running job on remote agent

Well here is the latest debug log on the Agent side and we are making some progress. I am still getting the socket closed error and everything i have looked up online points to that being the program having the socket open doesnt know what to do with the request and closes the socket. The errors on both master and agent have changed and as you see now the agent reports the master connection but only when i test the agent not when i submitt a job.

2007-11-07 12:36:59,425 [main] DEBUG com.softtreetech.jscheduler.JSchedulerStarter - main(...) : start
2007-11-07 12:36:59,428 [main] DEBUG com.softtreetech.jscheduler.JSchedulerStarter - startup() : start
2007-11-07 12:36:59,440 [main] DEBUG com.softtreetech.jscheduler.JSchedulerStarter - startup() : creating business objects
2007-11-07 12:36:59,692 [main] DEBUG com.softtreetech.jscheduler.JSchedulerStarter - startup() : creating UI controller
2007-11-07 12:36:59,979 [main] DEBUG com.softtreetech.jscheduler.JSchedulerStarter - startup() : initializing business objects
2007-11-07 12:36:59,994 [main] DEBUG com.softtreetech.jscheduler.business.preferences.AbstractPrefDatabase - Creating backup for preferences file preferences.xml
2007-11-07 12:37:00,001 [main] DEBUG com.softtreetech.jscheduler.business.preferences.AbstractPrefDatabase - Preferences file has been copied to preferences.bak
2007-11-07 12:37:01,217 [main] DEBUG com.softtreetech.jscheduler.JSchedulerStarter - startup() : loading job db
2007-11-07 12:37:01,217 [main] DEBUG com.softtreetech.jscheduler.JSchedulerStarter - startup() : ready to run jobs
2007-11-07 12:38:23,089 [RMI TCP Connection(5)-192.168.208.155] DEBUG com.softtreetech.jscheduler.business.runner.security.SecurityService - authNativeUser: jscheduler login ok
2007-11-07 12:40:07,764 [RMI TCP Connection(8)-192.168.208.155] DEBUG com.softtreetech.jscheduler.business.runner.security.SecurityService - authNativeUser: jscheduler login ok
2007-11-07 12:48:39,665 [RMI TCP Connection(22)-192.168.208.155] DEBUG com.softtreetech.jscheduler.business.runner.security.SecurityService - authNativeUser: jscheduler login ok

Well this looks a lot better. I think now we are getting very close. Please post debug.log from the master.

I just can seem to get it t work on a regular basis and have read java forums until i am blue in the face. I did get it to just run the ls command remote but one time it runs the next 15 it doesnt and i cannot figure out why. One silly question we have as when we run agent.sh nogui from a shell connection it starts and says press control c to abort but we cannot get the command line to do anything once the agent starts we have to leave the ssh connection open or when we close the connection it kills the agent. No commands except control c seem to work .

8:31 am
remote scheduler log

8-Nov-2007 08:31:44 AM 2 null 8 Open edit Remote job started.
8-Nov-2007 08:31:44 AM 2 null 8 Open edit Job started.
8-Nov-2007 08:31:45 AM 2 null 8 Open edit Job finished.
8-Nov-2007 08:31:46 AM 2 null 8 Open edit Remote job finished.

remote debug log

2007-11-08 08:31:44,766 [Job #8 - Open edit] DEBUG com.softtreetech.jscheduler.business.runner.ProgramJobRunner - runJob(): start
2007-11-08 08:31:45,270 [Job #8 - Open edit] DEBUG com.softtreetech.jscheduler.business.runner.security.SecurityService - authNativeUser: jscheduler login ok
2007-11-08 08:31:45,274 [Job #8 - Open edit] DEBUG com.softtreetech.jscheduler.business.runner.ProgramJobRunner - execProcess(): command line [ls] in work directory [null]
2007-11-08 08:31:45,275 [Job #8 - Open edit] DEBUG com.softtreetech.jscheduler.business.runner.ProgramJobRunner - runAs() username=jscheduler command=ls workDir=
2007-11-08 08:31:45,275 [Job #8 - Open edit] DEBUG com.softtreetech.jscheduler.business.runner.ProgramJobRunner - exec : ./runas.pl,jscheduler,ls,
2007-11-08 08:31:45,471 [Job #8 - Open edit] DEBUG com.softtreetech.jscheduler.business.runner.ProgramJobRunner - waitForProcess(): start
2007-11-08 08:31:45,472 [Thread-10] DEBUG com.softtreetech.jscheduler.business.runner.AbstractJobRunner$TimeoutVerifier - run(): start
2007-11-08 08:31:45,473 [Thread-10] DEBUG com.softtreetech.jscheduler.business.runner.AbstractJobRunner$TimeoutVerifier - run(): sleeping. Max wait time is 420000 milliseconds
2007-11-08 08:31:45,781 [Job #8 - Open edit] DEBUG com.softtreetech.jscheduler.business.runner.ProgramJobRunner - waitForProcess(): end
2007-11-08 08:31:45,782 [Job #8 - Open edit] DEBUG com.softtreetech.jscheduler.business.runner.AbstractJobRunner - isFailed(...) : exit code 0
2007-11-08 08:31:45,782 [Job #8 - Open edit] DEBUG com.softtreetech.jscheduler.business.runner.ProgramJobRunner - killProcess start
2007-11-08 08:31:45,782 [Job #8 - Open edit] DEBUG com.softtreetech.jscheduler.business.runner.ProgramJobRunner - runJob(): end
2007-11-08 08:31:46,188 [Thread-10] INFO com.softtreetech.jscheduler.business.runner.AbstractJobRunner$TimeoutVerifier - run():
java.lang.InterruptedException: sleep interrupted
at java.lang.Thread.sleep(Native Method)
at com.softtreetech.jscheduler.business.runner.AbstractJobRunner$TimeoutVerifier.run(Unknown Source)
2007-11-08 08:31:46,193 [Thread-10] DEBUG com.softtreetech.jscheduler.business.runner.AbstractJobRunner$TimeoutVerifier - run(): ended
2007-11-08 08:31:59,608 [Job #8 - Open edit] DEBUG com.softtreetech.jscheduler.business.runner.security.SecurityService - authN

A few hints, that might help you

1. Don't schedule/run ls, run sh -c ls. Besides you cannot see if ls succeeded or not if you don't send the output to a file, so you probably want to do sh -c ls > /home/mydir/ls_out.log

2. agent.sh is not an interactive console application. You cannot type any commands and you should not start it via SSH in the foreground unless it is not already running and you need to test something.
3. To run it in the background, use Linux standard method, which is adding & symbol at the end of the command line

Hope this makes sense and get you where

Correction, in all places referring to sh, please replace sh with /bin/sh
You can also run /bin/ls, but in this form you cannot send the output to a file, because output redirection is a function of shell

Last edited by SysOp on Thu Nov 08, 2007 12:11 pm; edited 1 time in total

I have taken that in to account and have tried running every command line possible including the one you just put in the last update. the problem is something java related. It closes the socket session right after the agent is contacted and never records the job even tried running or that it was contacted on the any of the agent logs. the master log just says.

2007-11-08 11:03:23,730 [Job #12 - New Job] ERROR com.softtreetech.jscheduler.business.runner.RemoteJobRunner - runJob
java.rmi.UnmarshalException: Error unmarshaling return header; nested exception is:
java.net.SocketException: Connection reset
at sun.rmi.transport.StreamRemoteCall.executeCall(StreamRemoteCall.java:203)
at sun.rmi.server.UnicastRef.invoke(UnicastRef.java:133)
at com.softtreetech.jscheduler.business.agent.remote.RemoteAgentImpl_Stub.executeJob(Unknown Source)

which is weird because every connection test i run on the agent shows in the agents debug log.

Are agent and scheduler from the same version and build?
Does the scheduler say in the job log that this job failed?
Are you isng JDK/JRE 1.4 for the scheduler?

og the agent.sh it reads
#!/bin/sh
JAVA_HOME=/usr/opt/j2sdk1.4.2_15

blh blah

on the master

#!/bin/sh
JAVA_HOME=/opt/j2sdk1.4.2_15

and i verified that is the correct paths for the java installs

I have the network guys haelping me and now we cannot get the agent to start getting

2007-11-08 12:07:09,439 [main] ERROR com.softtreetech.jscheduler.business.log.LoggerImpl - parseLogRecord(...) :
java.text.ParseException: Unparseable date: ""
at java.text.DateFormat.parse(DateFormat.java:335)
at com.softtreetech.jscheduler.business.log.LoggerImpl.o00000(Unknown Source)
at com.softtreetech.jscheduler.business.log.LoggerImpl.init(Unknown Source)
at com.softtreetech.jscheduler.business.O.o00000(Unknown Source)
at com.softtreetech.jscheduler.business.void.this.super(Unknown Source)
at com.softtreetech.jscheduler.JSchedulerStarter.o00000(Unknown Source)
at com.softtreetech.jscheduler.JSchedulerStarter.main(Unknown Source)
2007-11-08 12:07:09,518 [main] FATAL com.softtreetech.jscheduler.JSchedulerStarter - startup() :
java.lang.NullPointerException
at java.util.Date.before(Date.java:858)
at com.softtreetech.jscheduler.business.log.LoggerImpl$o.compareTo(Unknown Source)
at java.util.Collections$ReverseComparator.compare(Collections.java:2452)
at java.util.Arrays.mergeSort(Arrays.java:1278)
at java.util.Arrays.mergeSort(Arrays.java:1290)
at java.util.Arrays.mergeSort(Arrays.java:1290)
at java.util.Arrays.sort(Arrays.java:1219)
at java.util.Collections.sort(Collections.java:155)
at com.softtreetech.jscheduler.business.log.LoggerImpl.init(Unknown Source)
at com.softtreetech.jscheduler.business.O.o00000(Unknown Source)
at com.softtreetech.jscheduler.business.void.this.super(Unknown Source)
at com.softtreetech.jscheduler.JSchedulerStarter.o00000(Unknown Source)
at com.softtreetech.jscheduler.JSchedulerStarter.main(Unknown Source)

Sorry for confusion about sh, I meant references to /bin/sh in my previous post as the command line to run, not references to the contents of .SH files that come with 24x7. These files are fine.

In regard, to "Unparseable date..." please rename schedule.log on the computer where this message appears in the trace or just ignore that trace event, it is not really important.

By the way, I think at this point you should really start fresh: shutdown both agent and scheduler; delete all log files on both sides; start them again; wait for the scheduled remote job to run; check the log files and the output as I suggested earlier. This way you can get a clear picture of what is going on and see whether its working or not.

got it for some darn reason it just couldnt read it own logs, but the agent is running now

I have it working!!!! we are going to run a few test and hopefully be moving forward soon with the project. It all seemed to be network realted issues on our end. Program runs great, I appreciate all the help and it has been a huge factor in deciding what software we decide to purchase.

Thanks for the status update. I'm glad you were able to figure out and fix the route cause.