Print

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

[xcj]

Hi everybody,

I have a DNS 343 with 4 disks installed:

Volume 1 is a Western Digital Green 2TB
Volume 2 is a Seagate 2TB
Volume 3 is a also a Western Digital Green 2TB
Volume 4 is a a Western Digital Green 1TB

All configured as single volumes ( if needed i can get the exact models later).

I had no problem moving data into the disks, the speed was constant on the transfer, but when i started to use the data on the NAS (watching videos over Samba) i got a problem only specific to Volume2, the data troughput simply stops for a while and i have start playing the video again (resuming from where it stopped), i don't have this problem on videos stored in the other volumes.

I decided to move the data to my computer and reformat the disk but moving the data to my computer is being a real pain since the problem is exactly the same, i am using windows 7 and have a gadget on the desktop that monitors network activity, and i see the data flow stopping and after a while starting up again.

I also tried to move data from volume2 to one of the other volumes accessing the NAS via SSH (i have fun_plug installed) and what i could see running the "top" command is that the process responsible for the copy would become dormant ("D" state) and active again after a while.

Does anyone have any ideas where i should start to look for the problem? Or did anyone ever had this problem?

I am considering taking the drive out and using some util to read EXT2/3 on windows to get the data off from the drive...

[JavaLawyer]

Given that this issue is isolated to HDD2/Slot-2, and all HDDs are formatted as "Standard Volumes", my initial thought is that there may be a physical issue with HDD #2 or Slot-2 of the DNS-343. You can rule out the DNS-343 by formatting a spare HDD in slot-2 to see if data throughput goes back to normal.

There are apps that you can use to physically mount the HDD in a Windows PC: DNS-343 - Data Recovery (Windows PCs). Although this method can be used to extract data, there's always an inherent danger of having the HDD unreadable by the DNS-343 if the host PC/mounting software performs any writes modifying the filesystem.

It sounds like you don't have a backup of your data. I don't want to start preaching here, but when you get a chance, please read the following sticky post: DNS-343 - Data Backup vs. Redundancy

[xcj]

Thanks for the reply, i have the data on volume 4 four mirrored on another disk connected to a computer , volume 3 is a backup of the hard disks of two game consoles, volume 1+2 are where i store videos that after being watched are either deleted or moved back to the computer for dvd/blueray burning.

Focusing on the theory that it's either a problem the SLOT 2 BUS or the disk itself, shouldn't it register on a log? I browsed trough the system LOG on the webUI and on the /var/log, and there is nothing that registers the problem.

Regarding getting the data out via a PC i always follow the rule of no writing to the disk, i only use tools that treat the disk as read-only.

[JavaLawyer]

Focusing on the theory that it's either a problem the SLOT 2 BUS or the disk itself, shouldn't it register on a log? I browsed trough the system LOG on the webUI and on the /var/log, and there is nothing that registers the problem.

Ideally, the DNS-343 should pickup on a failing HDD. Did you run a comprehensive S.M.A.R.T scan against the HDD? Practically speaking, I usually don't take system generated results at face value, especially if they don't make sense. I can't say whether the DNS-343 would pickup on a hardware issue with Slot-2.

[xcj]

Hi, i tried a different approach, yesterday i swapped disk 2 with disk 3, i edited the files mentioned on this post "http://forums.dink.com/index.php?topic=6473.30" and i tested copying files from both these disks and then the remaining 1 and 4, and all of them sustained the troughput, no pauses, so i can't really see the logic.

Now there is something a bit odd, the volume 2 share, even when the disk was moved from slot2 to slot3 still remains the volume2 share, in other words the volume2 share shows the content of disk3, i can live with this if there is no workaround.

The other issue, the one that bugs me is that when i enter the WebUi it tell's me that the disks are in the wrong order and i MUST to change them back.
If i run from ssh the "hd_verify" it also says that the order is incorrect and i have to change the order "* All disks are OK, but we need to change the order.see "/web/hdd_order" "

Is there anyway to make this the permanent order?

On a side-note, i feel I'm asking support for an unofficial software (fun_plug) but i had to resort to it when i added more disks, the first time i added a disk the DNS343 with stock firmware and no fun_plug installed it formated the already existent one!! , so i tried the procedure described on the same post i mentioned above and prevented that from happening again when i added the last disks.

[JavaLawyer]

A couple of rules that you need to keep in mind when manipulating volumes in standard configuration on the DNS-343:

• The volume name is linked to the HDD slot number, not the HDD itself (Slot_1 = Volume_1, Slot_2 = Volume_2). If you move a HDD from slot_1 to slot_2, the volume number will change
• The DNS-343 wants HDDs to remain in the slot in which they were originally configured/formatted. If a HDD was configured in slot_2 and moved to slot_3, the DNS-343 will complain

If memory serves, there are some very old threads on this board (or perhaps the DNS-323 board) where users figured out how to manually remove the HDD location dependency. The question is whether it's worth the effort and potential risk?

[xcj]

Well, if this order solves my previous problem i would say it is worth it!

I will search the forums later to see if there is a workaround.

Thanks for your help!

[JavaLawyer]

No problem. If/when you find the old posts and if you establish a working solution, please post the referenced (old) threads here along with the solution for other users to see.

Good luck!

[xcj]

Hi there, two things to report:

1-I tried changing back the disk2 and disk3, and disk2 started to have performance problems again, i don't know the reason for this and why this only happens in this disk order.

2-Swapped disk2 and disk3 again, performance problems gone, i can even get 30Megabyte/s sustained troughput (gigabit network with jumbo frames OFF), to solve the webgui complaining about the disk order , after editing the files i already mentioned all i needed to do was to run "hdverify -w", and it is now solved!

[xcj]

3 Things:

The order in which the shares are created (the relation bewteen "volume_#" and disk slot) is managed on the file /mnt/HD_a4/dsk_mapping , you can swap "sda, sdb, sdc and sdd" to change the order of the volume numbers shares.

[JavaLawyer]

Would you be willing to post more detailed step-wise instructions? I can add this on our FAQ board for reference by other DNS-343 owners experiencing similar issues.

Thanks! 

[xcj]

No problem, but i don't know if all the steps i took are required, i did them because they made sense to me, maybe all that was needed was the the "hd_verify -w" command.

All i can share as confirmed is that if you get speed issues on one disk only, changing the slots could help.

I was having speed issues on disk2 so i:

1. Swapped the disk2 with disk3

2. Accessed the NAS trough SSH, but if it was telnet it would make no difference (you need to have fun_plug intalled).

3.Edited the files on /mnt/HD_a4

3.1 File "dsk_mapping" <-This allows me to keep disk 3 showing up as the "volume_2" share.
sda
sdc <--was sdb
sdb <--was sdc
sdd

3.2 File "raidtab2web" <--Not sure if this was needed.

Was:

status,0
filesystem,ext3
nonraid,sda2,sdb2,sdc2,sdd2
md0,null,null,null,null,null,null,null
md1,null,null,null,null,null,null,null
md2,null,null,null,null,null,null,null
md3,null,null,null,null,null,null,null

Is now:

status,0
filesystem,ext3
nonraid,sda2,sdc2,sdb2,sdd2
md0,null,null,null,null,null,null,null
md1,null,null,null,null,null,null,null
md2,null,null,null,null,null,null,null
md3,null,null,null,null,null,null,null

3.3 File ".x37c9E" (MAC's, serials and Magic# censored) <--Also not sure if this was needed.

Was:

magic: cxxxxxxxxxxxxxxxxxxxxx7
mac: 1xxxxxxxxxF
node: sda
p2: NORMAL * Z1xxxx2E * WD-WCAxxxx3129  * WD-WMAxxxxx083

Is now:

magic: cxxxxxxxxxxxxxxxxxxxxx7
mac: 1xxxxxxxxxF
node: sda
p2: NORMAL * WD-WCAxxxx3129 * Z1xxxx2E * WD-WMAxxxxx083

4. Run command "hd_verify -w"

5. Access WebGUI, and reboot the system

6. Access the WebGui and confirm that it no longer bugs you to change the disks back to the previous order.

7. Test copying files from each volume to a PC and see if the speed transfer is normal and sustained.

[JavaLawyer]

Thanks for the quick turnaround. . . I'll work on putting together / formatting a FAQ entry over the weekend.

Thanks again.