balancer does not balance ... !

[Tex]

Could someone help me find what I did wrong with the following configuration :

<?xml version="1.0" encoding="UTF-8"?>
<virtualHostConfig>
<docRoot>/var/www</docRoot>
<enableGzip>0</enableGzip>
<adminEmails></adminEmails>
<extProcessorList>
<extProcessor>
<type>proxy</type>
<name>web01</name>
<address>web01.xxx.com:80</address>
<maxConns>1000</maxConns>
<pcKeepAliveTimeout>3600</pcKeepAliveTimeout>
<initTimeout>2</initTimeout>
<retryTimeout>2</retryTimeout>
<respBuffer>0</respBuffer>
</extProcessor>
<extProcessor>
<type>proxy</type>
<name>web02</name>
<address>web02.xxx.com:30001</address>
<maxConns>1000</maxConns>
<pcKeepAliveTimeout>3600</pcKeepAliveTimeout>
<initTimeout>2</initTimeout>
<retryTimeout>2</retryTimeout>
<respBuffer>0</respBuffer>
</extProcessor>
<extProcessor>
<type>loadbalancer</type>
<name>static</name>
<workers>proxy::web01, proxy::web02</workers>
</extProcessor>
</extProcessorList>
<contextList>
<context>
<type>loadbalancer</type>
<uri>/</uri>
<handler>static</handler>
<allowOverride>0</allowOverride>
<realm></realm>
<authName></authName>
<required></required>
<accessControl>
<allow></allow>
<deny></deny>
</accessControl>
<authorizer></authorizer>
<addDefaultCharset>off</addDefaultCharset>
<defaultCharsetCustomized></defaultCharsetCustomized>
</context>
</contextList>
</virtualHostConfig>


The balancer redirects every request to the first proxy host. What did I miss ?

Thanks.

[mistwang]

configuration looks fine to me, maybe you can try increase the concurrent level, LSWS uses the first idle cluster node.

[Tex]

OK, I'll check this way. Shouldn't it use round robin ? Would balance better the load across nodes, isn't it ?

[nathanc]

I have the same issue. It seems that only one of the nodes will get the traffic.
Here is what I have in my Node Address List:

(node1) 10.39.121.1->10.39.121.2:80, (node1) 10.39.121.1->10.39.121.7:80
Strategy is Round Robin

It seems the last node address gets the load because, for a test I swapped out the .2 and .7 addresses to confirm this behavior.

Perhaps I have something messed up with in the address list.

[mistwang]

should
(node1) 10.39.121.1->10.39.121.2:80, (node1) 10.39.121.1->10.39.121.7:80
be
(node1) 10.39.121.1->10.39.121.2:80, (node2) 10.39.121.1->10.39.121.7:80
?

[nathanc]

Awesome thanks!,
I see now.
You guys ROX!

[mistwang]

There is a LSLB 1.2 release, you can get it by changing the version number in the download link.