Custom 1.0.7.8 firmware is ready to download

[mumsoft]

Hi,

I'm doing the update of this great firmware the difficult way. I want to update my opt directory, not to remove it.
So I'm busy doing that and meanwhile I prepare sort of a howto for others.

Of course, if one of you has it complete already, please overrule me.

Marc
O, BTW: I like to THANK mister Furge for his excellent job!

[kfurge]

I'm doing the update of this great firmware the difficult way. I want to update my opt directory, not to remove it.
So I'm busy doing that and meanwhile I prepare sort of a howto for others.

Thank you Marc! I look forward to your how-to contributions.

- K.C.

P.S. I'm in the same boat you are. I have quite a substantial /opt on my own primary router that will be brought kicking and screaming to 1.0.7.8 soon.

[tuepen]

will i lose all my files on the box when i reflash/upgrading firmware?

[fireflash]

will i lose all my files on the box when i reflash/upgrading firmware?

99.9% of the time, no. There is always that .1% chance that something will go ker-plooie though. If you're the paranoid type, backup your goods.

[mumsoft]

will i lose all my files on the box when i reflash/upgrading firmware?

Not likely. But if you have a /shares/MYVOLUME1/opt folder and a running rc.local from your 1.0.4.6 KC Furge image, it is best to, BEFORE you do the update: mv /shares/MYVOLUME1/opt /shares/MYVOLUME1/opt_DO_NOT_LOAD
If you get a message about opt being busy
a) disable all own added software. Set # in front of such startups in rc.local, but at least leave the telnet stuff there unaltered.
b) reboot
c) mv /shares/MYVOLUME1/opt /shares/MYVOLUME1/opt_DO_NOT_LOAD (you need to have telnet running or this will not be possible)

I have updated my Asus without this and had several problems with not starting up, resetting etc. I even had to reset to factory defaults once. I blame the old rc.local.
But all my shares are untouched, music, pictures etc are not affected.

So take care, and wait a while till I finish my how to

Marc

[elgatoz]

First of all, I want to thank kfurge for doing such a great job. Without this custom image, this router has only small part its powers.

Now goes the question
I've successfully installed the image and set up MLDonkey. However, MLDonkey seems to eat all available CPU time:

  PID  PPID USER     STAT   VSZ %MEM %CPU COMMAND
  541     1 root     R    40340  65%  93% /opt/bin/mlnet.lite -allowed_ips 127.
  582   580 root     S N  40340  65%   3% /opt/bin/mlnet.lite -allowed_ips 127.
  191     1 root     S     4880   8%   1% upnp -D -L br0 -W eth0
  588   587 admin    R     1792   3%   1% top
    4     1 root     RWN      0   0%   0% [ksoftirqd_CPU0]
    5     1 root     SW       0   0%   0% [kswapd]
  580   541 root     S    40340  65%   0% /opt/bin/mlnet.lite -allowed_ips 127.

MLDonkey web interface loads for ages. It takes several minutes to fully load transfers page.

Is it normal, or I have misconfigured something?

i believe not, are you downloading files with many sources? or have you put files in your share folder and now its hashing? give more info's

[marcolino]

Two question:
1) Since mldonkey use a LOT of cpu resources would be very interesting having the possibility to start and stop it in a easy way. Stopping of course is only a matter of killing the core using Sancho, but restarting requires a telnet session and typing some annoying commands (not very comfortable).

2) I have the ftp server accessible from wan and would be very interesting to me knowing who is connected at that time, possibly in real time. Is there the possibility? Could I manage the user's connections?

Thanks in advance to anyone will help me.
(sorry for my ignorance and rusty english, but EXPERTS exists to be hopefully helpful)

[hweom]

i believe not, are you downloading files with many sources? or have you put files in your share folder and now its hashing? give more info's

I have ~10 files with >100 sources each.
I have set up share a day ago, so I suppose hashing should be done by now. CPU usage is sill up to 100%

[dalanik]

2) I have the ftp server accessible from wan and would be very interesting to me knowing who is connected at that time, possibly in real time. Is there the possibility? Could I manage the user's connections?

I wanted to ask the same thing! Or at least, be able to disable some accounts that have access to ftp?

D.

[elgatoz]

I have ~10 files with >100 sources each.
I have set up share a day ago, so I suppose hashing should be done by now. CPU usage is sill up to 100%

hmmm i think that is too much for the cpu
you can play with loop_delay to save cpuload and try to reduce disk access cause i think this is still the bottleneck playing with diskbuffer etc..
http://mldonkey.sourceforge.net/MemoryUsage maybe useful too

but imo this will not help thats just too much,
i started one big well seeded torrent ~40peers 500-700kb/s download= 80% cpuload, like to hear more mldonkey-experience on the wl700 here greetz

edit: setting share_scan_intervall=0 try this too

[monkey8]

*NEVER MIND THIS. Bear with me, i have hangovers and have only slept for like 5 hours :-P*
Your problem is that you have both the client_port and tracker_port set to the same value. The tracker_port defines what port MLDonkey will use if you want to run your own tracker, where the client_port is the port being used to download torrents. Just reset tracker_port to its default value (6881) and only change client_port.

have you tried to open the file starting with the sudo command? (Eg. "sudo nano bittorrent.ini")? Or change the permissions (sudo chmod bittorrent.ini 777)?

Thank you Henrikh!! this resolved my issue. I was in pain trying to get that fix and I only had 2 hrs of sleep that night.

Now, I have second problem. I have wl700ge behind my Linksys WRT54G (192.168.3.1). wl700ge is getting a static IP addr. of 192.168.3.100. Meaning, I have wl700ge treated as a client behind my WRT54G. before I mess up the ports 2 days ago, I was able to HTTP access MLdonkey by typing in 192.168.3.100:4080. The first time I had an error that asked me to add my PC's ip address (192.168.3.109) into mldonkey's 'allowed_ips'. I did, and it worked.

Now after i fixed the port issue, I can't access it anymore. i have even tried the 'nvram set kc_ssh_wan_access=Yes' still no go. I wonder what I did wrong? note, I can access WL700gE web Manager by just typing in 192.168.3.100 tho.. umm..

So, simply, hope the above arn't too complicated. I wish to access mldonkey's webui from outside... it was possible before, but now can't. Any suggestions?

thanks in advance.. and kfurge, keep up the great work! I appreciate what you offer to the community!

[kfurge]

The first time I had an error that asked me to add my PC's ip address (192.168.3.109) into mldonkey's 'allowed_ips'. I did, and it worked.

Interesting. You shouldn't have had to do this since the entire internal subnet 192.168.3.x is configured for full access by the startup script.

So, simply, hope the above arn't too complicated. I wish to access mldonkey's webui from outside... it was possible before, but now can't. Any suggestions?

What do you mean by outside? So far everything you mentioned has been "inside". I.e. on the LAN interface 192.168.3.x. Is this what you can no longer access?

Please log into the router via telnet or ssh and post the mldonkey startup output in /tmp/mldonkey.out.

- K.C.

[ice-cool]

Something must have happened when the file got copied... The 42000 GID is indicative of an Asus made user. Maybe you didn't copy it as root? The firmware image has the UID:GID for all files set to 0:0.

Fixing is easy with chgrp, though other sudo related files are probably also broken.

- K.C.

I have a problem same Neil

Downloading http://ipkg.nslu2-linux.org/feeds/optware/oleg/cross/stable/Packages.gz
wget: missing URL
Usage: wget [OPTION]... [URL]...

Try `wget --help' for more options.
An error ocurred, return value: 1.
Collected errors:
ipkg_download: ERROR: Command failed with return value 1: `wget    -q -P  http://ipkg.nslu2-linux.org/feeds/optware/oleg/cross/stable/Packages.gz'
WARNING: Unable to remove temporary directory: : No such file or directory

Sorry this is a dumb question and Im very newbie.
I don't know how to chgrp. ?
Could you give me more detail fro fix this problem ?

[monkey8]

Interesting. You shouldn't have had to do this since the entire internal subnet 192.168.3.x is configured for full access by the startup script.



What do you mean by outside? So far everything you mentioned has been "inside". I.e. on the LAN interface 192.168.3.x. Is this what you can no longer access?

Please log into the router via telnet or ssh and post the mldonkey startup output in /tmp/mldonkey.out.

- K.C.

Thanks for the prompt reply, kfurge.

mldonkey startup output in /tmp/mldonkey.out:

2008/01/29 19:05:44 [cO] Starting MLDonkey 2.9.2 ...
2008/01/29 19:05:44 [cO] Language EN, locale ASCII, ulimit for open files 1024
2008/01/29 19:05:44 [cO] MLDonkey is working in /shares/MYVOLUME1/MYSHARE1/.mldonkey
2008/01/29 19:05:44 [Gettext] Loading language resource mlnet_strings.EN_ASCII
2008/01/29 19:05:44 [cO] loaded language resource file
2008/01/29 19:05:44 [DNS] Resolving [WL700gE] ...
2008/01/29 19:05:45 [DNS] Resolving [www.mldonkey.org] ...
2008/01/29 19:05:45 [cO] Logging in /shares/MYVOLUME1/MYSHARE1/.mldonkey/mlnet.log
2008/01/29 19:05:49 [dMain] Core started

Here is my current connection:

Linksys (WAN IP= Public addr., LAN IP = 192.168.3.1)
|
|
|
+----- wl700ge (WAN IP= 192.168.3.100, LAN IP 192.168.1.1)
| |
| |
| |
| +------ System two (LAN IP = 192.168.1.100)
.
.
.

|
+----- System One (LAN IP = 192.168.3.109)

as you can see from the above diagram, from System two, I have no problem accessing mldonkey's http webui. However, I'm not able to from System One, which I was able to before. Any ideas?

thx.

[dalanik]

What about connecting to the Mldonkey web interface (or via sancho) from the outside (WAN)? Is that possible?