Sending SIGUSR1 to node v0.12.7 process stops it

[diversario]

I was trying to attach node-inspector to a running process and saw something odd – having a group of active processes, each using 10-20% CPU, sending kill -s SIGUSR1 <pid> makes that particular process go to 0% CPU, memory usage would not change either.

I was able to debug v0.10 processes successfully using the same technique.

This is happening on a CentOS 6 (kernel 2.6.32-504.el6.x86_64), node v0.12.7.

[diversario]

Ok, follow up.

I was trying to debug remotely so I had to forward one port by setting up an SSH tunnel. As it turned out, node-inspector (which I was using) needs two ports open on the target machine, one for HTTP and another one for websockets. Setting up a second tunnel for the WS port made node-inspector work and, once it connected to the stalled process, the process resumed.

[diversario]

Probably a non-issue.