I thought about this for a little while and here is what I think:
1) other sensors are still working (people are still sending data successfully) and your sensor seems like the only one that is busted. So, I don't think it is a sensor shell level problem or even a Soap problem.
2) this local versus live server problem is becoming a really big problem with the build process. This past couple days we had a couple of build failures that were the result of people (Me, Cedric, and Burt) not experiencing problems on their local machines but had problems either in hackydev and/or the public server. There were very different problems that made this happen. But, the bottom line is that we all either need to make sure that we do a freshStart junitAll before committing AND we need to make sure that our test cases are good enough to test the functionality of our code.
For your specific problem, I'm wondering why the TestJiraSensorShell unit test passed on hackydev and on the public server. It would seem that this unit test should test that the JiraSensor can send an Issue. If we can't send Issues to the public server using the actual sensor, then it would seem that this unit test should fail on the public server.
In any case... your problem can't be some mysterious problem. To solve these types of problems you have to analyze what is different. Are the sensorshells the same, do you have the latest versions from cvs, what version of java are you using, does the sensor only work when the sensor and the server are on the same machine, did you configured the usermap correctly, etc? In other words, if you provide us with a list of things that are the same and different and more specific information that you can gather the easier it will be for someone to help.
3) Something looks weird in your log file..
AutoSend not enabled. Server not available; offline data not recovered this session.
Why is AutoSend not enabled? I forget what AutoSend really indicates, however all the other sensors we have the following in the log.
#> AutoSend [10] AutoSend OK (set to 10 minutes) AutoSend enabled every 10 minutes. Checking for offline data to recover. No offline data found. #> Activity [setTool, Eclipse] setTool OK
Yours is different for some reason. Why?
4) The other weird thing is that when I run freshStart junitAll with (hackyKernel, hackyStatistics, hackyReport, hackyStdExt) I see that a Jira.log file is created. Why is that? I don't understand what is testing Jira code because everything that is Jira related should be in hackyJira (and I did *not* build that). None of the other sensors (like the Commit sensor) does anything like that.
Hackystat Version: null (null)
SensorShell started at: 11/14/2004 02:47:03
Type 'help' for a list of commands.
Host: http://hackystat.ics.hawaii.edu/ is available and key is valid.
Defined SDT: UnitTestDefined SDT: BuffTransDefined SDT: CoverageDefined SDT: BadData (no shell command)Defined SDT: BuildDefined SDT: ActivityDefined SDT: FileMetricDefined SDT: DependencyDefined shell command: Activity
Defined shell command: Dependency
Defined shell command: Coverage
Defined shell command: BuffTrans
Defined shell command: UnitTest
Defined shell command: Build
Defined shell command: FileMetric
AutoSend not enabled.
Checking for offline data to recover.
No offline data found.
#> Issue [setTool, Jira]
Command 'Issue' not found. Try again.
thanks, aaron
At 05:02 PM 11/15/2004 -1000, you wrote:
Hello Everyone,
The story goes that today I was up on hackydev trying to install the Jira Sensor. It constantly gave a SOAP invocation error upon the SensorShell send. At the end I decided that it had something to do with the SDT and went back home.
When I got back home I did a cvsUpdateAll and freshStart. I tried to reproduce the error using my local Hackystat host (LHS) without success! In other words sending of Jira sensor data and receiving of it was successful on my personal computer.
So as a debugging step I tried editing my sensor.properties file to point to the official hackystat server (OHS) instead of my LHS. However, I got a strange log entry. As can be seen from the below excerpt the initial ping of OHS indicates that the server is online and that the key is valid. However, upon subsequent SensorShell send attempts it says that "Server not available. Storing commands offline."
Does anyone have any input on this contradictory log entry?
Thanks! Burt
Hackystat Version: 6.3.1115 (November 15 2004 16:24:03) SensorShell started at: 11/15/2004 16:54:27 Type 'help' for a list of commands. Host: http://hackystat.ics.hawaii.edu/ is available and key is valid. Defined shell command: Issue Defined shell command: Perf Defined shell command: FileMetric Defined shell command: ReviewIssue Defined shell command: Activity Defined shell command: Cli Defined shell command: ReviewActivity Defined shell command: Coverage Defined shell command: BuffTrans Defined shell command: Commit Defined shell command: UnitTest Defined shell command: Build AutoSend not enabled. Server not available; offline data not recovered this session. #> Issue [setTool, Jira] setTool OK #> Issue [add, bug 1216, Bug, Major, ISSUE_WORKSTARTED, , reporter=bleung;assignee=bleung;project=10000;workflowId=10021;status=In Progress] Issue add OK (1) #> send Server not available. Storing commands offline. >>
