Unfortunately no one can be told what fun_plug is - you have to see it for yourself.
You are not logged in.
Hi All,
I've search through the forums and of course google and haven't been able to come up with anything. I'm posting here seeking some guidance. I realize this is a forum about the DNS-323 (which I also have one of and is running ffp flawlessly) but was hoping someone could point me in the right direction to diag the problem.
I have a DNS0-343 and am trying to get ffp 0.5 working on it. I've been able to get ffp installed and the packages downloaded and installed. My problem is that after I reboot the system I'm no longer able to access the device via telnet or SSH.
So, what I've done.
1) Downloaded and installed ffp as per the instructions here: http://www.inreto.de/dns323/fun-plug/0.5/
2) Telnet into the system and execute the following commands:
passwd (to change root password)
usermod -s /ffp/bin/sh root (to use ffp shell)
store-passwd.sh (to store the password)
Reboot
Can't telnet in.
3) Reinstall ffp to get telnet access again. (no password has been stored)
4) passwd
usermod -s /ffp/bin/sh root
cd /ffp/start
chmod 755 sshd.sh
store-passwd.sh
Reboot
No Telnet nor ssh access. I can still access the web interface and can navigate to the folder.
So I went even more basic.
5) reinstall ffp to get tenet access
6) passwd
then rebooted. No telnet access.
In short, everything is working fine as long as I don't reboot the box. It appears that as soon as I do a store-passwd.sh the settings are not saved. When I execute store-passwd.sh I get no errors. I've even manually executed each command in the store-passwd.sh script and everything seems to work just fine.
Any ideas.
Thanx in advance for any help.
Offline
So to complicate the issue ...
over the past several days I've been copying files over from my system. I tar them up, scp them over, then untar them on the DNS-343. From time to time the DNS just locks up. No apparent reason or error. My connection just terminates and I can't t access the system via telnet, ssh, or http. I have to physically press the power button to turn it off. When I do this the system says it's powering down but really isn't. I've left it for over 20 mins and it never shuts down. I eventually have to pull the plug...which of course corrupts the drives meaning they have to resync which takes literally 13 hours.
I'm very unimpressed overall.
Offline
I have similar problemes. My goal is to run rsnapshot on a remote DNS-343 to provide off-site backup of my DNS-323, but everything beyong the fun_pug (v.0.5 by fonz) seems to create problems.
Installing the fun_plug works just fine, and telnet-ing too. However, rsnapshot requires Perl, which I install before installing rsnapshot itself. I use the Perl version that is part of fun_plug 0.5, and rsnapshot 1.3.0 provided by manilalink here: http://dns323.kood.org/forum/t2348-Rsna … ackup.html
For some strange reason, it seems that Perl may be part of the problem, because typing Perl on the telnet command line results in a massive hang requiring a physical un_plug to get things straight again. Executing rsnapshot gives various results, such as "Segmentation fault", the massive hangover (un_plug), and also successful runs. Quite inconsistently, though.
So, I am very interested in a resolution to this issue - even if it implies that I have to crosscompile fun_plug, Perl and rsnapshot (I say "even if" because I really don't know how to do that, but somehow I will if I have to).
The DNS-343 runs on firmware 1.02 (and the 323 is on 1.05).
Offline
Update: I forgot that I had commented out a line in the fun_plug script: the line that removes the fun_plug tarball after install. Thus, fun_plug was reinstalled at each boot, which explains why Telnet kept working between boots. When the fun_plug tarball is gone, so is Telnet after a reboot.
Looking at the ffp.log file, it seems that there is something fundamentally wrong with the fun_plug itself, as Segmentation fault is not something you want...:
ln -snf /mnt/HD_a2/ffp /ffp
* Running /ffp/etc/fun_plug.init ...
Segmentation fault
* Running /ffp/etc/rc ...
/ffp/bin/sh: '/ffp/lib/' is not an ELF file
Segmentation fault
* OK
Although I don't know much about Linux, I'd guess that the DNS-323 fun_plug source is not correctly compiled for the DNS-343, or during build there are some dependencies that differ between the two NAS'es, rendering the fun_plug unusable on both boxes at the same time.
Does anyone know how to work around this?
Offline
bump
same here..
Offline
Hey, got the same sort of problem here, but for me after rebooting telnet remains functional (unless I disable it...) while ssh'ing in just 'access denied's me...its kinda the same, as though none of the settings are being stored.
I used the tutorial here - http://nas-tweaks.net/CH3SNAS:Tutorials/fun_plug - and consequently posted on their support forum -
http://forum.nas-tweaks.net/modificatio … ddons/ssh-(access-denied)-and-telnet-andgtlogin-(incorrect-login)-after-reboot/ (yes it's one big link)
- which lead me to http://www.shadowandy.net/2008/07/fun_p … ns-343.htm so hopefully one of these avenues will lead to a solution
and ya, sorry for link spam
Offline
cliaz wrote:
Hey, got the same sort of problem here, but for me after rebooting telnet remains functional (unless I disable it...) while ssh'ing in just 'access denied's me...its kinda the same, as though none of the settings are being stored.
I am having the same problems also. I should have searched a little harder first. I have posted in another thread
http://dns323.kood.org/forum/p23698-Tod … tml#p23698
But it should probably be moved.
The problem seems to be caused by the 343 retrieving its "default config" from a different location than the 323.
What we need to try and figure out is where the default config is coming from, and how we can save changes back to it.
Offline
Found a possible solution to the DNS-343 problem where it fails to store the root account and password when the device is powered off.
http://dns323.kood.org/forum/viewtopic. … 702#p23702
Offline
Attached is my adjusted store-passwd.sh file.
It does some _very_ basic checking to determin if you have a DNS-323 or a DNS-343 and then copies the group, passwd and shadow files to the correct location for your device.
Essentially it just checks if you have 128Mb of system ram. As that is what the 343's seem to have, the 323 and 321 have ~64mb.
I am looking for some more cross references to make its model checking a little more robust.
I think the 323 and 321 have slightly slower CPU's so if someone with a 321 or 323 could post the details of `cat /proc/cpuinfo` i might be able to make the checking a little more reliable.
Offline
OneArmedMan wrote:
Essentially it just checks if you have 128Mb of system ram. As that is what the 343's seem to have, the 323 and 321 have ~64mb.
Purpose of store-passwd.sh is to overwrite existing copies of the files. How about checking the various known places for passwd and only copying the files if there already exists a copy?
Offline
OneArmedMan wrote:
if someone with a 321 or 323 could post the details of `cat /proc/cpuinfo` i might be able to make the checking a little more reliable.
The cpuinfo of the CH3SNAS: http://nas-tweaks.net/CH3SNAS:CPUINFO
fonz wrote:
How about checking the various known places for passwd and only copying the files if there already exists a copy?
Exactly what i thought. I did a quick rewrite: http://nas-tweaks.net/CH3SNAS:Store-passwd.sh
Works flawlessly on my CH3SNAS, how about the DNS-343 and DNS-323?
Offline
awesome!!
now if it can be figured out why every command segfaults after a few minutes, i'll be deliriously happy.
also wanted to point out a different thread on a slightly tangental topic, apparently the firmware building (assembling??) tools have been updated for the DNS-343, too? http://dns323.kood.org/forum/viewtopic. … 702#p23702
I'm not smart enough to have a crosscompiling environemnt (or even know if thats the point of this), so i haven't tried the new tools or anything, but if this info helps get the DNS-343 ffp stuff up to snuff, then yay.
Thanks to all you smart folks for the fixes!
-m
Offline
/ # store-passwd.sh
/ffp/sbin/store-passwd.sh: BACKUPLOCATION[0]=/sys/mtd1: No such file or directory
/ffp/sbin/store-passwd.sh: BACKUPLOCATION[1]=/sys/mtd2: No such file or directory
/ffp/sbin/store-passwd.sh: BACKUPLOCATION[2]=/mnt/HD_a4/.systemfile/AccountFile: No such file or directory
/ffp/sbin/store-passwd.sh: BACKUPLOCATION[3]=/mnt/HD_b4/.systemfile/AccountFile: No such file or directory
/ffp/sbin/store-passwd.sh: BACKUPFILE[0]=/etc/shadow: No such file or directory
/ffp/sbin/store-passwd.sh: BACKUPFILE[1]=/etc/group: No such file or directory
/ffp/sbin/store-passwd.sh: BACKUPFILE[2]=/etc/passwd: No such file or directory
/ffp/sbin/store-passwd.sh: BACKUPFILE[3]=/etc/samba/smbpasswd: No such file or directory
mount: mounting /dev/mtdblock0 on /sys/mtd1 failed: Device or resource busy
mount: mounting /dev/mtdblock1 on /sys/mtd2 failed: Device or resource busy
/ffp/sbin/store-passwd.sh: 30: Syntax error: Bad substitution
/ # ls /sys/
crfs custom mtd1 mtd2
on execution, the script threw the above errors, although the files/folders existed
think it's just a sh vs bash thing, on sh not liking []'s
so, after installing bash
/ # bash store-passwd.sh
mount: mounting /dev/mtdblock0 on /sys/mtd1 failed: Device or resource busy
mount: mounting /dev/mtdblock1 on /sys/mtd2 failed: Device or resource busy
shadow found in /sys/mtd1, copyingdone
shadow found in /sys/mtd2, copyingdone
shadow found in /mnt/HD_a4/.systemfile/AccountFile, copyingdone
shadow found in /mnt/HD_b4/.systemfile/AccountFile, copyingdone
group found in /sys/mtd1, copyingdone
group found in /sys/mtd2, copyingdone
group found in /mnt/HD_a4/.systemfile/AccountFile, copyingdone
group found in /mnt/HD_b4/.systemfile/AccountFile, copyingdone
passwd found in /sys/mtd1, copyingdone
passwd found in /sys/mtd2, copyingdone
passwd found in /mnt/HD_a4/.systemfile/AccountFile, copyingdone
passwd found in /mnt/HD_b4/.systemfile/AccountFile, copyingdone
smbpasswd found in /sys/mtd1, copyingdone
smbpasswd found in /sys/mtd2, copyingdone
smbpasswd found in /mnt/HD_a4/.systemfile/AccountFile, copyingdone
smbpasswd found in /mnt/HD_b4/.systemfile/AccountFile, copyingdone
Flushing unwritten filesystem I/O buffers...done
Backup complete
not sure why you(s) can run it as
/ # store-passwd.sh
while I have to run it as
/ # bash store-passwd.sh,
and why the script specifies using sh instead of bash and somehow works, but meh, it works!
cheers guys
Offline
I just recently picked up a DNS-343 and am running into the same darn problem!
I have 4x250GB drives. Upgraded to Firmware v1.02.
I've copying the ffp 0.5 fun_plug and fun_plug.tgz to the Volume_1 (which is /mnt/HD_a2)
fun_plug.log reports....
**** fun_plug script for DNS-323 (2008-08-11 tp@fonz.de) ****
Fri Jul 17 11:00:25 CDT 2009
ln -snf /mnt/HD_a2/ffp /ffp
* Installing /mnt/HD_a2/fun_plug.tgz ...
tar: WARNING: Converting hard link to symlink
... Lots of these, so I've trimmed them from this message..
tar: WARNING: Converting hard link to symlink
* OK
* Running /ffp/bin/busybox...
* Running /ffp/etc/fun_plug.init ...
* Running /ffp/etc/rc ...
* /ffp/start/syslogd.sh inactive
* /ffp/start/SERVERS.sh inactive
* /ffp/start/portmap.sh inactive
* /ffp/start/unfsd.sh inactive
* /ffp/start/nfsd.sh inactive
* /ffp/start/ntpd.sh inactive
* /ffp/start/LOGIN.sh inactive
* /ffp/start/telnetd.sh ...
Starting /ffp/sbin/telnetd -l /ffp/bin/sh
* /ffp/start/sshd.sh inactive
* /ffp/start/rsyncd.sh inactive
* /ffp/start/mediatomb.sh inactive
* /ffp/start/kickwebs.sh inactive
* /ffp/start/lighttpd.sh inactive
* /ffp/start/inetd.sh inactive
* OK
Going through all the steps of the wiki to set it up work fine. SSH works. I can login and the password is correct.
The next time I reboot,
**** fun_plug script for DNS-323 (2008-08-11 tp@fonz.de) ****
Fri Jul 17 11:11:11 CDT 2009
ln -snf /mnt/HD_a2/ffp /ffp
* Running /ffp/bin/busybox... (I added this line.
* Running /ffp/etc/fun_plug.init ...
* Running /ffp/etc/rc ...
Segmentation fault
* OK
So I installed BASH (needed to because of the var[0] stuff) and ran that (normal results) and then rebooted to see the same thing above about segfaults...
No Telnetd, no SSHd. Something is still not being saved. Life is sad.
Any thoughts/help?
-Ben
p.s. recopying fun_plug.tgz puts me back to square 1.
p.p.s. I also have:
/mnt # ls -l
drwxrwxrwx 7 root root 4096 Jul 17 16:34 HD_a2
drwxrwxrwx 5 root root 4096 Jul 17 16:32 HD_a4
drwxrwxrwx 5 root root 4096 Jul 17 16:32 HD_b4
drwxrwxrwx 4 root root 4096 Jul 17 16:32 HD_c4
drwxrwxrwx 4 root root 4096 Jul 17 16:32 HD_d4
drwxr-xr-x 2 root root 1024 Feb 16 2007 web_page
So when the new "store-password.sh" file is copying out of places like A4 and B4, do I need to include C4 and D4?
Last edited by bkamen (2009-07-17 19:54:02)
Offline
Use the top store-passwd.sh from here: http://nas-tweaks.net/CH3SNAS:Store-passwd.sh and read through this: http://tinyurl.com/ndhs6z
Offline
ok, did that and still having problems.
I'm going to try the upgrade to 1.03 since that seemed to fix it for some... and I'll get back to this list from there.
-Ben
Offline
Well, interestingly enough -- without changing a thing except upgrading to 1.03 (beta)...
It works.
Go figure.
Thanks for the help though!
-Ben
p.s. Here's a before and after of ffp.log for those who might want a sanity check
**** fun_plug script for DNS-323 (2008-08-11 tp@fonz.de) ****
Fri Jul 17 12:32:41 CDT 2009
ln -snf /mnt/HD_a2/ffp /ffp
* Running /ffp/bin/busybox...
* Running /ffp/etc/fun_plug.init ...
Segmentation fault
* Running /ffp/etc/rc ...
Segmentation fault
* OK
**** fun_plug script for DNS-323 (2008-08-11 tp@fonz.de) ****
Fri Jul 17 20:24:32 CDT 2009
ln -snf /mnt/HD_a2/ffp /ffp
* Running /ffp/bin/busybox...
* Running /ffp/etc/fun_plug.init ...
* Running /ffp/etc/rc ...
* /ffp/start/syslogd.sh inactive
* /ffp/start/SERVERS.sh inactive
* /ffp/start/portmap.sh inactive
* /ffp/start/unfsd.sh inactive
* /ffp/start/nfsd.sh inactive
* /ffp/start/ntpd.sh inactive
* /ffp/start/LOGIN.sh inactive
* /ffp/start/telnetd.sh ...
Starting /ffp/sbin/telnetd -l /ffp/bin/sh
* /ffp/start/sshd.sh ...
Starting /ffp/sbin/sshd
* /ffp/start/rsyncd.sh inactive
* /ffp/start/mediatomb.sh inactive
* /ffp/start/kickwebs.sh inactive
* /ffp/start/lighttpd.sh inactive
* /ffp/start/inetd.sh inactive
* OK
Last edited by bkamen (2009-07-18 06:57:31)
Offline
bump again, i just purchased my 2nd 343 last night and am having the exact same problem when following the same instructions that worked for my other 343 and my 323. i still don't see a solution but a lot of chatter. did i miss it?
i got so hosed up last night, i had to back down to b66 fw as the drives wouldn't even initialize after ffp got "corrupted". right now i have a fresh set of drives, and b70 code in a R5 so i'd like to do this one more time and have it work
Last edited by drick1 (2009-07-18 18:39:46)
Offline
FFP didn't work correctly on FW version lower than 1.0.3bXX because of the kernel or the toolchain on the older versions.
I'm not sure which one, I'm still learning about linux and the DNS-343 slowly.
All my problems with FFP went away when I switched to the beta firmware.
Drick1 are you flashing using network cable or doing it over wireless? Could be a bad flash.
Offline
Smallstack wrote:
FFP didn't work correctly on FW version lower than 1.0.3bXX because of the kernel or the toolchain on the older versions.
I'm not sure which one, I'm still learning about linux and the DNS-343 slowly.
All my problems with FFP went away when I switched to the beta firmware.
Drick1 are you flashing using network cable or doing it over wireless? Could be a bad flash.
i'm flashing over cable.
what i found is that b70 got hosed, i downgrade to b67 then upgrade back to b70 which seemed to do the trick
Offline