Kernel Thread that won't exit, Load Average +1

0

So I have a new server that is up and running with Centos 6, and the load average is stuck hovering around 1, even when the machine is nearly idle. I reviewed the process table and found this running:

root@myserver:~]# ps -efl|grep " [D] "
1 D root     43252     2  0  80   0 -     0 rpc_wa Apr01 ?        00:00:00 [10.125.50.56-ma]

The timing (Apr01) lines up perfectly with my trending graphs, and it was right around the time I was working with mounting an NFS share from that IP address. However, the thread is in uninterruptible sleep, and I don't know how to get rid of it. I have unmounted any/all NFS mounts I had, and it still won't go away.

Does anyone know what this process/thread is, and if there is a way to help it exit without restarting the box? It is already in production so I'd really like to avoid that.

Aaron R.

Posted 2015-04-09T19:11:39.073

Reputation: 101

The D flag (2nd column) says that it's in uninterruptable sleep and I think you can't kill that with -9. You can try -SEGV. And you can try to find out the exact WCHAN, rpc_wa is just the beginning of it. – ott-- – 2015-04-09T20:01:06.033

Answers

-1

If it's ok just to kill the thread, you can try to use command

kill -9 43252

It will send SIGKILL signal to the process, which should immediately terminate it.

jjurm

Posted 2015-04-09T19:11:39.073

Reputation: 101

I think you missed that the process is in an uninterruptible sleep, so interrupting it isn't an option. – David Schwartz – 2015-04-09T20:25:53.120

Asked: 2015-04-09T19:11:39.073

Viewed: 199 times

Active: 2019-08-20T01:21:41.540