Hi,
In the issue-tracker I couldn't find a way to create a new issue, which I assume is due to a lack of permissions for my user.
I also searched through the issues and didn't find the following report yet.
When a CPAS pipeline fails for whatever reason (e.g. Mascot failed due to a server restart) the state of the job changes to ERROR as expected. Unfortunately, when trying to start the job again with the same files and search-protocol it won't start the job as labkey indicates the job is RUNNING.
It seems the only way to get it to work again, is to force the job to COMPLETE, delete it and remove the corresponding directory structure from the server.
If another method exists to get a job to run again, I would love to know about it.
This bug is present in Labkey 8.1, which runs on:
* Debian Etch
* Postgres 8.1
* Tomcat 5.5.20-2etch2
-Aschwin |
| |
| wnels2 responded: |
2008-04-21 05:23 |
Hi Aschwin,
Each run needs to have a unique experiment protocol so you will need to rename the protocol to do a second run. That is choose <new protocol> and copy and paste the input xml. The running status means that some of the artifact file from the run weren't cleaned up because it failed in the middle. CPAS uses those files to determine the state of the run.
There will be a copy protocol button soon.
Bill |
| |
| brendanx responded: |
2008-04-21 07:20 |
Hi Aschwin,
Sounds like you have found what is currently required, though it is only deleting the created directory. Forcing the job status to complete is not necessary.
This is only the case with the single-machine, trial installation. The current Perl Cluster Pipeline supplies a "Retry" button that allows you to pick up where you left of, in case there was a system error.
We are actively working on the pipeline, and will fix this issue with the trial installation in a future release.
--Brendan |
| |
| aschwin.vanderwoude responded: |
2008-05-20 05:07 |
Bill and Brendan,
Thanks for your responses.
Good to know you are working on this issue. I am not sure what is meant with a 'trial installation' as I do mean this installation to be used in production.
I also wondered if there is some documentation on the mentioined perl-pipeline.
Thanks,
Aschwin |
| |
| eknelson responded: |
2008-05-20 11:01 |
|
| |
|
|