Purple leds, can't access to my NAS webpage

[Aurélien GRACIA-CAMPO]

Hello,

My Nas doesn't show its webpage, the left array containing my HDD is blinking in a purple colour, the right one blinks in orange
I let it that way for more than 24 hours, before doing a reset (pressing back button for 20s), then I could access the NAS webpage back again and I fulfilled the wizard, rebooted it, and then the same problem happened again, the nas would show no webpage and the purple and orange leds where blinking again, it seems the only way for me to access the NAS webpage is by resetting it,

I held the power button to be sure ALT F was active and I got confirmation, Left disk blinked with orange colour for 3 times, then right one,

I think the source of my issue is when I initially tried to set up the "folders setup", I wanted users not to access the entire sda2 so I unticked the 3 boxes "browsable" "read only" and "inherit perms", to tick them only in a specific shared folder in sda2,

My NAS is composed of only one WD Caviar Green 3Tb, last time I saw its status it was degraded,
I also followed one of your tutos:

In the RAID web page, RAID Maintenance section, I selected in the md0 line,

   Raid Operations->Create Bitmap.

I read this tutorials of yours without knowing what information to use:
http://sourceforge.net/p/alt-f/wiki/About%20Buttons%20and%20Leds/
https://code.google.com/p/alt-f/issues/detail?id=59
https://groups.google.com/forum/#!msg/alt-f/5fdbE0Jxdts/9ZgDr4gxLmwJ

Is it normal, will my NAS come back to normal state if I wait more? If then, how much time approximately do I have to wait?
By the meantime, my folders are still accessible under Windows, it seems my Samba share isnt broken

Here is what the Status page looks like:

Disks

Bay	Dev.	Model	Capacity	Power Status	Temp	Health
left	sda	WDC WD30EZRX-00DC0B0	3.0TB	active or idle	39°C/102.2°F	passed

RAID

Dev.	Capacity	Level	State	Status	Action	Done	ETA
md0	0.5GB	raid1	clean	degraded	idle

Mounted Filesystems

Dev.	Label	Capacity	Available	FS	Mode	Dirty	Automatic FSCK in
sda2		2.7TB	1.2TB	ext3	RW		10 mounts or 175 days
sda4		484.3MB	469.0MB	ext3	RW		16 mounts or 178 days

[Aurélien GRACIA-CAMPO]

I managed to get the Web interface back by resetting the NAS and loading an old save, I suspect I didn't tick the Static ip option though,

anyway, now the NAS still shows a purple led on the left array and orange led on right one, drive is still working, but I have no information about any maintenance operation in my web interface, what should I do?

[João Cardoso]

On Monday, December 8, 2014 11:47:16 PM UTC, Aurélien GRACIA-CAMPO wrote:

Hello,

My Nas

Model? Alt-F version?

 

doesn't show its webpage, the left array containing my HDD is blinking in a purple colour, the right one blinks in orange

orange means a degraded RAID, blinking might means RAID reconstruction, purple is the result of blue+orange.

how is the blue power led? blinking? heartbeat? steady? how it behaves depends on your box model.

 

I let it that way for more than 24 hours, before doing a reset (pressing back button for 20s), then I could access the NAS webpage back again and I fulfilled the wizard, rebooted it, and then the same problem happened again, the nas would show no webpage and the purple and orange leds where blinking again, it seems the only way for me to access the NAS webpage is by resetting it,

I held the power button to be sure ALT F was active and I got confirmation, Left disk blinked with orange colour for 3 times, then right one,

OK. Then you probably have a network (IP) issue

 

I think the source of my issue is when I initially tried to set up the "folders setup", I wanted users not to access the entire sda2 so I unticked the 3 boxes "browsable" "read only" and "inherit perms", to tick them only in a specific shared folder in sda2,

Don't think that to be the issue cause.

 

My NAS is composed of only one WD Caviar Green 3Tb, last time I saw its status it was degraded,

If you have only one disk, then the RAID (1?) is degraded, thus the orange led

 

I also followed one of your tutos:

In the RAID web page, RAID Maintenance section, I selected in the md0 line,

   Raid Operations->Create Bitmap.

From the info bellow, your RAID is in really for a swap device, a D-Link idiosyncrasy, and there is no point in adding a "write-intent bitmap" in such setup.

I recommend you to destroy that RAID and create a standard swap device instead.

 

I read this tutorials of yours without knowing what information to use:
http://sourceforge.net/p/alt-f/wiki/About%20Buttons%20and%20Leds/
https://code.google.com/p/alt-f/issues/detail?id=59
https://groups.google.com/forum/#!msg/alt-f/5fdbE0Jxdts/9ZgDr4gxLmwJ

Is it normal, will my NAS come back to normal state if I wait more?

Even if you had a RAID1 and it was reconstructing itself or if filesystem is being checked you should have access to the box. That is not the issue.

Yes, you have swap over a RAID1, there is nothing wrong with that, but it is a waste of resources.

You have to use the command line to "fix" it, if you want to.

First verify that swap is indeed being used over RAID1:

cat /proc/swaps 

If you are using swap over RAID1 the above command will show something similar to

Filename                                Type            Size    Used    Priority

/dev/md0                                partition       524272  0       -1

The important is the RAID device name, /dev/md0 in the above example. You should use what is reported for your box.

If it has the form /dev/sd... and not /dev/md... then it is not a RAID, don't continue with this procedure.

Now obtain the disk partitions over which the RAID is build:

mdadm --examine --scan --verbose

you will see something similar to

ARRAY /dev/md0 ...

   devices=/dev/sdb1,/dev/sda1

ARRAY /dev/md1 ...

   devices=...

ARRAY /dev/md2...

   devices=...

The important is the disk devices names matching the RAID device, in the above example /dev/sda1 and  /dev/sdb1

As you have only one disk, only one device will appear, probably /dev/sda1. In the following use only the device that is reported above.

swapoff /dev/md0 # stop swap over RAID1 device md0
mdadm --stop /dev/md0 # stop the md0 RAID. amber/orange led will be turned off
mdadm --zero-superblock /dev/sda1 /dev/sdb1 # destroy RAID information from the RAID disk components sda1 and sdb1
mkswap /dev/sda1 # create swap on sda1
mkswap /dev/sdb1 # create swap on sdb1
swapon -p1 /dev/sda1 /dev/sdb1 # activate swap on both

Mounted Filesystems

Dev.	Label	Capacity	Available	FS	Mode	Dirty	Automatic FSCK in
sda2		2.7TB	1.2TB	ext3	RW		10 mounts or 175 days
sda4		484.3MB	469.0MB	ext3	RW		16 mounts or 178 days

...

I managed to get the Web interface back by resetting the NAS and loading an old save, I suspect I didn't tick the Static ip option though,

The default is to ask and IP to your DHCP server, which is normally your router. If you want to use a dynamic IP, you should verify that you router is supplying one, and what IP was supplied by the router web page. Normaly this will appear under lease or client on the router DHCP web page.

 

anyway, now the NAS still shows a purple led on the left array and orange led on right one, drive is still working, but I have no information about any maintenance operation in my web interface, what should I do? 

The above procedure to "fix" the swap "issue". If it works, and based on your own experience, add a wiki entry on "how to remove swap from a RAID1" :-)

[Aurélien GRACIA-CAMPO]

Hello João, thx for your quick reply,

I did a bit of investigation meanwhile, so I'll give you a first part of answer, about what I think was causing the "orange + blue led = purple",

I will study and apply your advice about raid configuration and give you part2 as soon as I assimilate all you told me (I'm noob to setting NAS and raid)

So my NAS is a Dlink DNS320, fw ALT-F 0.1RC4

I managed to find the service that was causing my NAS to overcharge:

forked_daapd DAAP and RSP media server (iTunes)

When this service is active I have 100% of cpu usage and 2.40 load

When disabled I come back to a steady 10% cpu and 0 to 0.1 load

This overcharge was causing the left blue led to blink like crazy and give this purple colour,

Now that I disabled forked_daapd, I may confirm that my HDD doesn't excessively vibrate anymore,

My 3 NAS leds are now in a steady orange colour, so I still have an issue and must study your RAID suggestions above.

Btw, here is what is displayed in my Disk/RAID menu:

Dev.	Type
md1

RAID Maintenance

Dev.	Capacity	Level	Ver.	Components
md0	0.5GB	raid1	0.9	sda1

[João Cardoso]

On Tuesday, December 9, 2014 8:48:36 PM UTC, Aurélien GRACIA-CAMPO wrote:

Hello João, thx for your quick reply,

I did a bit of investigation meanwhile, so I'll give you a first part of answer, about what I think was causing the "orange + blue led = purple",

I will study and apply your advice about raid configuration and give you part2 as soon as I assimilate all you told me (I'm noob to setting NAS and raid)

So my NAS is a Dlink DNS320,

Not the 320L model? a rev-A1 board? See the box bottom attached label.

 

fw ALT-F 0.1RC4

I managed to find the service that was causing my NAS to overcharge:

forked_daapd DAAP and RSP media server (iTunes)

When this service is active I have 100% of cpu usage and 2.40 load

When disabled I come back to a steady 10% cpu and 0 to 0.1 load

This overcharge was causing the left blue led to blink like crazy and give this purple colour,

Now that I disabled forked_daapd, I may confirm that my HDD doesn't excessively vibrate anymore,

After the initial scanning of all your files, for finding the media files to serve, forked-daapd should not consume CPU anymore.

 

My 3 NAS leds are now in a steady orange colour, so I still have an issue and must study your RAID suggestions above.

Btw, here is what is displayed in my Disk/RAID menu:

Dev.	Type
md1

RAID Maintenance

Dev.	Capacity	Level	Ver.	Components
md0	0.5GB	raid1	0.9	sda1

Yes, you have swap over RAID1, and it is degraded as you have only one disk.

Please verify if you have a DNS-320L and not a DNS-320 board. I though that only the 320L had swap over RAID1.

[Aurélien GRACIA-CAMPO]

Hello,

Sorry for late reply, I confirm my previous post, my NAS is a (not L) DNS320,
I did the first part of your tutorial, and removed the raid 1 swap partition, this successfully removed the orange leds,
I still have a red message in my status page that states no swap, but I fear from using the partition menu to create a new swap partition, I could erase my sda2 data partition by doing wrong, so I'll let it untouched while I'm not more documented,

Here is a look at my status page now:

System

Temperature 37.0°C/98.6°F	Fan speed 0	Load 0.43	CPU 10%	Memory 60% of 128MB	Swap None
Name: dns325			Model: DNS-320-A1A2
Date: Thu Dec 18 10:32:06 CET 2014			Uptime: 6 day(s) 15 hour(s)

Network

Speed:1000Mbps Duplex:full MTU:1500 TX:2.7GiB Rx:158.4MiB IP:192.168.0.32

Disks

Bay	Dev.	Model	Capacity	Power Status	Temp	Health

left

sda

WDC WD30EZRX-00DC0B0

3.0TB

active or idle

29°C/84.2°F

passed

Mounted Filesystems

Dev.	Label	Capacity	Available	FS	Mode	Dirty	Automatic FSCK in

sda1		509.7MB	493.2MB	ext3	RW		30 mounts or 172 days

sda2

2.7TB

1.2TB

ext3

RW

32 mounts or 167 days
sda4		484.3MB	439.4MB	ext3	RW		13 mounts or 169 days

[João Cardoso]

You have a filesystem on sda1, which is not a standard setup. sda1 is usually reserved for swap on both the D-Link firmware and Alt-F.

First verify that you have no data of yours on /mnt/sda1, as the next procedure will delete it.

Then execute the following commands:

umount /dev/sda1 # unmount the filesystem

mkswap /dev/sda1 # create swap on sda1

swapon /dev/sda1 # activate swap on it

 

That's it. You don't need to reboot, swap is now active. On the next reboot it will be automatically available.

[Aurélien GRACIA-CAMPO]

Thanks for your help, swap is now active, and I haven't lost any data, but still have to reconfigure some settings though, sda1 has now disappeared from partitions, can you just take a look at my status screen to tell me if it's ok?

System

Temperature 38.0°C/100.4°F	Fan speed 0	Load 0.05	CPU 17%	Memory 79% of 128MB	Swap 0% of 517MB
Name: dns325			Model: DNS-320-A1A2
Date: Mon Dec 22 12:34:53 CET 2014			Uptime: 0 day(s) 0 hour(s)

Network

Speed:1000Mbps Duplex:full MTU:1500 TX:8.5MiB Rx:1.5MiB IP:192.168.0.32

Disks

Bay	Dev.	Model	Capacity	Power Status	Temp	Health

left

sda

WDC WD30EZRX-00DC0B0

3.0TB

active or idle

34°C/93.2°F

passed

Mounted Filesystems

Dev.	Label	Capacity	Available	FS	Mode	Dirty	Automatic FSCK in

sda2

2.7TB

1.2TB

ext3

RW

31 mounts or 163 days
sda4		484.3MB	433.5MB	ext3	RW		12 mounts or 165 days

Remotely Mounted Filesystems

Host	Share	FS
donparadise	Partage	cifs

[Daniel Mladek]

Hello there, and many thanks to João,

I have just successfully followed the steps and created a Wiki page for it: https://sourceforge.net/p/alt-f/wiki/How%20to%20remove%20swap%20from%20a%20RAID1%20(AKA%20Your%20md0%20raid1%20device%20is%20degraded)/

Cheers,

-d