Hi @xuanz01
Hopefully your job completed by now!
The timing for jobs can vary for both the queue phase and the execution phase. It can vary by day, tool, data, server. Underneath every Galaxy server is a set of clusters. How busy they are depends on how many other people are working that day!
Once a job is started it is best to allow it to fully process. Rerunning a new job starts the entire process over again.
Whenever possible, try to use a workflow. This will start up all your tools at once. This allows them to process then start up the next tool while you are away and can significantly speed up processing times. Processing many samples at once inside of collections does something similar.
Hope this helps! 