Another LaCie 5big Network 2 problem

[Mijzelf]

Indeed 4 disks are needed to assemble the array.

Booting with only one disk won't harm the raid member. mdadm will simply fail to assemble the array, and leave the partition alone. The question is, is it safe to put the disk in a dying box? Don't know. Depends what is actually failing.

How to check the power supply is another question.

Yeah, that's a problem. According to the specs posted by tenhouse, it has 4 pins (+shield), but it only outputs 12V. OK, at 12.5A. So maybe it has that much pins to divide the current. 12.5A is a lot!
You could measure the voltage, if it's lower than, lets say, 11V, the supply is broken. Problem is that when it outputs 12V you still don't know if that still will be 12V under load. The voltage should be measured while the box is booting. But doing so will void the warranty.

What I dont see is:
Why is the humming soudn only there when I remove 2 disk -shouldn't it be the exact opposite?

Maybe the box is degrading while we speak?

[fvdw]

indeed strange why fdisk doesn't find these disks, but the output in dmesg for port 4 & 5 is ok to my opinion. That would mean the mainboard seems to function ok.

The problem could be disk 3. I think it is save to put in disk1+2 and 4+5 and then boot without a risk for the raid. But lest wait for Mijzelf inputs

[fvdw]

Rereading your first post

and 1 disks (its always the "middle" one by the way) was missing again

this is disk 3. if that takes too much power (because it is broken) it could cause the problem.
Reading the post from Mijzelf I think you can do the test with 4 disks and leave the central one out

The standalone kernel will not build the array it will just boot in ram and load the the disk drivers nothing more.

[tenhouse]

Good evening ladies and gentlemen (well or only gentlemen)

I just put out the middle disk and booted first to the dashboard: 3 disks were gone (3,4,5) and then to the tool which took its 120 seconds as always. So no luck there.

@Mijzelf

Well the "sound" was back to normal as soon as I put the 2 disks back. I don't think the sound has too much to do with it, maybe only there because of the lack of an even balanced weight or something :P

And I don't know how to measure the power - I don't have a measurement-device or anything.

[Jocko]

Hi tenhouse,

I just put out the middle disk and booted first to the dashboard: 3 disks were gone (3,4,5) and then to the tool which took its 120 seconds as always. So no luck there.

Why did you plug a disk in the slot 3

I am pretty sure that when Fvdw wrote

The problem could be disk 3. I think it is save to put in disk1+2 and 4+5 and then boot without a risk for the raid. But lest wait for Mijzelf inputs

He asked you :
- not use the disk3
- keep the other disks in their slot

[Mijzelf]

And I don't know how to measure the power - I don't have a measurement-device or anything.

Then it stops, I guess. You can connect 4 disks to another Linux device (PC), and try to assemble the array:

Code: Select all

mdadm --assemble /dev/md0 /dev/sda2 /dev/sdb2 /dev/sdc2 /dev/sdd2 --run

where you have to check in /proc/partitions which devices actually to use.
When that gives no error, you can mount it:

Code: Select all

mkdir /tmp/mountpoint
mount -o ro /dev/md0 /tmp/mountpoint

Now you can use your GUI to copy the files from /tmp/mountpoint

@Jocky: He did: "I just put out the middle disk".

[Jocko]

@Mijzelf
Ok, sorry I read too quickly :whistle and understood that 3 disks were still plug in slot 3,4 and 5

[fvdw]

the only thing that could cause a sound is the fan. however I can not see a reason why the sound should be influenced by putting in two disks more.
If the fan is blocked and cannot rotate it will draw more power but to my opinion never so much that the disks cannot start.

A new power supply is rather expensive, so can imagine that you are reluctant too go that route.
I think measuring the voltage of the power supply with no load has not so much use it seems to be capable to let at least two disk start and run.You must measure it under full load (with 5 disk at boot0.
The the problem could also be on the mainboard as the power will be led to the disks. See for a picture of the mainboard here. As far as I know a harddisk needs a 12 and 5 Volt DC supply . You could disconnect the HDD power connectors from the mainboard and connect them to another power supply for HDD (for instance an PSU form an old PCU.
But measuring the voltage first under load with 5 disks at boot is by far the best way to start the investigations. (simple DC voltage meter is not too expensive)

[tenhouse]

Hello.

I think it's over.
I attached 4 of the 5 disks to a mainboard and booted Gentoo linux and did what you said Mijzelf.

First I attached disk 1,2,4 and 5 because I was pretty sure 3 is the broken one.
But with this setup I can't even past the POST BIOS stuff. I hear the (well known) "clicking"-sound coming from the disk 4 and it stays forever on "detecing SATA Device on port x" - pretty sure it's broken.

Then I attached disk 1,2,3,5 and I can boot with the live cd. All HDs seem fine in the procs/partitions. But when I try to assemble it fails.

See my setup and the failure in the screenshots.

http://i42.tinypic.com/20uzkom.jpg
http://i40.tinypic.com/in95sg.jpg

[fvdw]

:shock: impressive setup.

If disk 4 is the problem and the 4 remaining are ok then putting disk,1,2,3 and 5 into the 5big2 (in slot 1,2,3, and 5) should also let it boot using the standalone kernel. (I would not let it boot in the Lacie firmware with the message as shown in your screen shot)

In the picture of your screenshot it mentions that it ignores "sdb2" because it reports sda2 as failed.
This might be wrong, before giving up wait for the comment of Mijzelf, he was able to recover a raid array from unlucky1 where we had two disks that were reported as failed due to a problem with the partition tables. if the partition is untouched then copying the partition info from one of the disks that are ok could resolve this problem. We did it once...