2 x RED, 1 x Green - Dual WAN Problem.

[Guest]

[jump]

Hi Endians

Many Thanks for this distribution - looks awesome.

My Question resp. my problem is as follows:

I have 3 ports (eg. 3 NICs) on my system;
2 are RED / WAN (DHCP each) eth1 and eth2 - idea is to have two 2 isps for redundancy - if one is down,
the other one is still up...

1 is Green / LAN with all my clients in there. eth0

*****
Now, if i configure one Red (eth1 f.ex.), the green one eth0 and fire up, it works as a charme:
dhcp acquisation on red eth1 is done, i can ping and trace to wherever i want.
Now, as EFW 2.2-Rc1 allows me to have more than one red interface (outstanding!) i configure
through Network?>Interfaces>Create an uplink another uplink with eth2.

I then configure the type (Ether DHCP) say, that it is enabled, started on boot.
I active "if this uplink fails activate "main..." the update.
Same on eth1 vice versa (if uplink fails activate uplink1), update.

Problem 1:
The uplinkeditor then stalls (uplinkeditor.cgi is at 100% - top/shell) on 2 tested systems - all
the same - only a hard kill helps, uplinkeditor stays unreachable ever after.

Problem 2:
the uplinks do not take over automatically when i f.ex. take off eth1 redethercable for testing.
eth2 is connected, but i no longer can ping/trace/web. if i de-activate manually (through system/home)
the main uplink on eth1 and vice versa, it works, but not automatically eg. if the main uplink fails
(simulated through a disconnected cable), the uplink1 should automatically take over, eth2 is disonnected
such way, and eth1 is connected it should work - but it doesn't here with me.

Or, f.ex. both red are connected (both with a diffrent ip from dhcp) - if main is disconnected, i'm no longer
able to ping, trace.... allthough uplink1 is connected. Etc..

I entered some routes for testing, but that won't change it.

I have tested this on 2 diffrent systems (with diffrent standard harware) - 2 times the same.

What am I doing wrong? Is there any howto? does this work somwhere? is this supposed to work
in this release? I saw, that the paid devices allready are on 2.2 (final?) does this work there?

Is there a fix release schedule for 2.2 community stable or any new rcs?

Many thanks in advance for any replies or hints - anything concerning this toppic is appreciated.

Best regards
jump

[c4rdinal]

What is the System status of your Primary and Secondary Link?
And click the "Managed" option on the two interfaces.

Also, ensure that you have selected the right interface for your uplinks in the Network Configuration.

HTH

[jump]

Hi HTH - Thanks for replying...

What is the System status of your Primary and Secondary Link?
And click the "Managed" option on the two interfaces.

The status of both links is connected, both with their respective ip-adresses - everything seems ok.
I activated "managed" (once through the uplinkeditor, and for testing also on the system status screen).

Also, ensure that you have selected the right interface for your uplinks in the Network Configuration.

HTH

Allthough the interface naming changed somwhere (eth0 became eth2, eth2 bacame eth1 and eth1 became eth2) i'm convinced, that i choosed the right interfaces.
I'm allways able to ping eth2 (my lan/green port), ifconfig represents the right values compared to the web-gui.

I'm going to reinstall the whole thing today from scratch to re-test.


sofar
jump

[jump]

Hi endians

Oh great lord of the interfaces and electrons - made it... it works - what a relief

So, what i did in a few short steps:

1) Reinstalled 2.2RC1 from scratch. eth0 is green eth 1 is red1, eth2 is unconfigured
2) created an additional uplink (uplink1) with eth2 - active/managed
3) told mainuplink to fire up uplink1 if itself fails. managed.
4) told uplink1 to fire up mainuplink if itself fails => uplinkeditor.cgi stalls
5) found 2 corrective rpms on efw-mantis, that i downloaded and installed
6) rebooted
7) uplinkeditor does not stall anymore ! on problem less.
did a test - not working mainupllink does not fall to uplink1
9) deconfigured backuplink on uplink1 (no fall to mainuplink)
10) did a reboot
11) it works: when i disconnect the mainuplink, it automatically (after some time) takes uplink1
12) if the mainuplink gets back, it "routes" all traffic back through this interface.

After all i saw in efw-mantis, some bugs (uplinkseditor does not show activated backuplink - .js error etc..), are corrected, and i'm now convinced, that i they will hopefully soon flow into another public rc or maybe even a final...

Regards and thanks to HTH
jump

[pillin_slk]

Hi endians

Oh great lord of the interfaces and electrons - made it... it works - what a relief

So, what i did in a few short steps:

1) Reinstalled 2.2RC1 from scratch. eth0 is green eth 1 is red1, eth2 is unconfigured
2) created an additional uplink (uplink1) with eth2 - active/managed
3) told mainuplink to fire up uplink1 if itself fails. managed.
4) told uplink1 to fire up mainuplink if itself fails => uplinkeditor.cgi stalls
5) found 2 corrective rpms on efw-mantis, that i downloaded and installed
6) rebooted
7) uplinkeditor does not stall anymore ! on problem less.
did a test - not working mainupllink does not fall to uplink1
9) deconfigured backuplink on uplink1 (no fall to mainuplink)
10) did a reboot
11) it works: when i disconnect the mainuplink, it automatically (after some time) takes uplink1
12) if the mainuplink gets back, it "routes" all traffic back through this interface.

After all i saw in efw-mantis, some bugs (uplinkseditor does not show activated backuplink - .js error etc..), are corrected, and i'm now convinced, that i they will hopefully soon flow into another public rc or maybe even a final...

Regards and thanks to HTH
jump

its works?
works with Endian 2.5?