risk of disrupting batch job data

1 次查看（过去 30 天）

Rafael 2013-11-13

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/105968-risk-of-disrupting-batch-job-data

评论： Rafael 2013-11-14

Hi all.

I am using Matlab 2013a with Parallel Computing Toolbox on multiple machines and connecting each of those machines to the same remote HPC cluster with Matlab Distributed Computing Server. Now I observe that the batch job data is usually stored using the jobname Job1, Job2, Job3, etc. And that number seems to depend on the job history of each individual machine.

Do I have to workaround the risk of disrupting batch job data from another machine having accidentally the same name manually ?

How should I do this in the easiest and best way ? (use for example different job data storage folders on the HPC cluster for every single connecting machine... or what else?)

I hope I could make it clear to you what I mean!

Thanks for your comments on that, Rafael