No automatic connect to ISP
-
Antiben,
Just tell me what to do, I'll try your mods over version 20100221-0133.
-
I have a Dell 1740. Send your mods! I'm ready
-
precondition complied,
so i sent you the file with a short instruction via pm. please reboot after that.
i soon post the changes, and some diffs for the devs.i did the update, rebooted and a short time after the i reload the pfsense welcome screen my
wan pppoe was up and hopefully runs until the next hardware or power failure comes gwhat features of 2.0 are you using? multi dyndns? multi-wan?
what type of internet-connection? (not only the pfsense hw) -
Oh sorry. At the moment its a simple and stupid config: Single-wan, single-lan. No opt, no packages. Only built-in features. My inet is a 100MBit-DSL.
At the moment it looks good. No disconnect since reboot. Only strange thing i have is a very weird connection. Endless delays and packet-loss. Due to that I have a really poor connection, but I'll try tomorrow with 1.2.3 to see if this changes.
But all relevant things work as expected for the moment: Dyndns-update occured and is shown right in the dyndns-page, on the dashboard wan-gateway tells me about the state of the gw, not only "gathering data" which was the only info before.
So at the moment your changes look very good! Thanks again! -
ok, thx for the feedback.
if you look into your '/var/etc/mpd_wan.conf', you notice that i re-add the no noretry bundle option, disabled the tcpmssfix, removed the ppp-linkdown call and set the max-redial to 7 instead of unlimited, cause if i enter a wrong password, my t-dsl account may locked.delays and packet-loss while browsing the web?
100MBit-DSL - that's not really poor?! gtry 1.2.3 and keep some system.log files…
you may want to test the 'old' mpd 3.18 in pfsense 2.0b (like i did... ^). -
Now the auto-reconnect works! After a while the line disconnected due to the packet-loss and it reconnected automagically. First time since long time! Very nice.
Normally my web is fast! But now all websites are as slow as ISDN. Not nice. Remote working is nearly impossible. :-(
Packet-loss is nearly by 50%.
-
Normally my web is fast! But now all websites are as slow as ISDN. Not nice. Remote working is nearly impossible. :-(
tcpmssfix may help, if it's disabled.
cheers,
-
Used the last patch from antiben with tcpmssfix enabled. Looks as if this was the problem!
Packet-delays and losses are gone completely.
Web is as fast as again, I can send Mails again (was impossible with tcpmssfix disabled) and much more things work right now again!
I'm excited again. Thanks again for all your help! Great appliance, great community! -
im also having such issues, can u plz guide what file needs modification to solve this?
-
I got this info from antiben:
just download this tgz file and upload it through the web interface (diagnostics/command prompt).
than simply run the following shell command (in top) and hit execute
Code:
tar xzf /tmp/antiben_mod.tgz -C / && rm /tmp/antiben_mod.tgzit's not tested, but i'm testing this now, so don't blame me if something goes wrong.
always backup important settings and please report/post/link your setups, as i asked before…Then patch for tcpmssfix:
since nocer has suggested to enable the tcpmssfix, so i re-added this option...
same procedure... just download this small tgz and run via diagnostics/command prompt
(you are still on the same build?)
Code:
tar xzf /tmp/tcpmssfix.tgz -C / && rm /tmp/tcpmssfix.tgzHope that helps you.
The patch is made for the snap from Feb 20/21. I have it working flawlessly with snap from Feb 22 04:29:20.Good luck!
antiben asked me to remove the links. So please better wait for an official fix.
-
im currently on the 23rd Feb nanobsd snapshot
-
as i wrote before (^), attached is a diff (for the most important things concerning auto-reconnect) against the 20100221-0133 snap…
(more infos soon) -
Any updates to this ??
-
i applied both the patches and rebooted, works fine so far but ill have to wait a bit more till my line disconnects, its been up for 14:17:28 so far, shall wait till 24hrs and if it doesnt disconnect then will plug the wire out to create a disconnect scenario and shall check if it reconnects on plugging it back in.
-
That would be cool if you could do a write up for this - As yeah, I have the same issue also. I have tried editing mpd.conf etc - but this has not worked.
Look forward to a solution. :)
-
after around 16hrs there was a disconnect from isp and it seems the reconnect was tried by mpd and successfully connected so for me its is just great so u can go ahead and apply the patch.
i wonder if this will be patched in main pfsense builds coz it seems to work just fine and also the connection up time field in interfaces is what i really like.
-
one thing i noticed was, with and without the patch, the periodic reset never works for pppoe
reset at each day ('0 0 * * *')
it never happens coz if it did then my WAN ip would change but it never does.
-
now all of a sudden when i login into the firewall and click interfaces i get this
Warning: shell_exec(): Unable to execute 'ifconfig -l' in /etc/inc/globals.inc on line 44 Warning: shell_exec(): Unable to execute 'ifconfig -m | grep "capabilities=.*HWTAGGING"' in /etc/inc/globals.inc on line 49 Warning: shell_exec(): Unable to execute 'ifconfig -l' in /etc/inc/globals.inc on line 44 Warning: shell_exec(): Unable to execute 'ifconfig -m | grep "capabilities=.*VLAN_MTU"' in /etc/inc/globals.inc on line 49 Warning: shell_exec(): Unable to execute 'uname -m' in /etc/inc/globals.inc on line 130 Warning: file_get_contents(/tmp/config.cache): failed to open stream: Too many open files in system in /etc/inc/config.lib.inc on line 115 Warning: session_start(): Cannot send session cache limiter - headers already sent (output started at /etc/inc/globals.inc:44) in /etc/inc/auth.inc on line 863
No page assigned to this user! Click here to logout.
-
Hrmm, doesn't sound good. It would be nice to hear something from the pfSense dev guys in order to get this fully fixed.
-
Is there an actual diff for these changes instead of replacing whole files? That would go a long way toward getting the problem fixed.
And I am not quite sure, does this fix involve the use of the old mpd 3.x binary or is it still mpd4? If so, it would be better to make it work with mpd5 instead of 3, as that brings other options to the table as well.
