plug you may have the tools required to make a thorough mess of the RAID1 subsystem - there is really no way to determine the interactions of software from an unknown source on the system, which is one of the reasons that manufacturers don't support third party software.

But - since you ask - let's have a shot at it.

I no longer run fun_plug on my NAS - if I were in the situation you are in, I would try to force a drive failure on the bay that is being shown as empty, and once the unit sees that failure, it should allow a rebuild.

By the way - you follow the procedure outlined at your own risk, I am not responsible for any data loss or equipment damage that may occur - having said that - I am not suggesting you do anything that I have not done personally at some time in the past.

1 - BACKUP the data.
2 - Remove the unit front panel and hot unplug the drive that is not shown in the status.  The unit will hopefully sense the drive removal and give you a failed drive message and a degraded status.
3 - Test the removed drive in a PC - if it passes, delete all partitions - if it fails, get a replacement.
4 - Power off the NAS and install the clean/replacement drive.
5 - Power on the NAS, login and follow the prompts to rebuild the array.

If the unit does not give a degraded array message when the drive is removed, you'll need to reformat the disks and restore the data from the backup that you made.

[THN]

Hi fordem,

thank you very much, your procedure is fully clear and understood. In fact, you suggest to "break" the RAID forcefully, reformat/repartition the invisible disk and then initiate a rebuild through the NAS' internal OS. I agree that that should be a straight forward option.

It just comes to my mind that I have several free hotswap bays available in my PC allowing me to mount either or both drives locally. I have several LINUX OS available to boot from (*buntu, OpenSUSE, MINT, Debian etc.).

Could you think of a way to diagnose, and if needed more "gently" repair the RAID in such a local configuration, that would later on also be accepted by the NAS?

Thank you in advance.

THN

[THN]

Some interesting news, I hope:

In another menu (Tools -> RAID), I discover that another SAMSUNG HD204UI drive is reported here; it has a serial number different from the visible one I saw in the earlier screen, so I suppose it's the invisible drive in the left bay of the NAS.

This other drive is reported with a value of "0GB", the check box "Enable Auto-Rebuild" is checked, and just above I see "Current RAID Type" is RAID 1.

Now what does that mean:

• this other drive reported here is well part of the RAID mirror, is fully used, leaving no free space for other things
• this other drive is idle (0GB used), and not yet integrated in the RAID mirror

Thank you for any clarification you may have.

THN

[THN]

I didn't want to wait for answers, so I started the process suggested by fordem.

- Unplugged the left drive while NAS was powered on, left LED turns orange, sync time says "Degraded"
- Plugged the drive in my PC, boot Kubuntu; drive shows up as /dev/sde and fdisk reports:

Disk /dev/sde: 2000.4 GB, 2000398934016 bytes
255 Köpfe, 63 Sektoren/Spur, 243201 Zylinder, zusammen 3907029168 Sektoren
Einheiten = Sektoren von 1 × 512 = 512 Bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disk identifier: 0x00000000

   Gerät  boot.     Anfang        Ende     Blöcke   Id  System
/dev/sde1              63     1060289      530113+  82  Linux Swap / Solaris
/dev/sde2         2088450  3904903484  1951407517+  83  Linux
/dev/sde4         1060290     2088449      514080   83  Linux

- e2fsck /dev/sde2 says that drive wasn't unmounted cleanly, and starts the consistency check:

e2fsck 1.41.14 (22-Dec-2010)
/dev/sde2 wurde nicht ordnungsgemäß ausgehängt, Prüfung erzwungen.
Durchgang 1: Prüfe Inodes, Blocks, und Größen
Durchgang 2: Prüfe Verzeichnis Struktur
Durchgang 3: Prüfe Verzeichnis Verknüpfungen
/lost+found nicht gefunden.  Erstelle<j>? nein
Durchgang 3A: Optimiere Verzeichnisse
Durchgang 4: Überprüfe die Referenzzähler
Durchgang 5: Überprüfe Gruppe Zusammenfassung
/dev/sde2: ***** DATEISYSTEM WURDE VERÄNDERT *****
/dev/sde2: ********** WARNUNG: Noch Fehler im Dateisystem  **********                                  
/dev/sde2: 27757/121970688 Dateien (57.0% nicht zusammenhängend), 115259821/487851856 Blöcke

- now I can mount the cleaned drive, but need to force the filesystem to avoid unbekannter Dateisystemtyp „linux_raid_member“.

sudo mount /dev/sde2 /somewhere -t ext2

I am currently browsing the drive content to see if I find something strange.

By the way: the Tools -> Raid screen still reports the exact same status in the "AUTO-REBUILD CONFIGURATION".

...to be continued...

[THN]

OK, so here we are again.

Checking the drive content thoroughly didn't reveal anything wrong. I nevertheless decided to continue the procedure, also for my own training purposes.

So I scratched all partitions and recreated one new physical partition spanning the entire drive. I did it once with ext2 format under Linux, and ran shred for some hours. But it wouldn't finish in any reasonable amount of time, so I abandoned that idea.

At that moment I realize that my initial worries might have come from the fact that I didn't only install fun_plug, but along with it another external USB-attached disk, which of course isn't part of the RAID mirror and might have created a fake incosistency in the RAID configuration. Anyway it was too late to stop the procedure at this moment.

I then reverted to WinXP, repartitioned the drive to NTFS, and ran some more smartctl and wiping operations on it without detecting anything wrong or even strange. So I decided to consider that all is OK with that drive.

I removed the NTFS partition again, and reinserted the "virgin" drive back into the empty left bay of the NAS.

When logging back into the NAS, I get the expected prompt inviting me to rebuild the RAID, but the first attempt to format the drive fails with a formatting error 107. As can be read here http://forums.dlink.com/index.php?topic=5547.msg50152#msg50152, one should disactivate fun_plug, which I did by inserting and mounting the good (right) drive locally in my workstation.

After that, I was able to successfully complete the formatting step, and I am now waiting the sync process to rebuild the RAID. Expected duration time is 5h20.

And I already have one element of success: I now see both left and right drives reported on the bottom of the Status -> Device Info page!!!


...to be continued...

[THN]

This morning the RAID was rebuilt correctly. I ran a comparison against an earlier backup an couldn't find any differences. The inconsistencies I had noticed earlier in the web interface were gone.

I then renamed the fun_plug script again in order to have it started during boot and mount my external USB-disk. After rebooting the NAS, this disk was again accessible as before under \\...\Volume_1\Extern.

But after activating fun_plug, as already suspected before, the web interface again displays the same inconsistency, with only the right drive being reported under Status -> Device Info, and the left drive being visible under Tools -> Raid with a value of "0GB". So indeed, fun_plug mounting an external USB disks seems to be the cause for this discrepancy in the web display.

This finding appeases me, as I now at least can explain the correlations. And my thorough testing also shows that it does not seem to influence at all the good operation of the RAID mirror. I think I will continue with this setup, I may only revisit the fun_plug documentation at some point in time to see if this behaviour is explained somewhere.

Thanks for reading.

THN