Hi Igor,

I played with PhaROS quite a bit while I was working at Mines Douai. But at the time, cleanning instances was quite an issue.. I use to tend to work with rosMaster instances that already exists or build a new image everytime it get too slow because of those ros master instances...��

You can try:

'''

PhaROSSystem cleanAll.

'''

But I think it keeps some instances anyway.

As far as I remember, I think they were working on rebuilding the networking interface to solve this issue.��

And Santi was also working on PhaROS2 looking for robustness...

I dont know the statut of those projects.

regards,

maxmattone

Le��lun. 26 d��c. 2016 ��00:09, Igor Stasenko <siguctua@gmail.com> a ��crit��:

Hmm okay, i created own ros package��
and started an example nodelet that prints 'hello world' , when i push messages from outside by ros..
now the question is how i should gracefully stop these nodelets and let them unsubscribe/die gracefully..
because i now have 4 instances of ROSMaster in my image and all locked by waiting for some mutex/semaphores and��
telling them 'stop' waits forever and blocks everything..
How i can clean this up?

--
Best regards,
Igor Stasenko.