WebLogic 集群平衡器始终路由到第二个节点

发布于 2024-12-28 11:35:51 字数 1801 浏览 4 评论 0原文

我已将 Apache 配置为负载平衡两个 WebLogic 节点（使用 mod_weblogic），但所有请求都转发到第二个节点。

这是我添加到 httpd.conf 的内容：

LoadModule weblogic_module modules/mod_wl_22.so
<IfModule mod_weblogic.c>
MatchExpression /soa-infra WebLogicCluster=node1.domain.com:8001,node2.domain.com:8001
</IfModule>
<Location /soa-infra>
SetHandler weblogic-handler
WebLogicCluster node1.domain.com:8001,node2.domain.com:8001
</Location>

我尝试过但没有效果：

1) swap node1.domain.com and node2.domain.com
2) change balancing algorithm in WebLogic Administration Console

一些说明： 1. 客户端正在寻址 apache - 我正在测试自己。 2.我看到apache仅在来自客户端/我的请求之后才创建连接，并且它仅连接到第二个节点（它正在其上执行）。 3. 是的，两个节点都在本地网络中并且可以访问。我发现了有趣的事情：如果我从httpd.conf中删除除了用于加载mod_weblogic的行之外的所有行并重新启动apache，那么我无法通过负载平衡器进行访问 - 所以平衡器实际上正在使用我的配置行，但如果我从列表中删除第二个节点，则负载均衡器仍会重定向第二个节点，而不是仅重定向一个列出的第一个节点。

看来我发现问题了！

经过几次请求后，我在日志中看到：

Wed Jan 25 19:56:49 2012 <480013275070093> Trying a pooled connection for '192.168.0.111/8001/8001'
Wed Jan 25 19:56:49 2012 <480013275070093> getPooledConn: No more connections in the pool for Host[192.168.0.111] Port[8001] SecurePort[8001]
Wed Jan 25 19:56:49 2012 <480013275070093> general list: trying connect to '192.168.0.111'/8001/8001 at line 2724 for '/soa-infra/services/partition/my_composite/my_composite_client_ep'

Wed Jan 25 19:57:05 2012 <480013275070256> Trying to locate Primary or Secondary using SrvrInfo with JVMID: -76285319
Wed Jan 25 19:57:05 2012 <480013275070256> getPreferredFromCookie: Found Primary 192.168.0.112:8001:0
Wed Jan 25 19:57:05 2012 <480013275070256> getPreferredFromCookie: Found 1 servers

192.168.0.111 是第一个节点，192.168.0.112 - 第二个节点。

有什么建议吗？

原文

I have configured Apache to load balance two WebLogic nodes (using mod_weblogic),
but all requests are forwarded to second node.

Here what I have added to httpd.conf:

LoadModule weblogic_module modules/mod_wl_22.so
<IfModule mod_weblogic.c>
MatchExpression /soa-infra WebLogicCluster=node1.domain.com:8001,node2.domain.com:8001
</IfModule>
<Location /soa-infra>
SetHandler weblogic-handler
WebLogicCluster node1.domain.com:8001,node2.domain.com:8001
</Location>

What I have tried with no effect:

1) swap node1.domain.com and node2.domain.com
2) change balancing algorithm in WebLogic Administration Console

Some clarifications:
1. client is adressing apache - I am testing myself.
2. I see that apache is creating connection only after request from client/me and it is only connection to second node (on which it is executing).
3. yes, both nodes are in local network and are reachable.
I have found interesting thing:
if I remove all my lines from httpd.conf except line for loading mod_weblogic and restart apache then I have no access via load-balancer - so balancer is really using my config lines,
but if I remove second node from list then load balancer are still redirecting second node instead of only one listed first node.

Looks like I have found issue!

After several requests I see in log:

Wed Jan 25 19:56:49 2012 <480013275070093> Trying a pooled connection for '192.168.0.111/8001/8001'
Wed Jan 25 19:56:49 2012 <480013275070093> getPooledConn: No more connections in the pool for Host[192.168.0.111] Port[8001] SecurePort[8001]
Wed Jan 25 19:56:49 2012 <480013275070093> general list: trying connect to '192.168.0.111'/8001/8001 at line 2724 for '/soa-infra/services/partition/my_composite/my_composite_client_ep'

Wed Jan 25 19:57:05 2012 <480013275070256> Trying to locate Primary or Secondary using SrvrInfo with JVMID: -76285319
Wed Jan 25 19:57:05 2012 <480013275070256> getPreferredFromCookie: Found Primary 192.168.0.112:8001:0
Wed Jan 25 19:57:05 2012 <480013275070256> getPreferredFromCookie: Found 1 servers

192.168.0.111 is a first node, 192.168.0.112 - second.

Any suggestions?

分享到QQ

分享到微博