FAQ Search Today's Posts Mark Forums Read
» Video Reviews

» Linux Archive

Linux-archive is a website aiming to archive linux email lists and to make them easily accessible for linux users/developers.


» Sponsor

» Partners

» Sponsor

Go Back   Linux Archive > CentOS > CentOS

 
 
LinkBack Thread Tools
 
Old 07-28-2008, 05:32 AM
John R Pierce
 
Default PF_RING crashed the CentOS5 - BUG: soft lockup - CPU#7

Ian jonhson wrote:

I googled the internet and found similar bug occurred in Ubuntu
on the same hardware platform: Dell poweredge.

...

PowerEdge is Dell's brand name for ALL their Server products. That
Ubuntu bug was specific to the i450NX chipset, which was new in 1999 and
obsoleted circa 2001, and supported dual Xeon Pentium II/III processors,
around 300-800Mhz.. How old is your server??!?



_______________________________________________
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos
 
Old 07-28-2008, 05:49 AM
"Ian jonhson"
 
Default PF_RING crashed the CentOS5 - BUG: soft lockup - CPU#7

> CentOS 5 uses kernel 2.6.18-xx ... If this PF_RING thing requires a different kernel, I think a more accurate statement would be, .
> PF_RING is not supported on CentOS.

Indeed. I wonder whether there exists some difference between CentOS
kernel and genral kernel from www.kernel.org?
If so, maybe I should download the CentOS kernel source code and patch
it to run PF_RING.

Any help?

>
> PowerEdge is Dell's brand name for ALL their Server products. That Ubuntu
> bug was specific to the i450NX chipset, which was new in 1999 and obsoleted
> circa 2001, and supported dual Xeon Pentium II/III processors, around
> 300-800Mhz.. How old is your server??!?
>

My machine used to be installed PF_RING is exactly DELL PowerEdge 1950,
with 8-core CPUs and 4G RAM, 80G disk. I don't think it is out of date.
_______________________________________________
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos
 
Old 07-28-2008, 06:33 PM
Scott Silva
 
Default PF_RING crashed the CentOS5 - BUG: soft lockup - CPU#7

on 7-27-2008 10:19 PM Ian jonhson spake the following:

Hi all,

The PF_RING seems not to work smoothly in CentOS 5.
Several day before, I patched the kernel 2.6.25.3 and
installed the PF_RING-patched kernel in my CentOS5.
Based on the PF_RING, I developed my program to capture
the network packages. I wished it can work until the
machine power is off. Unforturnately, no matter how to
adjust my program, whole system can not run more than
48 hours. Finally, whole system crashed and syslogd said
it found the following kernel output.

PF_RING seems to be used for the newest version of ntop for faster packet
capture and analysis. Is that what you are trying to accomplish?

Or did I just get a bad google?



--
MailScanner is like deodorant...
You hope everybody uses it, and
you notice quickly if they don't!!!!

_______________________________________________
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos
 
Old 07-29-2008, 05:29 PM
Scott Silva
 
Default PF_RING crashed the CentOS5 - BUG: soft lockup - CPU#7

on 7-29-2008 9:48 AM Ian jonhson spake the following:

On Tue, Jul 29, 2008 at 3:25 PM, Luca Deri <deri-Q5Vr9lIqpUA@public.gmane.org> wrote:

Ian
from the trace below I don't see an error related to PF_RING but a
soft-lockup (see below).

I have tested PF_RING on 2.6.24 and .26 and have not experienced the problem
you reported.


hmm... I have restarted the operating system with CentOS's original kernel and
done again my experiments. I found that bug has not occurred again.

I don't know why. Maybe, the traditional libpcap has no provided
enough capability
like PF_RING. Or, perhapse something wrong is indeed in patched kernel or Dell
hardware drivers.

BTW, could you tell me what OS distribution you used in your testing PF_RING,
RedHat AS4 or Others? And what kernel source codes are downloaded, from
vendor's websites or www.kernel.org?

I very thank you if you can recommend what OS distribution and what
kernel version
are put together can achieve the maximal stability.

Thank anybody!


Ian
If you need a newer kernel to use that patch, I would recommend something that
runs more cutting edge for that server, maybe Gentoo or Fedora 9. I don't
recommend patching CentOS with a new kernel as you lose the main benefit for
using an Enterprise distro .. stability.
I love CentOS, but no one distribution can be everything for everyone. That is
why there are so many choices.


--
MailScanner is like deodorant...
You hope everybody uses it, and
you notice quickly if they don't!!!!

_______________________________________________
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos
 
Old 07-30-2008, 04:31 PM
Scott Silva
 
Default PF_RING crashed the CentOS5 - BUG: soft lockup - CPU#7

on 7-29-2008 11:10 PM Ian jonhson spake the following:

If you need a newer kernel to use that patch, I would recommend something
that runs more cutting edge for that server, maybe Gentoo or Fedora 9. I
don't recommend patching CentOS with a new kernel as you lose the main
benefit for using an Enterprise distro .. stability.


You meant I should not patch the newest kernel but original kernel, right?
For example, I should download the source codes of kernel -2.6.18 used
in CentOS 5 from www.centos.org and do patching. right?

I am thinking the same method to achieve the highest stability. If you
also recommend the way, I am willing to try again.

Thanks!

Ian
It seems that you tried to patch the CentOS kernel and it didn't work. A patch
for a newer kernel probably wouldn't apply clean, and the CentOS kernels
coming from RHEL are heavily patched already. The last time I played in the
kernel there were at least a hundred patches. Probably more. Any one of those
could contribute to your new patch failing.
I usually wouldn't recommend using a non-standard kernel on CentOS. If the
kernel in the main repo or the plus repo doesn't suit your system, and you add
a kernel from the main kernel.org sources, you risk de-stabilizing your
system. An Enterprise linux like CentOS is designed for stability and long
life and the pieces are fairly intertwined. If you need a custom patch or a
newer kernel, I would just use a distribution like Gentoo or Slackware that
seems to use newer parts, or see if someone already adds your patch to their
current kernel.


--
MailScanner is like deodorant...
You hope everybody uses it, and
you notice quickly if they don't!!!!

_______________________________________________
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos
 
Old 07-31-2008, 04:30 PM
Scott Silva
 
Default PF_RING crashed the CentOS5 - BUG: soft lockup - CPU#7

on 7-30-2008 9:54 PM Ian jonhson spake the following:

It seems that you tried to patch the CentOS kernel and it didn't work.


I am doing this experiment. And, I don't know what would happen.
Different from previous work, I don't patch the kernel from www.kernel.org
but patch the one download from www.centos.org.

I would use the same version (kernel-2.6.18) to do patching in CentOS5.

From the view of Scott and Johnny, I think this is proper way to accomplish

my work.

Thanks.

Ian
That should be safer, but your best bet would be to add the patch to the
bottom of the patch section of the spec file so you don't exclude the
multitude of other patches there. Then cross your fingers and see if it builds
clean.


--
MailScanner is like deodorant...
You hope everybody uses it, and
you notice quickly if they don't!!!!

_______________________________________________
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos
 

Thread Tools




All times are GMT. The time now is 01:42 AM.

VBulletin, Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
Content Relevant URLs by vBSEO ©2007, Crawlability, Inc.
Copyright 2007 - 2008, www.linux-archive.org