High CPU usage in [migration/X] process

[alberto0o]

Hi, I am getting this:



Kernel: 3.2.13-grsec-xxxx-grs-ipv6-32 #1 SMP

I am getting this after moving from 2.6 kernel, the CPU usage its low but I dont know if this threads really use CPU or they are just a bug.

Regards

[cPanelTristan]

Are you seeing this same activity in top in SSH? Also, what is this specific kernel as it appears to be a very custom one with grsec and ipv6 in it.

[alberto0o]

Using top I dont see this CPU usage, the Kernel is from OVH custom kernels.

I think this shouldnt be a problem from cPanel but any help would be great, I search and couldnt find any fix.

[cPanelTristan]

If it's only appearing in cPanel and not in top, then something is reading the output wrong possibly. Please submit a ticket using WHM > Support Center > Contact cPanel or using the link in my signature. Please post the ticket number here upon opening one. Thanks!

[NetMantis]

Using top I dont see this CPU usage, the Kernel is from OVH custom kernels.

I think this shouldnt be a problem from cPanel but any help would be great, I search and couldnt find any fix.

Actually, I have seen something similar to this in a lot of different servers each with the latest Cpanel release updates so maybe some sort of bug in and of itself.

Got clients who claim their servers are massively overloaded and when I take a look I see that WHM is showing high load numbers but when I look at SSH I see the loads are actually extremely low and not problems whatsoever and when I check server monitoring that watches the stats on the same servers 24 hours a day, I also see nothing but very low loads.

That tells me there is apparently some sort of bug in Cpanel/WHM in the newest releases which for whatever reason every now and then shows higher loads than what actually exist on the server.

Don't think this has to do with Linux Kernel because I've seen it on all sorts of different kernels and different CentOS versions.

[cPanelTristan]

Please submit a ticket as suggested. If you suspect it is a bug, then it would be reported to http://go.cpanel.net/bugs area or using the link at the top titled Bugs. Thanks!

[hostservicenet]

We've the same issue:
------------------------------
6 (Trace) (Kill) root - 1279872 0.0 [migration/0]
14 (Trace) (Kill) root - 94.9 0.0 [migration/3]
11 (Trace) (Kill) root - 91.2 0.0 [migration/2]
7 (Trace) (Kill) root - 82.2 0.0 [migration/1]
------------------------------

Note: same kernel ---> 3.2.13-grsec-xxxx-grs-ipv6-64

[hostservicenet]

I've submitted a support ticket
-----------------------------
Your Request id is: 2802458.
-----------------------------

Thanks,
Paolo

[Hostmavi]

We've the same issue
and same kernel ---> 3.2.13-grsec-xxxx-grs-ipv6-64
-------------------------------------------------------


17 (Trace) (Kill) root - 99.9 0.0 [migration/4]
6 (Trace) (Kill) root - 99.1 0.0 [migration/0]
11 (Trace) (Kill) root - 98.1 0.0 [migration/2]
23 (Trace) (Kill) root - 97.3 0.0 [migration/6]
7 (Trace) (Kill) root - 95.7 0.0 [migration/1]
26 (Trace) (Kill) root - 94.8 0.0 [migration/7]
20 (Trace) (Kill) root - 82.2 0.0 [migration/5]
14 (Trace) (Kill) root - 69.1 0.0 [migration/3]


Thanks
yueksel

[cPanelTristan]

Hello yueksel,

grsec kernels apparently show a high number of migration processes as part of the kernel. You should likely discuss this with your host and possibly using a different kernel if you do not wish to see these processes.

I'd suggest logging into SSH and running "top" (without the quotes) on the machine itself to check if it shows any actual high memory or CPU, since it likely doesn't.

Please note that checking for high usage in WHM is not as beneficial as logging into root SSH as the server administrator and checking in top. Anytime you are troubleshooting possible load or latency, logging into SSH to check top is the best way to do so.

Thanks!

[gtraxx]

hello I currently encounter the same bug with a VPS
PRocess manager :

7 (Trace) (Kill) root - 93.0 0.0 [migration/1]

[cPanelTristan]

This thread discussed what is causing it and provides suggestions. If they aren't actually showing in top in SSH, then there isn't an issue with the processes. If they are, then you'd need to speak with your provider. You can always go to a different kernel also. All was mentioned in the thread already.

[hostingcom]

I am using the same kernel on a OVH server. And also same problem here.
When viewing top I see no CPU load, but the time that it is working on the process is extremely long. It seems it doesnt close the task migration (threads).

any news to solve this problem as it is a long time ago?