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

Unless you wrote over flash or you are calling the changes on bootup from funplug nothing should be permanent. Have you tried removing the drives from the device and then doing the factory reset? Also make sure you are holding the paperclip(better than a pinhead) in the back fro 10-15 seconds. If you just press it in thats a simple reboot of the device without the factory reset. Also try plugging the DNS-323 directly into your pc and setting a static ip on your PC at 192.168.0.X network and then attempting to access the DNS-323 at 192.168.0.32(The default IP Address when no DHCP is detected).

-Dlink

ooo by the way to elaborate on why the system reset did not work... if you edited a system file threw ssh is because this files are OS dependent IE: nsswitch.conf, etc... files which are not meant to be modified so the reset command will not erase your changes. correct me if I am wrong...

Just for clarification (zmarius correct me if I misunderstand) what happened is...

zmarius is running a stock DNS-323 D-link FWv1.04 (no fun_plug)

Using the web GUI - Tools->System  then Save Configuration Settings
zmarius saved a text config file on the system the browser runs from.

zmarius edited the config file (with a windows text editor changing the CR/LF
characters), then uploaded the edited config file, via the web GUI and
rebooted the DNS-323

So, Dlink, does the upload config file function from the web GUI write to flash? 
If so, will the manual reboot button on the back fix the problem?

I made the same mistake; my unit was still successfully picking up a DHCP address (same as before the mistake) and was pingable - but no web interface and no telnet, etc (mine was fun_plugged). The reset approach didn't seem to accomplish anything.

I ended up going the serial route (already had a 3.3v serial USB cable; just had to slice up a little connector to make it fit the plug). Reloaded the same config with proper LFs and it worked again fine.

The following USB serial cable is great for these projects (handles 5v or 3.3v fine): http://www.dionysusdesign.com/product_i … ucts_id/72

-Jeff

Since mine already had an IP address and was otherwise fine; I just ran the web server manually; then uploaded via the webui as usual:

"/web/webs &"

I didn't sort out all of the details; but somehow the screwed up config files must have kept startup from reaching that point.

-Jeff

I added a warning to the wiki page about this problem; as well as some technical details of where these SMB share settings actually get saved on the hard drive (/mnt/HD_a4/.systemfile/.smb.ses). With telnet access; downloading, editing, and re-uploading of whole config is unnecessary ... just edit .smb.ses directly.

-Jeff

As far as I know, there are two circumstances under which the DNS-323 will create a JBOD volume along with a RAID volume, the first is is you instruct it to create the RAID volume at less than the maximum available capacity, and the second is if the disks are of different capacities - this  creates a situation where the maximum possible RAID capacity is less than the maximum available storage capacity.

I know you said both disks were 500GB disks, but that does not mean the capacity is identical, even when the disks are the same make & model.

A hardware reset will cause the RAID array to be broken and the disks to appear as separate volumes - usually no data is lost - however, I would be interested to see what happens in a case like yours, since it theoretically involves creating three volumes.

Sorry I'm unable to answer questions 3 & 4, although I believe the answer to 3 is to use vi (I'm in between disks on my DNS-232 and don't have fun_plug installed).

One suggestion - avoid experimenting with any server holding critical data unless you have it backed up.  Make the changes you want, before putting the data on it.

just did - so it ouputs this:

# # /web/webs &

# # Quota Stop
Set Daylight :/usr/sbin/daylight f &
country = 66
Daylight mode : NONE
/web/ddns.conf:2: no such option '-----------------------------7d8e515db0606--'

and running ps after this

[1] + Segmentation fault         /web/webs

at the end of listed processes

thanks mig!
if I got the process right - 1) restore ddns.conf to its original state via Telnet 2) start webserver with /web/webs & 3) login anf change some unimportant changes to NVRAM 4) reboot to have correct ddns.conf saved in NVRAM 5) enjoy
the only thing that's missing, apart from causes of the problem (I'm sure I never messed up with CR/LF as I used UltraEdit32 for editing) is that I donno how to edit the file with PuTTY (or is there any other way to place the correct content into the ddns.conf ?) I know it sounds stupid< but that's what I am at the moment....

OK - positive outcome - managed to edit ddns.conf with VI and got the webs working the way described above!!!! thanks 2 you all!!!!
now, what concerns the reasons - the assumption about the possible cause is windows line ends as opposed to unix ones in Cfg file we download, edit and reupload
I agree, UltraEdit is not a guarantee but somehow I always used to check line ends by opening files in notepad - if it looks like one continious line (as this file looks now) - I assumed it should be OK
is there a way to make ABSOLUTELY SURE the file you are about to upload has correct CR/LF?

this might give you another clue to what is actually happening - I noticed that after my troubles each ddns.conf line turned into an FTP user seen at FTP ACCESS LIST like this CF-DDNS-ENABLE = "0"  and so on all 6 of them but they are not present in pulldown menue - hope this helps somehow

I too just freaked myself out thinking I had bricked my new DNS-323 which I've had for all of 2 weeks ( which I immediately upgraded to fw1.05; running 2 x 750GB in Raid1 config ).  Like the original poster, I also disliked the Volume_1 naming default and decided to give the proc mentioned on the Wiki to edit and reload the config file ( yes I read the warning that hadn't been tested on fw1.05 but inspecting the config file the volume name is listed so far down that it seemed harmless enough - hah! ).   In my case I'm pretty darn sure that I didn't introduce a CR/LF when editing ( I used notepad++ as reco'd by the wiki article and selected 'view newline on' just to be absolutely certain ).   

And reboot, and nada!   Ping of the last known IP worked but nothing else.  So in a panic I started hunting here and found this thread ( thank goodness ).   With limited expectations, I decided to try the hardware reset procedure mentioned before panicking further, BUT what I did differently was hold the reset button for probably a good sixty seconds rather than only 10 seconds ( my logic being it seems to take at least 60 seconds to boot ).   I was MUCHO relieved to get back the web interface with literally factory settings ( name = dlink-c8f281, no password etc. ) and it went straight into the format disks dialogue.  No biggie in my case as I had barely anything on there.

So agree with the conclusions that the reload config proc mentioned on the wiki does NOT work with fw1.05.  Secondly the hardware reset does seem to work in this case where it can't process the config file provided you hold the button much longer than suggested by the wiki article.