I am experiencing similar but not exact problem with 2.2.rc2:

NO VMWare...

P4 2.0Ghz, 1GB Ram, 80GB IDE
5 Nics

Green/Orange/Blue/Red(Main Uplink)/Red(2nd Uplink)
192.168.2.1/192.168.3.1/192.168.4.1/192.168.1.139/10.28.210.251

With Main Uplink and 2nd Uplink in Managed mode, unplugging Main Uplink
causes 25 second delay (25 second packet loss) before failover routes
traffic over 2nd Uplink.

With Main Uplink and 2nd Uplink Managed unchecked, I can uncheck active next
to Main Uplink and there is no packet loss as traffic routes over 2nd Uplink
immediately.  Manual ticking Active next to either link is instantaneous.

What is causing delay when in managed mode? Is this a bug or a setting
issue?  How to resolve?


Ruald Andreae wrote:
> 
> I think you are correct. Outgoing connections will load balance, it 
> seems. incoming connections will load balance with some kind of dns load 
> balancing or and internal system like heartbeat. How would dyndns work 
> with multiple connections though? Which interface ip will it use to 
> update dyndns with?
> 
> May I ask where is the docs for the new 2.2 version. The online docs 
> seems to be still for the old one. I've been playing around with the 
> beta version and especially with the failover feature. Pretty much the 
> main reason I would use endian. I have a vmware vm built with the 
> following network setup:
> 
> Green: Connected to a NAT vmnet and allows my host machine to talk to 
> the firewall
> Red1: Connected to a Bridged vmnet and allows my firewall to talk to the 
> net via the local lan's dhcp
> Red2: Connected to a Bridged vmnet and allows my firewall to talk to the 
> net via the local lan's dhcp as a second link
> Red3: Connected to a Bridged vmnet and allows my firewall to talk to the 
> net via the local lan's dhcp as a third link
> 
>  From I tests it looks like bandwidth load balancing is working ok. How 
> does it do the load balancing exactly? per tcp connection?
> 
> How does the failover work, or is it still in progress as it does not 
> seem to be working well. When I disconnect one of the virtual nic's in 
> vmware the firewall does not detect that the link is down. In fact you 
> have to manually tick "manage" and then untick "active". I can then 
> again tick manage and the link will detect that it is down and keep 
> attempting to reconnect until I re-enable the vmnic. When the main link 
> is down the 2nd link will then take the load and pings will continue, 
> but when i manually disconnect the 2nd link, leaving 3rd link the only 
> active one failover will not happen. I have to then manually disconnect 
> the 3rd link and reconnect it and pings will continue.
> 
> 
> Should all of this not be automatic or am I missing something? While 
> playing with the links in non-managed mode failover is seamless. I can 
> drop link 1, no packet loss, drop 2, no packet loss, drop 3, no 
> connection as expected, reconnect 1 or 2 or 3 and connection back again. 
> I'm even when I change the red nics to static and disable the nic in 
> vmware it still does not detect that it has gone down. how does endian 
> actually detect a down link? I tried doing ifconfig eth1 down and it 
> still does not detect that it's gone down.
> 
> Thanks
> 
> 
> 
> 
> Gregory Machin wrote:
>> Allie Syadiqin wrote:
>>   
>>> Hi, I intend to try and install Endian Firewall 2.2 but I need help 
>>> understanding the load balancing feature as there is really not much 
>>> info about it (or I probably just don't understand what I am reading 
>>> in the
>>> documentation :P ).
>>>
>>> Anyway, assuming that I have 2 webservers, both running the same 
>>> sites, with different internal IP addresses (kind of a redundant 
>>> setup), can the Endian Firewall load balance the external traffic 
>>> going to the webservers?
>>>
>>> Webserver 1 : Listening on IP 10.1.1.2 <http://10.1.1.2/> port 80
>>> Webserver 2 : Listening on IP 10.1.1.3 <http://10.1.1.3/> port 80
>>>
>>> Basically, what I am asking is whether using Endian Firewall 2.2 load 
>>> balancing feature eliminate me from having 2 separate dedicated 
>>> high-availability load-balancers (Heartbeat/HAProxy)  behind the
>>> firewall.
>>>
>>> Thanks and hope someone can enlighten me.
>>> ------------------------------------------------------------------------
>>>
>>> -------------------------------------------------------------------------
>>> This SF.net email is sponsored by the 2008 JavaOne(SM) Conference 
>>> Don't miss this year's exciting event. There's still time to save $100. 
>>> Use priority code J8TL2D2. 
>>> http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone
>>> ------------------------------------------------------------------------
>>>
>>> _______________________________________________
>>> Efw-user mailing list
>>> Efw-user@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/efw-user
>>>   
>>>     
>> By my understanding of the text the load balancing is for balancing lan 
>> connectivity to the internet ie u have 2 adsl lines and want to spread 
>> the load of interested access across the two lines. to load balance 
>> between to http servers or any other for that matter would require dns 
>> load balancing or one incoming line connecting to a load balancing 
>> server in front of the servers. I'm open to correction but thats the 
>> short story ..
>>
>>   
> 
> -------------------------------------------------------------------------
> This SF.net email is sponsored by the 2008 JavaOne(SM) Conference 
> Don't miss this year's exciting event. There's still time to save $100. 
> Use priority code J8TL2D2. 
> http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone
> _______________________________________________
> Efw-user mailing list
> Efw-user@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/efw-user
> 
> 

-- 
View this message in context: 
http://www.nabble.com/Understanding-Endian-Load-Balancing-Feature-tp16745629p19136126.html
Sent from the efw-user mailing list archive at Nabble.com.


-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Efw-user mailing list
Efw-user@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/efw-user

Reply via email to