[gridsam-discuss] Repeated GridSAM job submissions fail: org.xml.sax.SAXException: Bad envelope tag: html

[Garry S. <gar...@po...>]

Hi William,

Are you seeing this behaviour in your latest version of GridSAM?

GridSAM 1.1.0 rc1, OMII 2.3.0. Java 1.5, default Java Heap size

Server: uname -a
Linux holly 2.4.32 #1 SMP Fri Nov 25 14:41:06 GMT 2005 i686 GNU/Linux
Dual processor, 2GB physical memory.


I am running into a problem with GridSAM and mutliple concurrent job 
submission.
The JSDL for the following tests is attached (Notice there a not any 
datastaging elements; when included the situation is worse: I get 
OutOfMemory Exceptions).

The command (jobCommand):

kryten:~/jsdl/1.1.0-rc1> 
~/bin/OMIICLIENT3_2_0/OMIICLIENT/gridsam/bin/gridsam-submit -s 
http://holly.dsg.port.ac.uk:8088/gridsam/services/gridsam?WSDL 
helloworld.jsdl helloworld.jsdl helloworld.jsdl helloworld.jsdl 
helloworld.jsdl helloworld.jsdl helloworld.jsdl helloworld.jsdl 
helloworld.jsdl helloworld.jsdl helloworld.jsdl helloworld.jsdl 
helloworld.jsdl helloworld.jsdl helloworld.jsdl helloworld.jsdl 
helloworld.jsdl helloworld.jsdl helloworld.jsdl helloworld.jsdl 
helloworld.jsdl helloworld.jsdl helloworld.jsdl helloworld.jsdl 
helloworld.jsdl helloworld.jsdl helloworld.jsdl helloworld.jsdl 
helloworld.jsdl helloworld.jsdl


GridSAM configured to use fork:

I can submit the jobCommand successfully a maximum of 5 times to the 
same GridSAM instance before the GridSAM refuses to accept any further 
connections.  Before submitting each jobCommand I wait for previous job 
instances to complete by monitoring the gridsam.log. The speed at which 
gridsam processes subsequent iterations gets slower with each iteration.
 
On the sixth attempt I get the following error:

2006-05-02 09:38:59,923 FATAL [GridSAMSubmit] (main:) unable to submit 
job: failed to submit job: WSDoAllReceiver: cannot convert into 
document; nested exception is:
        org.xml.sax.SAXException: Bad envelope tag:  html

There are no exceptions in Catalina.out and gridsam.log shows the end of 
the previous successful submission, e.g:

2006-05-02 10:44:09,805 INFO  [144516830af4709f010af4785bd7025a] 
deleting working directory: /tmp/gridsam-144516830af4709f010af4785bd7025a
2006-05-02 10:44:09,821 INFO  [144516830af4709f010af47866ed025f] 
deleting working directory: /tmp/gridsam-144516830af4709f010af47866ed025f


In order to get the GridSAM to accept jobs again I stop/start the tomcat.


Observing memory usage using top

 PID   USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND

 17600 omii       9   0  110m 109m  12m S  0.0  6.0   0:20.86 java  
#tomcat started

 17600 omii       9   0  120m 120m  13m S  0.0  6.0   0:20.86 java  
#After first iteration

 17600 omii       9   0  122m 122m  14m S  0.0  6.0   0:20.86 java  
#After the second iteration

 17600 omii       9   0  120m 120m  13m S  0.0  6.0   0:20.86 java  
#After the third iteration

 17600 omii       9   0  120m 120m  13m S  0.0  6.0   0:20.86 java  
#After the fourth iteration

 17600 omii       9   0  121m 121m  13m S  0.0  6.0   0:20.86 java  
#After the fifth iteration

# Sixth iteration fails consistently with the usual failure message:

2006-05-04 12:20:27,255 FATAL [GridSAMSubmit] (main:) unable to submit 
job: failed to submit job: WSDoAllReceiver: cannot convert into 
document; nested exception is:
        org.xml.sax.SAXException: Bad envelope tag:  html

The Gridsam server memory usage after reading this error message at the 
client is:

17600 omii       9   0  122m 122m  14m S  0.0  6.1   0:20.86 java


Thanks in advance

regards
Garry