Accumulating Processes

Get Help
1

One of my colleagues was interested in accumulating processes on the ondemand server. We run a load monitor on our machines and have noticed some spikes. These spikes aren’t an issue, just for our own knowledge. But my colleague was digging deeper and noticed that lsof of .sock files are causing “D” states and those processes are accumulating. There are ~14 of these for a single user.
Example:

# ps auxw | grep .sock
root      13621  0.0  0.0 119368  1132 ?        S    Mar09   0:00 lsof -F piu /var/run/ondemand-nginx/<user>/passenger.sock
root      13623  0.0  0.0 119368   252 ?        D    Mar09   0:00 lsof -F piu /var/run/ondemand-nginx/<user>/passenger.sock

Does anyone have any insight on these potentially causing usage spikes? This is something we were reading as well - http://erman-arslan-s-oracle-forum.2340467.n4.nabble.com/What-is-the-performance-impact-of-D-state-processes-td7494.html

2

Hey we haven’t had time to look into this but we will. We’re going to open a github issue and label it a bug because it appears to be. I’ll let you know the ticket number when I’ve created it.

3

Thanks Jeff. It wasn’t a big issue at all, mostly just wanted to ping the community about it as it seemed odd.

4

Yes it is odd. Thanks for the heads up! I’ve created this ticket to investigate if you’d like to follow it or provide more information to it.

5

This topic was automatically closed 180 days after the last reply. New replies are no longer allowed.