"eos cp" command hangs« Back to Questions List
|
Hi! I have started a command
eos cp -r -p -P /nica/mpd21/plotnikov/bmnroot/ /eos/nica/bmn/users/vplotnik/bmnroot/
at friday using qsub task. Part of directories have been copied. Then the command execution stailed. While the qstat status of the job is "r". Quota is not exceeded. I have checked the task on the next day. No new megabytes have been copied. The status of the command is the same ("r"). May be something wrong with the cluster.
|
| ▲ ▼ |
I have tried using eoscp with nohup. And I am facing problem: I cannot save eoscp logs to file. Can I do it somehow? |
| ▲ ▼ |
The copy task should only be run locally, without a batch Indeed, it is not. For instance, if you process data with writing result files to a temporary storage (e.g. /weekly) , then you need to copy them in batch mode to the final storage. I always use ”xrdcp, eoscp” command in batch mode on the supercomputer Govorun after writing to the temporary SSD Lustre storage. |
| ▲ ▼ |
Hi! Thank you for suggestions to use nohup or cron. I never used it before. I will try it. But sometimes I need to make backup from the batch task. Can I do that? |
| ▲ ▼ |
Why run the task of copying data in the batch system? Do you want to copy the same data about 100 times with different machines? The copy task should only be run locally, without a batch. For example, you can use nohup to continue the task in the background or cron with a rsync for synchronization files. |