log running nodes

asked 2021-02-09 04:11:47 -0500

felixN
108 ●15 ●20 ●25

Hi,

is there an easy way to know deduce from logs which nodes are running and when? Or a way to monitor it and to record it mannually (for example publishing the start and stop of nodes into some topic)?

In the end, what I would like to obtain, is something like

timestamp1   nodeA started
timestamp2   nodeB started
timestamp3   nodeA stoped
timestamp4   nodeC started
timestamp5   nodeB stoped
timestamp6   nodeC stoped

if instead of "stoped" I can know if it is a clean exit or a crash, it's even better.

If you have a solution to get this info from existing logs, it would be nice (right now, I have logs from a "crash" I couldn't find the cause yet), but if there is something to do beforehand, I'm also find with it.

A very ugly solution would be a bash script running "rosnode list" at regular intervals, then doing the diff between 2 successive runs and saving it to a file along with timestamp. But I would be surprised if there isn't any better solution

Thanks a lot in advance

Felix

edit retag flag offensive close merge delete

add a comment

answered 2021-02-09 08:40:50 -0500

tryan
1421 ●96 ●26

updated 2021-02-09 11:50:41 -0500

How accurate do you need the data to be? You can find out when each node (un)registers for subscriptions and publications in master.log, which is usually in ~/.ros/log/latest. Here's an example of some output:

[rosmaster.master][INFO] 2021-02-09 09:18:53,983: +PUB [/rosout] /node1 http://computer:34533/
[rosmaster.master][INFO] 2021-02-09 09:18:53,984: +PUB [/topic1] /node1 http://computer:34533/
[rosmaster.master][INFO] 2021-02-09 09:18:53,984: +PUB [/topic2] /node1 http://computer:34533/
[rosmaster.master][INFO] 2021-02-09 09:18:53,988: +PARAM [/node1/param1] by /node1
[rosmaster.master][INFO] 2021-02-09 09:18:53,988: +PARAM [/node1/param2] by /node1
[rosmaster.master][INFO] 2021-02-09 09:19:44,372: -SUB [/clock] /node1 http://computer:34533/
[rosmaster.master][INFO] 2021-02-09 09:19:44,372: -SUB [/tf] /node1 http://computer:34533/
[rosmaster.master][INFO] 2021-02-09 09:19:44,372: -SUB [/tf_static] /node1 http://computer:34533/
...
[rosmaster.master][INFO] 2021-02-09 09:19:44,369: -PUB [/rosout] /node1 http://computer:34533/
[rosmaster.master][INFO] 2021-02-09 09:19:44,369: -PUB [/topic1] /node1 http://computer:34533/
[rosmaster.master][INFO] 2021-02-09 09:19:44,369: -PUB [/topic2] /node1 http://computer:34533/
[rosmaster.master][INFO] 2021-02-09 09:19:44,372: -SUB [/clock] /node1 http://computer:34533/
[rosmaster.master][INFO] 2021-02-09 09:19:44,372: -SUB [/tf] /node1 http://computer:34533/
[rosmaster.master][INFO] 2021-02-09 09:19:44,372: -SUB [/tf_static] /node1 http://computer:34533/

Also, you can find info, warnings, and errors in rosout.log. There will be other files in the same folder, depending on your logging configuration. I'm not sure if the standard logs will contain all the information you need, but you can customize the logging: C++ wiki, Python wiki.

If you're using launch files, each of their logs may indicate more precisely what you want. An example:

...
[roslaunch][INFO] 2021-02-09 09:18:48,568: process[node1-6]: started with pid [16366]
[roslaunch][INFO] 2021-02-09 09:18:48,568: ... successfully launched [node1-6]
...
[roslaunch][INFO] 2021-02-09 09:19:44,010: [node1-6] killing on exit
[roslaunch][INFO] 2021-02-09 09:19:44,011: process[node1-6]: killing os process with pid[16366] pgid[16366]
[roslaunch][INFO] 2021-02-09 09:19:44,011: [node1-6] sending SIGINT to pgid [16366]
[roslaunch][INFO] 2021-02-09 09:19:44,012: [node1-6] sent SIGINT to pgid [16366]
[roslaunch][INFO] 2021-02-09 09:19:44,413: process[node1-6]: SIGINT killed with return value 0
...

edit flag offensive delete link

Comments

Thanks a lot! It's not exactly what I was looking for, but it does the job. I had a look into master.log before, but there was too much "trash". but doing a grep on PUB and/or SUB enables not get something reasonnably readable.

felixN ( 2021-02-09 08:56:10 -0500 )edit

I updated my answer with something that may be a little more on target. I agree; it can be hard to sift through the logs if you don't know exactly what to look for. grep is a good idea.

tryan ( 2021-02-09 11:59:03 -0500 )edit

add a comment

log running nodes

1 Answer

Comments

Question Tools

Stats

Related questions

log running nodes edit

1 Answer

Comments

Question Tools

Stats

Related questions

log running nodes