Hi,
I've detected that kernel version on machines were 2.4 instead of 2.6. So
during test It was used linuxthreads-0.10. After reading
http://www.unixville.com/~moazam/stories/2004/05/18/debuggingHangsInTheJvm.html,
I`ve decided to move to kernel 2.6 in order to install NPTL (Native Posix
Thread Library). So it would be very helpful to know which kernel version
and which library to manage threads was used when the bug was closed.
Thanks in advance.
2007/1/31, Alvaro Cabrerizo <[EMAIL PROTECTED]>:
Hi:
During a load test using Jmeter I get this error:
2007-01-31 09:27:47 StandardWrapperValve[jsp]: Servlet.service() para
servlet jsp lanzó excepción
java.lang.NullPointerException
at org.apache.nutch.searcher.DistributedSearch$Client.getSummary (
DistributedSearch.java:323)
at org.apache.nutch.searcher.NutchBean.getSummary(NutchBean.java
:349)
at org.apache.jsp.search_jsp._jspService(search_jsp.java:267)
at org.apache.jasper.runtime.HttpJspBase.service (HttpJspBase.java
:94)
at javax.servlet.http.HttpServlet.service(HttpServlet.java:802)
at org.apache.jasper.servlet.JspServletWrapper.service(
JspServletWrapper.java:324)
at org.apache.jasper.servlet.JspServlet.serviceJspFile (
JspServlet.java:292)
at org.apache.jasper.servlet.JspServlet.service(JspServlet.java
:236)
...
007-01-31 09:27:47 StandardWrapperValve[jsp]: Servlet.service() para
servlet jsp lanzó excepción
java.io.IOException : no responses
at org.apache.hadoop.ipc.Client.call(Client.java:344)
at org.apache.hadoop.ipc.RPC.call(RPC.java:174)
at org.apache.nutch.searcher.DistributedSearch$Client.getDetails(
DistributedSearch.java :307)
at org.apache.nutch.searcher.NutchBean.getDetails(NutchBean.java
:338)
at org.apache.jsp.search_jsp._jspService(search_jsp.java:266)
The test was done using three machines, in the same lan segment, with this
configuration:
-Pentium 4 3.0GHz
-1GB RAM
-Jdk 1.5
-Tomcat 5
-linux 2.6 kernel.
-nutch-0.8
-hadoop-0.4.0
The first machine runs a "search-server" that loads a Nutch index of 20000
documents, about 20MB of disk space for the Lucene index. The second one
runs a tomcat that communicates with the "search-server" and the third one
runs jmeter.
The request that launches the exception takes about 60 seconds, while rest
take about 650ms. Sysstat indicates that first and second machine have the
CPU idle, close to 100%, no swapping and still lot of free memory.
As I can see that is related with the closed bug nutch-14 (
http://issues.apache.org/jira/browse/NUTCH-14).
Any help will be apreciated in order to solve this problem.
