Hey Gabe yeah just need to synchronize the method I think... Sent from my iPhone
On Sep 8, 2011, at 6:48 PM, "Gabe Resneck (JIRA)" <[email protected]> wrote: > > [ > https://issues.apache.org/jira/browse/OODT-314?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13100850#comment-13100850 > ] > > Gabe Resneck commented on OODT-314: > ----------------------------------- > > While the addJob() method is synchronized in the XStreamJobRepository class, > it is not in the MemoryJobRepository class. While this could lead to the > causing bug, I actually ran into it during some unit testing for some > JobQueue classes. I was submitting one job after another and kept seeing > non-deterministic results from the tests. After searching for the cause of > this I found that consecutively submitted jobs sometimes had the same ID > because they were added to the repo during the same millisecond. > > I will provide unit tests to check against this bug (and for the rest of the > class, since it currently lacks unit tests) when I submit the patch to fix it. > >> It is possible for 2 jobs submitted at very similar times to be assigned the >> same ID by the MemoryJobRepository >> --------------------------------------------------------------------------------------------------------------- >> >> Key: OODT-314 >> URL: https://issues.apache.org/jira/browse/OODT-314 >> Project: OODT >> Issue Type: Bug >> Components: resource manager >> Reporter: Gabe Resneck >> Assignee: Gabe Resneck >> >> Currently, when a job is added to a MemoryJobRepository, it helps to ensure >> no duplicates by assigning a new ID to that job based upon the current time. >> However, it is still possible for jobs to be assigned the same ID if they >> are submitted at very similar times. > > -- > This message is automatically generated by JIRA. > For more information on JIRA, see: http://www.atlassian.com/software/jira > >
