I am trying to investigate why some of the requests from users are ending up on 
2 
different app-tier servers during the course of the session.

I turned on fine logging for com.caucho.servlets.HmuxLoadBalanceServlet on the 
load 
balancer and noticed these errors in the logs.

What could cause this and what settings in the conf would I need to fine tune?

[11:43:18.114] [web-tier->app-xxxxxx:154]java.net.SocketTimeoutException: Read 
timed out
[11:43:18.114] java.net.SocketTimeoutException: Read timed out
[11:43:18.114]  at java.net.SocketInputStream.socketRead0(Native Method)
[11:43:18.114]  at java.net.SocketInputStream.read(SocketInputStream.java:129)
[11:43:18.114]  at com.caucho.vfs.TcpStream.read(TcpStream.java:163)
[11:43:18.114]  at com.caucho.vfs.ReadStream.readBuffer(ReadStream.java:1015)
[11:43:18.114]  at com.caucho.vfs.ReadStream.read(ReadStream.java:313)
[11:43:18.114]  at 
com.caucho.servlets.HmuxLoadBalanceServlet.handleRequest(HmuxLoadBalanceServlet.java:647)
[11:43:18.114]  at 
com.caucho.servlets.HmuxLoadBalanceServlet.service(HmuxLoadBalanceServlet.java:224)
[11:43:18.114]  at 
com.caucho.server.dispatch.ServletFilterChain.doFilter(ServletFilterChain.java:103)
[11:43:18.114]  at 
com.caucho.server.webapp.WebAppFilterChain.doFilter(WebAppFilterChain.java:181)
[11:43:18.114]  at 
com.caucho.server.dispatch.ServletInvocation.service(ServletInvocation.java:266)
[11:43:18.114]  at 
com.caucho.server.http.HttpRequest.handleRequest(HttpRequest.java:269)
[11:43:18.114]  at 
com.caucho.server.port.TcpConnection.run(TcpConnection.java:603)
[11:43:18.114]  at com.caucho.util.ThreadPool$Item.runTasks(ThreadPool.java:721)
[11:43:18.114]  at com.caucho.util.ThreadPool$Item.run(ThreadPool.java:643)
[11:43:18.114]  at java.lang.Thread.run(Thread.java:595)
[11:43:46.225] load-balance for session efcRAZ1fbPAHC_c07ayns primary 
web-tier->app-bloodraven connection failed.
[11:47:10.778] load-balance for session eadLesk-QyalBehSgcyns primary 
web-tier->app-bloodraven connection failed.
[11:47:11.546] load-balance for session eadLesk-QyalBehSgcyns primary 
web-tier->app-bloodraven connection failed.
[11:47:11.546] load-balance for session eadLesk-QyalBehSgcyns primary 
web-tier->app-bloodraven connection failed.
[11:47:11.553] load-balance for session eadLesk-QyalBehSgcyns primary 
web-tier->app-bloodraven connection failed.
[11:47:11.604] load-balance for session eadLesk-QyalBehSgcyns primary 
web-tier->app-bloodraven connection failed.
[11:47:11.672] load-balance for session eadLesk-QyalBehSgcyns primary 
web-tier->app-bloodraven connection failed.
[11:47:11.690] load-balance for session eadLesk-QyalBehSgcyns primary 
web-tier->app-bloodraven connection failed.
[11:47:11.810] load-balance for session eadLesk-QyalBehSgcyns primary 
web-tier->app-bloodraven connection failed.
[11:47:12.368] load-balance for session eadLesk-QyalBehSgcyns primary 
web-tier->app-bloodraven connection failed.

Here are my load balance settings as of now.  I just added the 
load-balance-connect-timeout since I thought initially it was a connect 
timeout, but it 
seems to be a read timeout instead.  I will probably remove this to have it go 
back to 5s.

<load-balance-connect-timeout>10s</load-balance-connect-timeout>
<load-balance-recover-time>600s</load-balance-recover-time>
<load-balance-warmup-time>10s</load-balance-warmup-time>
<socket-timeout>120s</socket-timeout>
<keepalive-timeout>20s</keepalive-timeout>

Let me know if you need additional info.

Thanks,
Anthony


_______________________________________________
resin-interest mailing list
resin-interest@caucho.com
http://maillist.caucho.com/mailman/listinfo/resin-interest

Reply via email to