ALLNET 6600/THECUS 5200 Self Testing Problem ...

[dabreaka]

Hi folks,
I'm in big trouble. I use an ALLNET 6600 NAS with 4 discs in RAID10 setup. Everything worked fine for over 2 years since yesterday. The NAS wasn't reachable anymore in the net and so I did a restart via the power button in front of the panel. During the start the following messages appear: Self Test ... / Selft Testing / Booting / Assemble Raid (here on all four discs the green LEDs are lightning ...) and then the upper left led turns into orange and the display shows Self Testing ... and nothing else happens. The NAS isn't reachable via the browser or the system setup utility. I waited more than 15 hours and nothing happens. The next test was to eject all the four discs out of the NAS and then I booted the NAS - everything works fine - I used this test also to upgrade the firmware to 2.00.11.1 and this also worked fine. After reinstalling the 4 discs the same problem again ocurred: Self Testing ... / Then I bought a new Samsung 250 GB disc , ejected the 4 RAID discs and inserted it into slot 1: The NAS booted and I can setup a RAID on this single disc. This works also in SLOT 2 / 3 / 4 . So I decided there is no hardware error in my NAS and the problem is located in the RAID discs.
Now, has anyone an idea what else I can do to save my data on the four RAID discs.
Is it possbile to insert just 3 of 4 discs in different combinations to see what happens or will this destroy my raid ? ( Thinking of 1/2/3/- installed or 1/2/-/4 installed and so on ...)
Any help would be very appreciated!!!
Many thanks in advance.
Greetings , dabreaka

[pit234a]

disk1
             RAID1-A
disk2
                             RAID0
disk3
            RAID1-B
disk4

I hope, this schematic answers your question. I f not...
I am unknown to your NAS and your problem with it.
You should been able to take away up to two disks, one out of each RAID1, cause those are mirrors of each other. In this example, disk1 equals disk2 and disk3 equals disk4 and as soon as one of the equal disks inside a RAID1 is available, the RAID0 should be consistent and working, just slower. RAID0 needs correct data from both, RAID1-A and RAID1-B and RAID1A has the data doubled, as well as has RAID1-B and they are available in each, disk1 and disk2 for RAID1-A and disk3 and disk4 for RAID1-B.

[dabreaka]

thanks for the explanaiton of the RAID 10 setup - my main concern is that when i plug in just 2 or 3 discs in the nas if the nas automatically starts a rebuild try and in this case writes data and what happens when i inserted the defect disc in my first try
example : disc 1 is defect (just a guess i don't know) and i insert disc 1 + 2 + 3 in the nas
will it solve my problem when rebuilding or do I crash finally the whole RAID data in this case ...
greetings
dabreaka

[pit234a]

As said, without knowledge from my side.
The error that you experience can only (if caused by the raid) be caused by a hardware defect. It is no loss of information or something, but for some reason, the boot process hangs at a point and you assume, one defective hard drive as reason for that. Would only the software have a problem to find a disk, it would not hang during boot and you wouldn't probably have had this problems to reach the NAS over net before. One broken hard disk, that is not available, causes no further trouble. A disk, that produces a short circuit or something mad on data bus, so, does influence the rest of the hardware due to its fault, is probably what you are searching.
Without that hardware defect, that causes that hanging, your NAS should boot up and you should get access to your webif again and thus, be able to administrate that device as usual and also have a look at the raid and what it asks for.
Without hardware failure means, without that disk that produces the error. It will still be missed.
The raid only can handle the appropriate disk on its place: you are not allowed to swap them, let's say from slot two to slot three and probably not even from slot one to slot two, even when those are mirrors and carry the same dates, it could cause a further error when the wanted drive is not found, but an other one is present. It could, it is no must and depends on the abilities of the raid. There are different controllers and software out, you better consider it acting restrictive and let every disk on its original slot for the moment, until you found the one that is defective and can boot again and visit the webif.
You should have a chance to replace the defective disk before you start rebuilding the raid and you should not run into an automatic rebuild, when you leave every disk in the slot it was in before.
Hope, I understood what you are asking.
I would start taking out disk one, boot up and see what happens. No error: bingo. Still error: back in with it and out disk two. Again, boot up and see: no error: bingo, still error: next one...
That should never bring you in trouble loosing your raid or the data.

[dabreaka]

thanks a lot pit234 for your detailed answers and suggestions - I will try all your ideas tomorrow and I'll keep you posted ...

[dabreaka]

hi pit234 - today I did the tests - first I tried the following

1/-/3/4 (disc 2 ejected) : no change in behaviour
the next test was
-/2/3/4 (disc 1 ejected) : bingo! the system finished the booting sequence - and I was able to connect via webif - but then : NO BINGO! the raid was not there - the webif shows as if there was no raid declared - I only can press the "New raid" button (i didn't of course ...)

so it seems that there is a problem with disc 1 but I cannot access the raid with only the 3 other discs - very diappointing ...

any ideas left?
best regards

dabreaka

[pit234a]

At least it looks like you found the one disk that produced the error during boot up.
I don't know about your raid and the behavior. In my NAS, I have a software raid in and I build myself software raids a few times before. They all acted unharmed, when I lost a disk. Therefore, I really wonder what happened on your side. I can imagine, that there was information on that disk, that now is needed, but than this would be a very bad raid solution. I could imagine, you mixed up the disks during trouble shooting or I could imagine, the raid asks for a disk one to be able to work. I don't know, but you are easy able to replace the disk one with the disk two (in case you configured the raid according to the example above, disk kernel- numbers not necessarily represent their place in a raid!) just to see what happens and there you should at least see no further trouble.