On 27/10/16 00:51, Kaizaad Bilimorya wrote: > Hi Chris,
Hiya! > One way is to use "scontrol": > > scontrol --details show job <jobID> Ahh, I *knew* it should be somewhere, I hadn't spotted that option before. Thanks so much! [...] > So if you want actually core ids, I guess you will have to query the > cgroup hierarchy. No need, I'm just trying to automate the detection of bad jobs which are spanning nodes but not using the cores on the other nodes and I wanted a way to quantify how many cores were being wasted by the job. Thanks again! Chris -- Christopher Samuel Senior Systems Administrator VLSCI - Victorian Life Sciences Computation Initiative Email: [email protected] Phone: +61 (0)3 903 55545 http://www.vlsci.org.au/ http://twitter.com/vlsci
