Hi, It's a historical problem for me, I asked administrator and others, no really-solved solution I got.
In some cluster, when I use 1 node, no problem in writing, when I turned to 4 nodes, sometimes writing, sometimes not writing, I mean, the job shows running, but for 4 hours, no writing, totally abnormal. which was going to kill me, especially for some job I am desperately waiting for results. might be someone experienced with this, can provide me some possible answers, Thanks ahead, -- Best Regards, lina
