PDA

View Full Version : high load for one kernel migration on SLES 11 SP2



pkoch4bgc
03-Mar-2015, 13:48
Hi,

we are using SLES 11 SP2 on our file services server. The system is running quantum stornext(therefore we can not update it in the moment).
However since we move to new hardware we see the following behavior:

Only one migration process seems to work. all other are doing nothing.

~ # ps aux | grep mig
root 6 0.0 0.0 0 0 ? S 2014 0:00 [migration/0]
root 8 0.0 0.0 0 0 ? S 2014 0:00 [migration/1]
root 13 0.0 0.0 0 0 ? S 2014 0:00 [migration/2]
root 17 0.0 0.0 0 0 ? S 2014 0:00 [migration/3]
root 21 0.0 0.0 0 0 ? S 2014 0:00 [migration/4]
root 25 69.4 0.0 0 0 ? S 2014 159181:27 [migration/5] <######
root 29 0.0 0.0 0 0 ? S 2014 0:00 [migration/6]
root 33 0.0 0.0 0 0 ? S 2014 0:00 [migration/7]



We are using a default installation. Only nfs is running on the machine.
We did not configure anything kernel related.

Any idea ?
How can we split the work to all cpus ?

Bye, Peer

jmozdzen
04-Mar-2015, 17:22
Hi Peer,

Hi,

we are using SLES 11 SP2 on our file services server. The system is running quantum stornext(therefore we can not update it in the moment).
However since we move to new hardware we see the following behavior:

Only one migration process seems to work. all other are doing nothing.

~ # ps aux | grep mig
root 6 0.0 0.0 0 0 ? S 2014 0:00 [migration/0]
root 8 0.0 0.0 0 0 ? S 2014 0:00 [migration/1]
root 13 0.0 0.0 0 0 ? S 2014 0:00 [migration/2]
root 17 0.0 0.0 0 0 ? S 2014 0:00 [migration/3]
root 21 0.0 0.0 0 0 ? S 2014 0:00 [migration/4]
root 25 69.4 0.0 0 0 ? S 2014 159181:27 [migration/5] <######
root 29 0.0 0.0 0 0 ? S 2014 0:00 [migration/6]
root 33 0.0 0.0 0 0 ? S 2014 0:00 [migration/7]



We are using a default installation. Only nfs is running on the machine.
We did not configure anything kernel related.

Any idea ?
How can we split the work to all cpus ?

Bye, Peer

to me, this looks like the kernel bug referenced here: https://bugs.gentoo.org/show_bug.cgi?id=394487

Are you in a position to open a service request (I guess for SLES11SP2 this would mean having an LTSS contract)?

Best regards,
Jens

mpibgc
05-Mar-2015, 12:55
Hi Peer,


to me, this looks like the kernel bug referenced here: https://bugs.gentoo.org/show_bug.cgi?id=394487

Are you in a position to open a service request (I guess for SLES11SP2 this would mean having an LTSS contract)?

Best regards,
Jens

Yes,

I will open a ticket.
But what I don't understand:
A new DELL R320 does NOT show this behavior (running identical setup), while the old DELL 1950 shows it ?!?

Thanks and bye, Peer

jmozdzen
05-Mar-2015, 13:07
Yes,

I will open a ticket.

That should get you the help neccessary to fix this issue.


But what I don't understand:
A new DELL R320 does NOT show this behavior (running identical setup), while the old DELL 1950 shows it ?!?

As even the kernel bug ticket (https://bugzilla.kernel.org/show_bug.cgi?id=47341) only reports that 3.6.11 fixes the issue, but doesn't mention any root cause, I have no idea what triggers this bug.

Regards,
Jens