Tensorflow: how to close tensorboard server

Question

Once I've started tensorboard server with the command

tensorboard --logdir=path/to/logdir

is there a command that explicitly close it or can I just kill it without any harm?

Thanks

Answer 1

You can kill it without any harm! TensorBoard simply reads your log files and generates visualizations based on them in memory, so you don't need to worry about file corruption, etc.

Answer 2

This command will find the tensorbroad process and terminate it:

kill $(ps -e | grep 'tensorboard' | awk '{print $1}')

Answer 3

In my case, CTRL+C doesn't work. The following works for me:

1. CTRL+Z halts the on-going TensorBoard process.

2. Check the id of this halted process by typing in the terminal

   jobs -l

3. kill this process, otherwise you can't restart TensorBoard with the default port 6006 (of course, you can change the port with --port=xxxx)

   kill -9 #PROCESS_ID

Answer 4

I solved this problem by this way - (actually in my ssh, sometimes CTRL+C don't work properly. Then I use this)

1. Get the running tensorboard process details

   ps -ef|grep tensorboard

   Sample Output: uzzal_x+ 4585 4413 0 02:46 pts/4 00:00:01 bin/python /bin/tensorboard --logdir=runs/

2. Kill the process using pid (process id)

   kill -9 <pid>

   first number 4585 is my current pid for tensorflow