Joined: 16 Jun 2006 Posts: 2 Location: Barsinghausen
Posted: Tue Jun 27, 2006 10:46 Post subject:
Alberto,
it seems that we running exactly the same config:
WRT54GL 1.1, WDS, WPA PSK -TKIP, one router connects to ISP and acts as DHCP server. I am running v23 SP1 from 05/16/06 (Std version). One wlan client connected to the second not ISP connected router.
I does not need to have high load on the wlan, the nas service dies after 10 to
30 mins but only on the router connected to the ISP.
While runnunig a telnet session I saw the nas terminating with a "segementation fault".
Joined: 06 Jun 2006 Posts: 7492 Location: Dresden, Germany
Posted: Tue Jun 27, 2006 11:10 Post subject:
i introduced a codechange in the build last night which hopefully solves this problem. _________________ "So you tried to use the computer and it started smoking? Sounds like a Mac to me.." - Louis Rossmann https://www.youtube.com/watch?v=eL_5YDRWqGE&t=60s
Joined: 07 Jun 2006 Posts: 198 Location: Parma - Italy
Posted: Wed Jun 28, 2006 14:07 Post subject:
Snoopy wrote:
I have installed the Multilanguage Mini generic from the 27.06.2006
Lets see how it works.
Thanks !
Snoopy
I'm testing this release too. First thing I noticed: I was thinking that the change was that nas was restarted if it died. Well, I kill it and it does not restart. Maybe BrainSlayer instead went the right way and now it will not die anymore.
Any hint about where the problem was?
same here nas wasn't running anymore and after restarting it with
/usr/sbin/nas /tmp/nas.lan.conf /tmp/nas.lan.pid lan
the users were able to connect again.
I wonder what to start when nas dies.
Himmi suggested
/usr/sbin/nas /tmp/nas.lan.conf /tmp/nas.lan.pid lan &
That does the job, but there is neither /tmp/nas.lan.conf nor /tmp/nas.lan.pid on my router. After a restart I'll find this on the ps output
nas -P /tmp/nas.wl0lan.pid -H 34954 -l br0 -i eth1 -A
I'm not sure whether this is the full line, as "nas -P /tmp/nas.wl0lan.pid -H 34954 -l br0 -i eth1 -A" takes all the lines available on my console.
While nas is still running, I'll get the following error on running the above line:
eth1: ignore i/f due to error(s)
I'd love to introduce some script in the startup that will keep nas running.
Himmi suggested:
#!/bin/sh
killall nas
while [ 1 ]
do
/usr/sbin/nas /tmp/nas.lan.conf /tmp/nas.lan.pid lan
sleep 5
#write something in the log
done
Alberto suggests:
If it is helpfull the right place for it would be immediately after booting, with
nvram set rc_startup="killall nvram..... &"
(mind the final &. Also there should be spaces around the 1 in [ 1 ]. Sorry if you are already a linux guru))
So, what now?
I create a file /tmp/keepnasalive with
#!/bin/sh
killall nas
while [ 1 ]
do
nas -P /tmp/nas.wl0lan.pid -H 34954 -l br0 -i eth1 -A
sleep 5
#write something in the log
done
and I'll then execute once:
nvram set rc_startup="killall nvram; /tmp/keepnasalive &"
Is this "killall nvram" really correct??
Is "nas -P /tmp/nas.wl0lan.pid -H 34954 -l br0 -i eth1 -A" the correct command line?
Is there a problem restarting it over and over again? Will the PID take care of that? Should there be a grep nas on the ps output??
I'm terribly happy, that the solution is so near now. I'd just love this last input before I dare excute killall and put something in the startup script for the first time.
I had random PING timeouts with stock firmware, dd-wrt, hyperwrt and tomato. All use the same pppd (v2.4.1)
Then I installed OpenWRT, which uses pppd v2.4.3 and the problems went away! Please update the PPPOE of DD-WRT because it really caused me a lot of headaches!
Joined: 07 Jun 2006 Posts: 198 Location: Parma - Italy
Posted: Wed Nov 01, 2006 10:16 Post subject:
VPN-User wrote:
Another approach:
I had random PING timeouts with stock firmware, dd-wrt, hyperwrt and tomato. All use the same pppd (v2.4.1)
Then I installed OpenWRT, which uses pppd v2.4.3 and the problems went away! Please update the PPPOE of DD-WRT because it really caused me a lot of headaches!
I am on OpenWRT for now because this works.
So what? We are talking about wpa.
JackHambabo: at the end I went back to wep. Since then I never had the time to check if SP2 was better. By the way this is something I should have a look at, since wep is not very safe.
Another info: I bought a Linksys WVC54G wireless camera, which works perfectly with wep but crashes after several hours with wpa and requires a power cycle. I guess the problem might be the same.
Where can I find the start up script that starts nas.
even with no nas running i get the above mentioned error
eth1: ignore i/f due to error(s)
on running
nas -P /tmp/nas.wl0lan.pid -H 34954 -l br0 -i eth1 -A
Moreover, I just tried to run the script that I put together from Himmi and Alberto. Before starting it, I wanted to run the commands by hand. I run nas -P /tmp/nas.wl0lan.pid -H 34954 -l br0 -i eth1 -A several times. To my surprise this started nas serveral times. I thought the script would take care of not restarting nas twice.
Oh, I just get it! There is no & at the end of the line so the script halts till nas ends. k. great.
Anyways, I then ran killall nas and that made me loose my wds connection to the first router. . Is killall nas the correct command.
And: If I want to put this script in the startup. Where can I save it, for it to be executable? The script I put in /tmp was deleted after a restart. (in contrast to all the other files in there. But maybe they are really just tmp-files and get created everytime.)
So: What's that correct nas command not yielding in any errors? And is killall nas correct?
It seems I've got the same problem.
I've got 2 routers : an Asus WL500G and a WRT54GLv1.1.
I've flashed the 2 routers with the VOIP V23SP2 version.
The Asus is connected to Internet throw ADSL, and is DHCP server.
The Linksys as Internet Settings set to Disable, DCHP is in transfert mode (to the ASUS).
WLAN is set with WPA-PSK TKIP.
I've got a WDS connection.
After some hours (or minutes in some cases), the Linksys don't respond for Wifi connection. After a Reboot, the connections come back.
The Asus have not this problem.
An other strange thing (probably normal) : when I use factory default on the Linksys, the Cisco logo is unlighted. When I put Internet Settings to disable, the Cisco logo become lighted.
when the problem occurs the next time, try to telnet into the wrt54g and excute ps in order to see whether nas is running. If your problem is also due to nas being crashed, all we have to do find someone who can tell us how precisely the script has to look like that restarts nas, and we've got a workaround.
I've got this problem for half an eternity now (I guess more then a year - though it's never been as bad as it is now. A year ago even the nightly started wasn't needed as the problem would occur only once a week or so), and I've never been that close to a solution.
In contrast to
nas -P /tmp/nas.wl0lan.pid -H 34954 -l br0 -i eth1 -A
which gives my the above mentioned error, I tried
/usr/sbin/nas /tmp/nas.lan.conf /tmp/nas.lan.pid lan &
this time again.
To my surprise I had both files today??? Dunno, whether I over looked them the last time?
nas.lan.pid reads:
20329
and that's the ps-id nas gets started under.
I'm not sure, what this nas does. As it seems the router still sends out the ssid over wlan and lets you connect incase you have the right passwort - even with nas down. you just won't get any data. But maybe thats the same behaviour as one would get on a wrong password, dunno.
So I suppose I've got the complete commandline for nas now, even in case files should vanish again.
But I'm still unsure
1) how to add a start up script without maybe deleting what's been in start up before.
2) where to put my script in case /tmp gets deleted on a regular basis
3) whether I can really start nas with the above parameter from /tmp/nas.lan.conf or whether nas.lan.conf is just an output file that is generated on starting /usr/sbin/nas /tmp/nas.lan.conf /tmp/nas.lan.pid lan & which fits with the date and time of creation of the files.
Oh, I guess the later is the case, isn't it? So my commandline for nas should really by
/usr/sbin/nas /tmp/nas.lan.conf /tmp/nas.lan.pid lan &
and non of the file need to exist before executing the command.
Joined: 07 Jun 2006 Posts: 198 Location: Parma - Italy
Posted: Mon Nov 06, 2006 17:00 Post subject:
JackHambabo wrote:
But I'm still unsure
1) how to add a start up script without maybe deleting what's been in start up before.
nvram set rc_startup= "/jffs/yourscript &"
nvram commit
Quote:
2) where to put my script in case /tmp gets deleted on a regular basis
enable jffs and put the script in /jffs
(have a look in the wiki. All this infos are there)
Quote:
3) whether I can really start nas with the above parameter from /tmp/nas.lan.conf or whether nas.lan.conf is just an output file that is generated on starting /usr/sbin/nas /tmp/nas.lan.conf /tmp/nas.lan.pid lan & which fits with the date and time of creation of the files.
Don't know. I remember having problems restarting nas. I was in a hurry so I reverted to wep and never had the time to test since then.
Hope you are able to get good results.