DSM-G600, DNS-3xx and NSA-220 Hack Forum

Hi Philpcs,

philipcs wrote:

The RAID volume is synchronizing now. Please wait for 1766.3 minute(s).

The RAID volume is synchronizing now. Please wait for 6476.3 minute(s).

Increased again

Eventually the minutes remaining should go down until the RAID is active again.  Let us know if/when that happens and the status of the amber light.  You may need to reboot after sync is complete to get rid of the amber light.

I have attached dmesg.out after turned to amber light

I had a look and the relevant lines:

md: md0 stopped.
md: bind<sdb2>
md: bind<sda2>
md: kicking non-fresh sdb2 from array!
md: unbind<sdb2>
md: export_rdev(sdb2)
raid1: raid set md0 active with 1 out of 2 mirrors
md: md1 stopped.
md: bind<sdb3>
md: bind<sda3>

In your case, I expect that you will be able to get back to normal operation after sync is completed.  Do keep us informed as to what really happens.

Jaya

Hi Philipcs,

philipcs wrote:

I have tried 1 thing. i have 2 volumes.

volume 1 = RAID1

volume 2 -JODB

When i unplugged the HDD with amber light, my volume 2 gone and i can't access my data. But when the HDD put it back, i can still access my data.

Doesn't meant my HDD is still healthy even with amber light?

Volume_1 being RAID1 will continue to work when one member disk is faulty.

Volume_2 being JBOD requires both disks to be online for it to work.

This is just how RAID1 and JBOD are design to work.

Jaya

PS:  I gather this is how you ended up with the orange light.  In this case, I am sure that it will all get back to normal when you reboot your DNS-323.

Hello Slash,

slash wrote:

After upgrading to 1.04 version it appeared to be two new options
in WEB interface at ADVANCED / NETWORK ACCESS menu: OPLOCKS and MAP ARCHIVE.

If nobody tried to do this, maybe try to change status of OPLOCKS checkbox for RAID-1 partition
and observe, if "amber led" problem will appear again, or not?

In my case (with three DNS-323's) both OPLOCKS and MAP ARCHIVE are enabled.  When the amber light developed on these boxes, I have always traced it to either an improper shutdown, or hot removal followed by hot insertion.

I was able to recover from these by doing mdmadm -a /dev/md0 /dev/sdx2 through CLI, where x is a or b for the RHS or LHS disks, respectively.

Hope this helps.

Jaya

slash wrote:

Hello!

I'm not too experienced in mdtools, and also not too experienced DNS-323 user..
For now I've used DNS323 for a very short time. And I've had no problems with RAID-1 partition
(RAID-1 + JBOD)

But I have one ..idea.. After upgrading to 1.04 version it appeared to be two new options
in WEB interface at ADVANCED / NETWORK ACCESS menu: OPLOCKS and MAP ARCHIVE.

If nobody tried to do this, maybe try to change status of OPLOCKS checkbox for RAID-1 partition
and observe, if "amber led" problem will appear again, or not?

Of course this is only theory, I know. Just try it, maybe it solves "amber led" problem...?

Sorry for my english,

Have a nice weekend
/Slash

I don't think the oplocks and or map archive check boxes have anything to do with it - I've seen the white/amber light once, and that was caused by my pulling and then reinstalling a drive in a JBOD configuration (no RAID) whilst working on some "data loss scenarios" - please note - I did not hot plug anything.

Apart from that my unit, currently in a RAID1 configuration with a pair of Seagate 7200.9 Barracudas, does a pretty good job, sitting in the corner, sharing files, and the occasional print job - I have fooled around with both the oplocks and the map archive (which still doesn't work) but no amber lights resulted.

Has anyone considered compiling a list of what drives were in the units that had amber lights?  Is there any similarity?

Thanks Adubin, I will try tonight

Position 1 - Firmware 1.03 and below, both LEDs Blue. All okay.
Position 2 - Since 1.04 with no formatting, left LED is off, all functionality good, Web GUI reports all okay with the Raid1 volume.
Position 3 - happened yesterday - Left LED off, and right LED Orange, Even the On/Off switch is Orange. All functionality good Web GUI okay too.

If I reboot device it goes back to Position 2, by next morning I am back at Postion 3.

I am going to fordem the data  [ie backup!] and swap the disks from left to right and reformat. if all is not well then I will at least have discovered it the disk with the orange light is faulty or if its the drive bay with the issue. I will revert to firmware 1.03 if necessary.
This is the first trouble I have had with this box, and am reluctant to think its the disks.
What if I end up putting the dodgy disk on the site where the LED does not light, then if things do go awry I will have to rely on the email (which I havent set up...)

Last edited by index monkey (2008-04-10 12:37:03)

Bits wrote:

Told DLink Customer Service, asked me to bring the NAS down, they tested it, no prob with the drives and replaced the NAS!

Where did you go to talk to D-Link Customer Service in person?

Oh, I now took the time to actually read the 'reformat message'.

It says:

"Click Next to begin formatting the replacement drive. Re-synch will take place after the restart"

So it's not a complete reformat like I thought.

I guess I could do that without the data on the working drive being erased and then it can start syncing again?

Or would you still advice me to back up? (asking the question I can somehow predict the answer :-)

Last edited by bagmanden (2008-04-14 12:20:24)

bagmanden wrote:

Oh, I now took the time to actually read the 'reformat message'.

It says:

"Click Next to begin formatting the replacement drive. Re-synch will take place after the restart"

So it's not a complete reformat like I thought.

I guess I could do that without the data on the working drive being erased and then it can start syncing again?

Or would you still advice me to back up? (asking the question I can somehow predict the answer :-)

Theoretically - it's quite safe to let if format the replacement drive and resync - practically, it seems to be a different thing.

Personally I have not lost data in RAID failure & re-build simulations, however other forum members have reported problems, and it seems that the common factor has been replacement disks that contained partitions &/or data.

One possible reason why I may never have experienced the problem (with the DNS-323) is that I learned many years ago (mid 90's) in a Novell NetWare lab simulation that a resync could go in either direction if the system was unable to determine which disk had valid data, and the easiest way to ensure a good resync was to eliminate the source of confusion.

I would suggest backing up (just to be on the safe side) and making sure that the replacement disk had no partitions &/or data.

jiggysmb wrote:

my 2c. 

After upgrading to 1.04 my 2nd drive went JBOD and the print server stopped working.

The print server issue is well documented - but the second drive went JBOD ?!?

What disk configuration did you have before the upgrade and what is your definition of JBOD?

I know some people think of JBOD as each disk being handled separately, but on this device JBOD is specifically a single concatenated volume spanning the two disks.

I can't really comment on the DNS-323, so to speak, losing track of the drive after (or as a result of) the firmware upgrade, I haven't had it happen personally, perhaps because I wasn't running RAID when I did my upgrade - unfortunately, firmware upgrades can be like that, and it may or may not happen again.

As to your second question - several companies make relatively expensive USB to SATA/PATA disk adapters, the last one I bought probably cost me $19.95 - there's no enclosure, just a power adapter and a little box with a USB cable on one side and three disk connectors on the others - you could try one of those, I bought mine for use as a troubleshooting tool.

Hi

I've also the pink/amber light issue with one of my drive. (The left one when looking at the front of my dns323)

However, my two drives and the dns-323 were still under the warranty so I've went back to the store and they test my 2 sata drives and gave me back a new dns-323. Still the same issue.

However, after some testing from the technicians, the 2 drives are stated are healthy (no issue).
On my side, with some testing, the same thing, the two drives seems ok.
In normal mode (not raid 1 or 0 nor JBOD) I never had that issue of one drive failing, weird.

I'm also on firmware 1.04
Formatted into raid 1 after applying the firmware 1.04 (on both dns-323)
my two drives are 500 gigs Seagate ST3500320AS

Hope to find the problem.

I'm more skeptical now about desynchronize data.

I've made the following test:

Format raid 1
Do nothing (both disk empty)
close computer
Switch off the dns-323
Wait (more than 8 hours)
open computer
Switch on the dns-323

Problem appear anyway


What I've remark also is once the formating is done, drives (raid 1) are usable normally. Closing the dns-323 for a few moments (2-3 minutes) and start it back does not trigger the issue regardless of having data in it. Tried it a few times, didn't trigger the issue. Having a long delay however trigger the issue.

I do believe I've waited long enough (2-3 minutes) between each successful attempts to suppose that both drives in the array stop spinning (to discard that hypothesis of having drives still spinning when switching it on again)