OLD STYLE works, new pools don't



  • Hi all, I'm stuck with a problem on a loadbalancing system.

    This system has been installed with 1.0.1. Recently I've reinstalled it with the lastest 1.0.1 snapshot (1.0.1-SNAPSHOT-03-27-2007) and then came the problem.

    On this system I have two WAN, the first one is on a pysical interface(bge0), the second one is on a vlan (vlan0).
                   WAN1(wan) : gateway  10.254.254.1/24
                   WAN2(opt3) : gateway 192.168.0.249/24

    On the 1.0.1 platform the pool looks like:
    10.254.254.1|10.254.254.1
    192.168.0.249|192.168.0.249

    All is fine trafic is balanced.

    On the 1.0.1-SNAPSHOT-03-27-2007 the pool looks like:
    wan|10.254.254.1
    opt3|192.168.0.249

    Trafic only goes through wan not opt3.

    First thing: If I use the configuration backup of the 1.0.1 platform, the pool is recreated and the logs are saying:
    "php: : Setting up route with 192.168.0.249 om vlan0 for monitor 192.168.0.249 on gateway 192.168.0.249"
    "php: : SLBD pool MULTIWANS is old style. Please recreate"
    "php: : Setting up route with 10.254.254.1 om bge1 for monitor 10.254.254.1 on gateway 10.254.254.1"
    But it does balance….all is ok except this warning.

    Second thing: When using newly created pool (so when trafic isn't balanced) logs are only saying:
    "php: : Setting up route with 10.254.254.1 om bge1 for monitor 10.254.254.1 on gateway 10.254.254.1"
    there is no line for the second wan.

    a full backup file shows:

    1.0.1-SNAPSHOT-03-27-2007 (the one not ok)
    <load_balancer><lbpool><type>gateway</type>
    <behaviour>balance</behaviour>
    <monitorip>10.254.254.1</monitorip>
    <name>MULTIWANS</name>
    <desc>Balance load over two wans</desc>
    <port><servers>opt3|192.168.0.249</servers>
    <servers>wan|10.254.254.1</servers></port></lbpool></load_balancer>

    1.0.1 (the one ok):

    <load_balancer><lbpool><type>gateway</type>
    <monitorip>192.168.0.249</monitorip>
    <name>MULTIWANS</name>
    <desc>Balance load over two wans</desc>
    <port><servers>10.254.254.1|10.254.254.1</servers>
    <servers>192.168.0.249|192.168.0.249</servers>
    <monitor>ICMP</monitor></port></lbpool></load_balancer>

    SLBD logs in the two cases:

    Apr 23 18:29:40 slbd[1719]: ICMP poll succeeded for 192.168.0.249, marking service UP
    Apr 23 18:29:38 slbd[1719]: ICMP poll succeeded for 10.254.254.1, marking service UP
    Apr 23 18:29:36 slbd[1719]: VIP 127.0.0.1:666 added real service 192.168.0.249:666
    Apr 23 18:29:36 slbd[1719]: VIP 127.0.0.1:666 added real service 10.254.254.1:666
    Apr 23 18:29:36 slbd[1719]: VIP 127.0.0.1:666 sitedown at 127.0.0.1:666
    Apr 23 18:29:36 slbd[1719]: VIP 127.0.0.1:666 configured as "127.0.0.1"
    Apr 23 18:29:36 slbd[1719]: Using configuration file /var/etc/slbd.conf
    Apr 23 18:29:36 slbd[1719]: Using r_refresh of 5000 milliseconds

    So for the moment I have made a full backup file and I modified the loadbalancer section to use "old style" configuration in order to have a working loadbalancing system. It would be cool if I could make the "new style" work…. If someone has the answer... I 've been reading the other posts  for an hour and haven't found the same problem.....

    Many thanks.



  • Go to the command prompt page.
    In the execute php section insert the following code.

    
    $status = interface_has_gateway("vlan0");
    if($status) echo "true"; else echo "false";
    

    Thanks.



  • Guess what ! result is "False"!

    Normal ?



  • Try a snapshot about 1-2 hours from now.  We just fixed this hopefully.



  • Do I have to use a 1.2 pre release snapshot or is there a 1.0.1 snapshot somewhere?

    Many thanks guys



  • Use a 1.2beta1 prerelease.



  • Great !

    Thanks hoba, scott and databeestje



  • So I understand it's fixed now?


Log in to reply