I have CentOS 4 running on my server, and I have been having some load issues. I have two fairly large vbulletin forums on the server (between 3k - 5k), and according to vbulletin forums, the CentOS kernel I am currently using (2.6.9-22.0.2.EL) has a major problem with vbulletin, causing hugh loads, etc.

So, I naturally want to upgrade the kernel. I have YUM on a cron, and last night, here is what I received: =============================================================================
Package Arch Version Repository Size
=============================================================================
Updating:
yum noarch 2.4.2-2.centos4 update 383 k

Transaction Summary
=============================================================================
Install 0 Package(s)
Update 1 Package(s)
Remove 0 Package(s)
Total download size: 383 k
Downloading Packages:
Running Transaction Test
Finished Transaction Test
Transaction Test Succeeded
Running Transaction

Updated: yum.noarch 0:2.4.2-2.centos4
Complete!

Now, I understand that 2.4.2-2 is the most current version, and according to this, it should be updated. I rebooted everything, and had my server physically restarted, and when I check, I still have the same kernel on the server.

I obviously am missing a step, but after 3 hours of searching, I can't find it. I appreciate any help you can give me. Thanks!

What kernel are you currently running, give us the output of 'uname -a'. If it is any of the following:

2.6.9-22.0.2.ELsmp #1 SMP
2.6.9-22.0.2.EL #1

You already have the latest kernel upgrade available from CentOS. You can either downgrade your kernel or compile it manually.

Here you go:

Linux 2.6.9-22.0.2.EL #1 Tue Jan 17 06:51:40 CST 2006 i686 athlon i386 GNU/Linux

I asked vbulletin support, was told to upgrade to 2.6.15 or downgrade to 2.6.11. I would prefer to upgrade rather than downgrade. Maybe I am misunderstanding terms.

Oh, new server, so I think to downgrade I have to compile it manually as well.

Do not downgrade. You will need to hand compile it.

Is there a guide or instructions somewhere on how to do that?

OK, Here's the revised edition. Updated to reflect the latest kernel at this time (2.6.16.1) and removed the "make mrproper" command since it was wiping out the old config and causing big issues. :smash: Also some info on "make menuconfig".


cd /usr/src/
http://www.kernel.org/pub/linux/kernel/v2.6/linux-2.6.16.1.tar.gz
tar -zxf linux-2.6.16.1.tar.gz
cd linux-2.6.16.1

cp /boot/config-2.6.9-22.0.2.EL .config
make clean (only necessary if you compile more than once)
make oldconfig


keep hitting enter and take the defaults until you get to the CPU type. Select your CPU. Also in the SMP section select SMP if you have multiple CPUs. Oh, and select "(No Forced Preemption (server))" under that section. Everything else, just hit enter to take the defaults.

(this is the "quick and dirty" method. You can also run "make menuconfig" instead, go to the bottom of the menu, load your old config file, and then select everything in a menu individually if you wish. For most systems, the "make oldconfig" method works fine.)


Then to compile and install:

make bzImage
make modules
make modules_install
make install


Then to configure your boot loader in case of disaster:

nano /boot/grub/grub.conf

default=1
timeout=5
splashimage=(hd0,0)/grub/splash.xpm.gz
hiddenmenu
title CentOS (2.6.16.1)
root (hd0,0)
kernel /vmlinuz-2.6.16.1 ro root=/dev/VolGroup00/LogVol00
initrd /initrd-2.6.16.1.img
title CentOS (2.6.9-22.0.2.EL)
root (hd0,0)
kernel /vmlinuz-2.6.9-22.0.2.EL ro root=/dev/VolGroup00/LogVol00
initrd /initrd-2.6.9-22.0.2.EL.img
title CentOS_ServerCD (2.6.9-22.EL)
root (hd0,0)
kernel /vmlinuz-2.6.9-22.EL ro root=/dev/VolGroup00/LogVol00
initrd /initrd-2.6.9-22.EL.img
..make sure it looks something like this with default=1 and the new kernel (2.6.16.1) on top. (Grub starts at 0, so "default=1" is still your old kernel. You change this later if it works out ok)




..then configure grub to boot to the new kernel one time. This way if it panicks or you can't get to the machine because the network service blows up, you can have the data center do a reboot and it will default back to the last kernel. If it does boot ok and everything looks good, edit grub.conf again and change it to "default=0" and it will boot the new kernel by default after that.

grub
savedefault --default=0 --once
quit

Reboot and see what happens. If it hangs, reboot again and you'll be back to your previous kernel.
Good luck.

sorry,

cp /boot/config-2.6.9-22.0.2.EL /usr/src/linux-2.6.15.4/.config

should instead be:

cp /boot/config-2.6.9-22.0.2.EL .config


(spent too much time in windows today) :P

dirtbag, thanks, just what I was looking for. I took your advice and hired someone to do it this time, didn't want to take any chance. I am going to bookmark this so that I can do it myself next time.

You're welcome. Let us know what kernel you end up with and how it performs in comparison if you could so other CentOS and RHE admins can get more info on this issue.

sorry,

cp /boot/config-2.6.9-22.0.2.EL /usr/src/linux-2.6.15.4/.config

should instead be:

cp /boot/config-2.6.9-22.0.2.EL .config


(spent too much time in windows today) :P hi could you update the previous post with the correct line too

thanks

A mod would likely have to do that eva :)
Anyone who looks at compiling their own kernels etc sould look in depth enough to know that they sould replace the linux-2.6.15.4 with their kernel version.
Else they souldnt be compiling it yet to begin with.

hi could you update the previous post with the correct line too

thanksPost updated. :wht:

A mod would likely have to do that eva :)
Anyone who looks at compiling their own kernels etc sould look in depth enough to know that they sould replace the linux-2.6.15.4 with their kernel version.
Else they souldnt be compiling it yet to begin with.


Yes, I should have mentioned that, thanks. And I couldn't fix the erroneous line because you can only edit your posts in the first 15 minutes and I didn't catch that until later ;) Basically your /boot/config-2.6.9-22.0.2.EL (or whatever version you're running) file needs to be copied into your /usr/src/linux-2.6.15.4/ (or whatever version you're using) directory and renamed ".config"

SSH in Commands : yum upgrade kernel


;)

SSH in Commands : yum upgrade kernel


;)


That only gives you the stock kernel instead of compiling a new one. The current default CentOS kernels have iowait problems and need to be compiled from source.

cd /usr/src/
Then to compile and install:

make bzImage
make modules
make modules_install
make install

i've got many warnings while making...is that matter?

You will get certain warnings about things being depreciated. You can ignore those. I don't know all the particulars about other things, so I can't really tell you for sure without seeing the errors myself. In my experience, you always get a few warnings like that and they can be safely ignored in general. As long as they are "warnings" and not "errors". If you follow the instructions about the grub bootloader, you are pretty safe because a hard reset from the DC will automatically revert it back to your previous kernel. But like I said before.. to be safe about it, I would recommend having Steven at Rack911 for $25 (or someone else that knows what they are doing) do it to be safe.

Wehad similar issues with one of our servers. We ended up having Jon from got-management.com to do it for us for $25. Took about 10 minutes.

Does upgrading kernel effect the data, can it be done on production box?

It doesn't affect the data. It can be done on a production box.

Does upgrading kernel effect the data, can it be done on production box?

Of course you can do it on a production box, just remember to reboot :p

wow... i made it by myself :) thanks for the guide...

btw,

can i delete /usr/src/linux-2.6.15.4 folder?

It will be a problem for the domlog file:
/dev/sda2 4.9G 3.5G 1.2G 76% /usr

not much space left

Don't spend hours optimising/securing your system if you're going to end up using a stock kernel, they are built for EVERY distribution and therefore are full of un-needed modules.

If you can't do it yourself, hire someone to do it for you.

btw,

can i delete /usr/src/linux-2.6.15.4 folder?

It will be a problem for the domlog file:
/dev/sda2 4.9G 3.5G 1.2G 76% /usr

not much space left

Yea you can

Everyone using this CentOS kernel should definitely consider dirtbag's guide - great performance improvement with apache and mysql.

Link would be useful ;)

Link would be useful ;)

I was referring to the guide in the first page of this thread ;)

Just to give this a bump.

I experienced the same symptoms with a new Dual Xeon with CentOS & the 2.6.9-34.ELsmp kernel. Upgraded using instructions in post #6 (changed the filenames ofcourse) to 2.6.15.6 (latest ??), and it also did wonders for my machine.

I've just done this upgrade (to 2.6.16), and load went down, but ip_tables stopped working. Apparently a lot of the netfilter-parameters in the old kernel config isn't valid parameters in the new .config.

Anyone experienced this?

I fixed it. Somehow the netfilter-parameters didn't make it from the old config.

"make menuconfig"
Then press "m" on every choice regarding netfilter and ip_tables.
"make modules"
"make modules_install"

And it worked.

To support the statements of default kernel problems, a small demo. I have two newly built boxes running CentOS 4.3 and .34 kernel and... well I won't say anything. 'Top' says all:

top - 10:25:23 up 1 day, 6:41, 1 user, load average: 0.22, 0.45, 0.53
Tasks: 153 total, 1 running, 152 sleeping, 0 stopped, 0 zombie
Cpu(s): 23.6% us, 5.6% sy, 0.0% ni, 55.1% id, 15.6% wa, 0.0% hi, 0.0% si
Mem: 1033696k total, 271108k used, 762588k free, 6952k buffers
Swap: 2096472k total, 58392k used, 2038080k free, 44924k cached

top - 10:26:20 up 7 days, 19:19, 2 users, load average: 0.39, 0.23, 0.13
Tasks: 162 total, 1 running, 161 sleeping, 0 stopped, 0 zombie
Cpu(s): 2.8% us, 1.0% sy, 0.0% ni, 72.9% id, 23.3% wa, 0.0% hi, 0.0% si
Mem: 1034464k total, 477364k used, 557100k free, 5276k buffers
Swap: 4192956k total, 105268k used, 4087688k free, 75020k cached

Swapping, swapping although plenty free RAM. If some service needs more RAM, stinkin swap rises, and occasional sshd or lmtp connection timeouts with them. On one box, you really need to wait when you kick yum.

To avoid any doubts about HIGH MEM support, it's there:

cat /boot/config-2.6.9-34.EL | grep HIGH
# CONFIG_X86_HIGH_ENTRY is not set
# CONFIG_NOHIGHMEM is not set
CONFIG_HIGHMEM4G=y
# CONFIG_HIGHMEM64G is not set
CONFIG_HIGHMEM=y
CONFIG_HIGHPTE=y
CONFIG_DEBUG_HIGHMEM=y

Several older boxes, upgraded over time to 4.3, now on same kernel:

top - 10:27:36 up 10 days, 21:16, 1 user, load average: 0.03, 0.04, 0.00
Tasks: 88 total, 1 running, 87 sleeping, 0 stopped, 0 zombie
Cpu(s): 0.3% us, 0.1% sy, 0.0% ni, 99.5% id, 0.2% wa, 0.0% hi, 0.0% si
Mem: 2074984k total, 1919608k used, 155376k free, 54004k buffers
Swap: 2096472k total, 80k used, 2096392k free, 1694936k cached

top - 10:28:08 up 14 days, 13:02, 1 user, load average: 0.00, 0.04, 0.06
Tasks: 60 total, 1 running, 59 sleeping, 0 stopped, 0 zombie
Cpu(s): 0.2% us, 0.1% sy, 0.0% ni, 99.3% id, 0.4% wa, 0.0% hi, 0.0% si
Mem: 1034616k total, 196340k used, 838276k free, 38488k buffers
Swap: 2096472k total, 96k used, 2096376k free, 110232k cached

top - 10:28:25 up 14 days, 12:30, 1 user, load average: 0.16, 0.06, 0.05
Tasks: 141 total, 1 running, 140 sleeping, 0 stopped, 0 zombie
Cpu(s): 3.5% us, 0.2% sy, 0.0% ni, 95.5% id, 0.7% wa, 0.0% hi, 0.0% si
Mem: 2074988k total, 1961532k used, 113456k free, 180472k buffers
Swap: 4192924k total, 80k used, 4192844k free, 1318688k cached

Absurd. :puke:

cd /usr/src/
wget http://www.kernel.org/pub/linux/kernel/v2.6/linux-2.6.15.4.tar.gz
tar -zxf linux-2.6.15.4.tar.gz
cd linux-2.6.15.4

cp /boot/config-2.6.9-22.0.2.EL .config
make clean
make mrproper
make oldconfig

... ...

Your order of command above is not using the old config file since you issue make mrproper which will clean .config file. Try the following to prove whether am I right:
tar -zxf linux-2.6.15.4.tar.gz
cd linux-2.6.15.4
cp /boot/config-2.6.9-22.0.2.EL .config
ls -al .config # see whether it is there... yes should be ;)
make clean
ls -al .config # see whether it is there... yes should be ;)
make mrproper
CLEAN .config # this will show to you
ls -al .config # see whether it is there... no, I don't think so :P
make oldconfig
In short, make mrproper will remove all generated files + config + various backup files... ...

Hope this helps ;)

Yes, I was about to ammend this to reflect that. The "make clean" and "make mrproper" can be skipped. In fact, it appears to be better to skip both of those steps altogether. I pieced together this guide from some stuff at Digital Hermit's site and some trial and error. I don't know why some of these guides have you do all this stuff and then tell you to wipe out your configs by running mrproper. I didn't know how thorough that command was until just recently. I never said I was an expert ;)

I followed the steps and noted that the new kernel showed in grub, oddly after reboot it still loaded the old one, even though the new one was listed first

default=1
timeout=5
splashimage=(hd0,0)/grub/splash.xpm.gz
hiddenmenu
title CentOS (2.6.16.1)
root (hd0,0)
kernel /vmlinuz-2.6.16.1 ro root=LABEL=/ selinux=0
initrd /initrd-2.6.16.1.img
title CentOS (2.6.9-34.ELsmp)
root (hd0,0)
kernel /vmlinuz-2.6.9-34.ELsmp ro root=LABEL=/ selinux=0
initrd /initrd-2.6.9-34.ELsmp.img



It loaded 2.6.9-34.ELsmp after reboot. Any idea why?

I followed the steps and noted that the new kernel showed in grub, oddly after reboot it still loaded the old one, even though the new one was listed first

default=1
timeout=5
splashimage=(hd0,0)/grub/splash.xpm.gz
hiddenmenu
title CentOS (2.6.16.1)
root (hd0,0)
kernel /vmlinuz-2.6.16.1 ro root=LABEL=/ selinux=0
initrd /initrd-2.6.16.1.img
title CentOS (2.6.9-34.ELsmp)
root (hd0,0)
kernel /vmlinuz-2.6.9-34.ELsmp ro root=LABEL=/ selinux=0
initrd /initrd-2.6.9-34.ELsmp.img



It loaded 2.6.9-34.ELsmp after reboot. Any idea why?

Grub starts with 0, so really, by choosing 1 you're telling it to use CentOS (2.6.9-34.ELsmp).

I figured so but dirtbag mentioned this:


nano /boot/grub/grub.conf

..make sure it looks like this with default=1 and the new kernel (2.6.15.4) on top

..then configure grub to boot to the new kernel one time. This way if it panicks or you can't get to the machine because the network service blows up, you can have the data center do a reboot and it will go default back to the last kernel. If it does boot ok and everything looks good, edit grub.conf again and change it to "default=0" and it will boot the new kernel by default after that.

grub
savedefault --default=0 --once
quit

Please clarify. I don't want to mess up

By the way, is there a proof safe way of loading the kernel only once in caes something goes wrong?

By the way, is there a proof safe way of loading the kernel only once in caes something goes wrong?


what do you think


grub
savedefault --default=0 --once
quit



does

Right, thanks.

I just noticed a problem:

title CentOS (2.6.16.1)
root (hd0,0)
kernel /vmlinuz-2.6.16.1 ro root=LABEL=/ selinux=0
initrd /initrd-2.6.16.1.img


It is not ELsmp (I have dual xeons). How can I correct this before rebooting?

ignore the elsmp, that is what rehdat puts so you can tell the difference.

if you configuration has smp enabled

cat .config | grep -i smp

then you are fine

root@server [~]# cat .config | grep -i smp
cat: .config: No such file or directory

you have to do it from inside the kernel source

Oh sorry, how do I get there? I've only done this once before (successfully). Plese refresh my memory. Thank you

I dont know. It spends on where you downloaded it to and untared it

Anyone had an issue with using this guide on Intel Dual Cores and Centos 4.x? I can get everything cruising just fine on my new boxes, but when they come back up iptables and APF die.

what kernel did you upgrade? if it's 2.6.16 try this:


I fixed it. Somehow the netfilter-parameters didn't make it from the old config.

"make menuconfig"
Then press "m" on every choice regarding netfilter and ip_tables.
"make modules"
"make modules_install"

And it worked.

OK, Here's the revised edition. Updated to reflect the latest kernel at this time (2.6.16.1) and removed the "make mrproper" command since it was wiping out the old config and causing big issues. :smash: Also some info on "make menuconfig".


cd /usr/src/
http://www.kernel.org/pub/linux/kernel/v2.6/linux-2.6.16.1.tar.gz
tar -zxf linux-2.6.16.1.tar.gz
cd linux-2.6.16.1

cp /boot/config-2.6.9-22.0.2.EL .config
make clean (only necessary if you compile more than once)
make oldconfig


keep hitting enter and take the defaults until you get to the CPU type. Select your CPU. Also in the SMP section select SMP if you have multiple CPUs. Oh, and select "(No Forced Preemption (server))" under that section. Everything else, just hit enter to take the defaults.

(this is the "quick and dirty" method. You can also run "make menuconfig" instead, go to the bottom of the menu, load your old config file, and then select everything in a menu individually if you wish. For most systems, the "make oldconfig" method works fine.)


Then to compile and install:

make bzImage
make modules
make modules_install
make install


Then to configure your boot loader in case of disaster:

nano /boot/grub/grub.conf

default=1
timeout=5
splashimage=(hd0,0)/grub/splash.xpm.gz
hiddenmenu
title CentOS (2.6.16.1)
root (hd0,0)
kernel /vmlinuz-2.6.16.1 ro root=/dev/VolGroup00/LogVol00
initrd /initrd-2.6.16.1.img
title CentOS (2.6.9-22.0.2.EL)
root (hd0,0)
kernel /vmlinuz-2.6.9-22.0.2.EL ro root=/dev/VolGroup00/LogVol00
initrd /initrd-2.6.9-22.0.2.EL.img
title CentOS_ServerCD (2.6.9-22.EL)
root (hd0,0)
kernel /vmlinuz-2.6.9-22.EL ro root=/dev/VolGroup00/LogVol00
initrd /initrd-2.6.9-22.EL.img

..make sure it looks something like this with default=1 and the new kernel (2.6.16.1) on top. (Grub starts at 0, so "default=1" is still your old kernel. You change this later if it works out ok)




..then configure grub to boot to the new kernel one time. This way if it panicks or you can't get to the machine because the network service blows up, you can have the data center do a reboot and it will default back to the last kernel. If it does boot ok and everything looks good, edit grub.conf again and change it to "default=0" and it will boot the new kernel by default after that.

grub
savedefault --default=0 --once
quit

Reboot and see what happens. If it hangs, reboot again and you'll be back to your previous kernel.
Good luck.

Selecting the M on everything for netfilter and iptables under the Network Config still comes up with:

Starting APF:Unable to load iptables module (ipt_state), aborting

I have installed and uninstalled APF attempting to correct, always comes back to iptables.

Ok, did some research and this is a pretty common issue with APF it seems.

http://www.webhostingtalk.com/archive/thread/284575-1.html

rfxn
Quote:
Originally posted by ownagesbot
when i start i get:

lsmod: QM_MODULES: Function not implemented

Unable to load iptables module (ip_tables), aborting.


enable the mono kernel option in conf.apf

I have the 2.6.16 running on an Intel 3.2 Dual Core server now with Raid 1 using the instructions above. Thanks a ton for the guide dirtbag.

I enabled the mono kernel option and now all passive FTP has stopped. Can't open it manually or anything. The kernel is running great and the boxes are super fast, just can't passively FTP due to problems with APF and this kernel.

Will the tutorial work with kernel 2.6.16.7 which was released today?

yes it will.

Yes I can not use ftp on passive mode too.

I fixed it. Somehow the netfilter-parameters didn't make it from the old config.

"make menuconfig"
Then press "m" on every choice regarding netfilter and ip_tables.
"make modules"
"make modules_install"

And it worked.

I'm also experiencing the same problem. could you please tell me if I must recompile the kernel again or do I just have to go into the folder I extracted the kernel to after wget and just doing those three commands?

As long as you choose "m" and not "*" on the missing modules, only the modules needs to be compiled and installed. It should work just fine without compiling the kernel or a reboot.

"service iptables restart" should load the modules and start the firewall, "tail /var/log/messages" for any errors afterwards.

Edit: I just noticed some people are having problems with APF and the new kernels in this thread. If you are using APF compiling the modules might not be enough, I'm only using iptables and it worked for me.

I'm getting this while doing make menuconfig :?

HOSTLD scripts/kconfig/mconf
HOSTCC scripts/kconfig/lxdialog/checklist.o
In file included from scripts/kconfig/lxdialog/checklist.c:24:
scripts/kconfig/lxdialog/dialog.h:31:20: curses.h: No such file or directory
In file included from scripts/kconfig/lxdialog/checklist.c:24:
scripts/kconfig/lxdialog/dialog.h:128: error: syntax error before "use_colors"
scripts/kconfig/lxdialog/dialog.h:128: warning: type defaults to `int' in declaration of `use_colors'
scripts/kconfig/lxdialog/dialog.h:128: warning: data definition has no type or storage class
scripts/kconfig/lxdialog/dialog.h:129: error: syntax error before "use_shadow"
scripts/kconfig/lxdialog/dialog.h:129: warning: type defaults to `int' in declaration of `use_shadow'
scripts/kconfig/lxdialog/dialog.h:129: warning: data definition has no type or storage class
scripts/kconfig/lxdialog/dialog.h:131: error: syntax error before "attributes"
scripts/kconfig/lxdialog/dialog.h:131: warning: type defaults to `int' in declaration of `attributes'
scripts/kconfig/lxdialog/dialog.h:131: warning: data definition has no type or storage class
scripts/kconfig/lxdialog/dialog.h:143: error: syntax error before '*' token
scripts/kconfig/lxdialog/dialog.h:143: warning: function declaration isn't a prototype
scripts/kconfig/lxdialog/dialog.h:146: error: syntax error before '*' token
scripts/kconfig/lxdialog/dialog.h:146: warning: function declaration isn't a prototype
scripts/kconfig/lxdialog/dialog.h:147: error: syntax error before '*' token
scripts/kconfig/lxdialog/dialog.h:147: warning: function declaration isn't a prototype
scripts/kconfig/lxdialog/dialog.h:148: error: syntax error before '*' token
scripts/kconfig/lxdialog/dialog.h:148: warning: function declaration isn't a prototype
scripts/kconfig/lxdialog/dialog.h:149: error: syntax error before '*' token
scripts/kconfig/lxdialog/dialog.h:150: warning: function declaration isn't a prototype
scripts/kconfig/lxdialog/dialog.h:151: error: syntax error before '*' token
scripts/kconfig/lxdialog/dialog.h:151: warning: function declaration isn't a prototype
scripts/kconfig/lxdialog/checklist.c:31: error: syntax error before '*' token
scripts/kconfig/lxdialog/checklist.c:33: warning: function declaration isn't a prototype
scripts/kconfig/lxdialog/checklist.c: In function `print_item':
scripts/kconfig/lxdialog/checklist.c:37: warning: implicit declaration of function `wattrset'
scripts/kconfig/lxdialog/checklist.c:37: error: `win' undeclared (first use in this function)
scripts/kconfig/lxdialog/checklist.c:37: error: (Each undeclared identifier is reported only once
scripts/kconfig/lxdialog/checklist.c:37: error: for each function it appears in.)
scripts/kconfig/lxdialog/checklist.c:38: warning: implicit declaration of function `wmove'
scripts/kconfig/lxdialog/checklist.c:38: error: `choice' undeclared (first use in this function)
scripts/kconfig/lxdialog/checklist.c:40: warning: implicit declaration of function `waddch'
scripts/kconfig/lxdialog/checklist.c:43: error: `selected' undeclared (first use in this function)
scripts/kconfig/lxdialog/checklist.c:44: warning: implicit declaration of function `wprintw'
scripts/kconfig/lxdialog/checklist.c:44: error: `status' undeclared (first use in this function)
scripts/kconfig/lxdialog/checklist.c:47: warning: implicit declaration of function `mvwaddch'
scripts/kconfig/lxdialog/checklist.c:47: error: `item' undeclared (first use in this function)
scripts/kconfig/lxdialog/checklist.c:49: warning: implicit declaration of function `waddstr'
scripts/kconfig/lxdialog/checklist.c:52: warning: implicit declaration of function `wrefresh'
scripts/kconfig/lxdialog/checklist.c: At top level:
scripts/kconfig/lxdialog/checklist.c:59: error: syntax error before '*' token
scripts/kconfig/lxdialog/checklist.c:61: warning: function declaration isn't a prototype
scripts/kconfig/lxdialog/checklist.c: In function `print_arrows':
scripts/kconfig/lxdialog/checklist.c:62: error: `win' undeclared (first use in this function)
scripts/kconfig/lxdialog/checklist.c:62: error: `y' undeclared (first use in this function)
scripts/kconfig/lxdialog/checklist.c:62: error: `x' undeclared (first use in this function)
scripts/kconfig/lxdialog/checklist.c:64: error: `scroll' undeclared (first use in this function)
scripts/kconfig/lxdialog/checklist.c:76: error: `height' undeclared (first use in this function)
scripts/kconfig/lxdialog/checklist.c:79: error: `item_no' undeclared (first use in this function)
scripts/kconfig/lxdialog/checklist.c:79: error: `choice' undeclared (first use in this function)
scripts/kconfig/lxdialog/checklist.c: At top level:
scripts/kconfig/lxdialog/checklist.c:95: error: syntax error before '*' token
scripts/kconfig/lxdialog/checklist.c:96: warning: function declaration isn't a prototype
scripts/kconfig/lxdialog/checklist.c: In function `print_buttons':
scripts/kconfig/lxdialog/checklist.c:97: error: `width' undeclared (first use in this function)
scripts/kconfig/lxdialog/checklist.c:98: error: `height' undeclared (first use in this function)
scripts/kconfig/lxdialog/checklist.c:100: error: `dialog' undeclared (first use in this function)
scripts/kconfig/lxdialog/checklist.c:100: error: `selected' undeclared (first use in this function)
scripts/kconfig/lxdialog/checklist.c: In function `dialog_checklist':
scripts/kconfig/lxdialog/checklist.c:117: error: `WINDOW' undeclared (first use in this function)
scripts/kconfig/lxdialog/checklist.c:117: error: `dialog' undeclared (first use in this function)
scripts/kconfig/lxdialog/checklist.c:117: error: `list' undeclared (first use in this function)
scripts/kconfig/lxdialog/checklist.c:117: warning: left-hand operand of comma expression has no effect
scripts/kconfig/lxdialog/checklist.c:121: warning: implicit declaration of function `endwin'
scripts/kconfig/lxdialog/checklist.c:122: warning: implicit declaration of function `fprintf'
scripts/kconfig/lxdialog/checklist.c:122: error: `stderr' undeclared (first use in this function)
scripts/kconfig/lxdialog/checklist.c:140: error: `COLS' undeclared (first use in this function)
scripts/kconfig/lxdialog/checklist.c:141: error: `LINES' undeclared (first use in this function)
scripts/kconfig/lxdialog/checklist.c:143: error: `stdscr' undeclared (first use in this function)
scripts/kconfig/lxdialog/checklist.c:145: warning: implicit declaration of function `newwin'
scripts/kconfig/lxdialog/checklist.c:146: warning: implicit declaration of function `keypad'
scripts/kconfig/lxdialog/checklist.c:146: error: `TRUE' undeclared (first use in this function)
scripts/kconfig/lxdialog/checklist.c:166: warning: implicit declaration of function `subwin'
scripts/kconfig/lxdialog/checklist.c:199: warning: implicit declaration of function `wnoutrefresh'
scripts/kconfig/lxdialog/checklist.c:201: warning: implicit declaration of function `doupdate'
scripts/kconfig/lxdialog/checklist.c:204: warning: implicit declaration of function `wgetch'
scripts/kconfig/lxdialog/checklist.c:211: error: `KEY_UP' undeclared (first use in this function)
scripts/kconfig/lxdialog/checklist.c:211: error: `KEY_DOWN' undeclared (first use in this function)
scripts/kconfig/lxdialog/checklist.c:221: error: `FALSE' undeclared (first use in this function)
scripts/kconfig/lxdialog/checklist.c:222: warning: implicit declaration of function `scrollok'
scripts/kconfig/lxdialog/checklist.c:223: warning: implicit declaration of function `wscrl'
scripts/kconfig/lxdialog/checklist.c:284: warning: implicit declaration of function `delwin'
scripts/kconfig/lxdialog/checklist.c:288: error: `KEY_LEFT' undeclared (first use in this function)
scripts/kconfig/lxdialog/checklist.c:289: error: `KEY_RIGHT' undeclared (first use in this function)
scripts/kconfig/lxdialog/checklist.c:117: warning: statement with no effect
scripts/kconfig/lxdialog/checklist.c: At top level:
scripts/kconfig/lxdialog/dialog.h:131: warning: array 'attributes' assumed to have one element
make[2]: *** [scripts/kconfig/lxdialog/checklist.o] Error 1
make[1]: *** [menuconfig] Error 2
make: *** [menuconfig] Error 2

Please help.

install the ncurses library

"yum install ncurses"

and

"yum install ncurses-devel"

yum install ncurses*

:)

thanks alot! saved me alot of headach!

this baby does magic.

if you have not upgraded yet, then you should ASAP.

:lovewht:

I have compiled new kernel(2.6.16.16), and when i have server computer and s selected new kernel to boot i get the following error(s):

mkrootdev: labele / not found
Mounting root filesystem
Mount: error 2 mounting ext3
Mount: error 2 mounting none
Switching to new root
Switchroot: mount faild: 22
umount /initd/dev faild: 2
Kernel panic - not syncing: Attempted to kill init!
Does anyone know what is problem?

Has anyone managed to fix the APF/iptables error once you upgrade?

Has anyone managed to fix the APF/iptables error once you upgrade?
I think nobody managed to do that. When i was compiling new kernel i first forget to enable it and then i have to recompile it.

Has anyone managed to fix the APF/iptables error once you upgrade?

Yeah I got it working. I just had to enable all instances of xtables in the config file. I think in a page or two back someone posted how they go iptables working. I also have APF installed and it's working fine. My only problem is that when I go service iptables status it states that my firewall is not working when it is.

because you have to enable apf to run in monokern, you have to open up a port range for ftp to use as passive and set that port range in the pure-ftpd configuration. the reason for this is, conntract_ftp does not work unless its loaded as a module.

I have compiled new kernel(2.6.16.16), and when i have server computer and s selected new kernel to boot i get the following error(s):

mkrootdev: labele / not found
Mounting root filesystem
Mount: error 2 mounting ext3
Mount: error 2 mounting none
Switching to new root
Switchroot: mount faild: 22
umount /initd/dev faild: 2
Kernel panic - not syncing: Attempted to kill init!
Does anyone know what is problem?

I do. You are trying to use a non-selinux kernel on a distribution that expects a kernel with selinux.
I really recomend against using custom kernels on CentOS. If you want to go for custom kernels, you should use LFS, Slackware or Gentoo.
But if you are really willing to get the extra work it WILL require to use non-selinux kernels on CentOS, the _first_step_ is to add "selinux=off" as a kernel paramter.

I'm also experiencing high iowait times (every 30mins on the dot) but am running 2.6.8-022stab070.9-enterprise

Anyone know if this build has the same problem and whether or not upgrading will help? Thanks.

I do. You are trying to use a non-selinux kernel on a distribution that expects a kernel with selinux.
I really recomend against using custom kernels on CentOS. If you want to go for custom kernels, you should use LFS, Slackware or Gentoo.
But if you are really willing to get the extra work it WILL require to use non-selinux kernels on CentOS, the _first_step_ is to add "selinux=off" as a kernel paramter.

Has nothing to do with selinux, you can see the exact same error on a rhel 3 machine, which doesnt have selinux, also if they are just copying over the kernel configuration file selinux will be included wouldnt it?

Has nothing to do with selinux, you can see the exact same error on a rhel 3 machine, which doesnt have selinux, also if they are just copying over the kernel configuration file selinux will be included wouldnt it?

Geez, my bad. Replying too late at night does that.

Anyway, looks like the kernel can't find a filesystem with label "/".
There are many possibilities on what caused that, but that is a secondary issue here.
The way to solve it is to change the root= parameter on the kernel line. Since I don't know how your disk is partitioned, I really can't tell. But it might me something like "root=/dev/sda1" or "root=/dev/hda2". Really depends.

Again, I don't know the kernel loading line. It might have a direct reference root=. In any case, when you build a new kernel, it will have hardcoded inside it a default root partition, which is the current root partition on the machine.

As a side note, it is not enough to copy the .config file to have SELinux, since SELinux is not included on the stock kernel.

Geez, my bad. Replying too late at night does that.

Anyway, looks like the kernel can't find a filesystem with label "/".
There are many possibilities on what caused that, but that is a secondary issue here.
The way to solve it is to change the root= parameter on the kernel line. Since I don't know how your disk is partitioned, I really can't tell. But it might me something like "root=/dev/sda1" or "root=/dev/hda2". Really depends.

Again, I don't know the kernel loading line. It might have a direct reference root=. In any case, when you build a new kernel, it will have hardcoded inside it a default root partition, which is the current root partition on the machine.

As a side note, it is not enough to copy the .config file to have SELinux, since SELinux is not included on the stock kernel.

if they copy over the configuration from the stock kernels (boot partition is where they are getting it from) as the tutorial in this post states selinux will be compiled in:


[root@server1 boot]# grep -i selinux config-2.6.9-22.0.1.EL
CONFIG_SECURITY_SELINUX=y
CONFIG_SECURITY_SELINUX_BOOTPARAM=y
CONFIG_SECURITY_SELINUX_BOOTPARAM_VALUE=1
CONFIG_SECURITY_SELINUX_DISABLE=y
CONFIG_SECURITY_SELINUX_DEVELOP=y
CONFIG_SECURITY_SELINUX_AVC_STATS=y
# CONFIG_SECURITY_SELINUX_MLS is not set
[root@server1 boot]#

The way to solve it is to change the root= parameter on the kernel line. Since I don't know how your disk is partitioned, I really can't tell. But it might me something like "root=/dev/sda1" or "root=/dev/hda2". Really depends.


You can find the partition by running df -h and finding the / partition.

if they copy over the configuration from the stock kernels (boot partition is where they are getting it from) as the tutorial in this post states selinux will be compiled in:

I wonder how, since the stock kernel doesn't have the SELinux code in it.
Just because it is on the config files doesn't mean it will be compiled.

All this is going to do is turn into a pissing fight for the rest of our lifes on wht.

Anyways since when is there no selinux code?

http://www.nsa.gov/selinux/code/download5.cfm



SELinux support is already included in the mainline Linux 2.6 kernel


But anyways:


[root@server1 linux-2.6.16.19]# find . -name "*selinux*"
./security/selinux
./security/selinux/selinuxfs.c
./include/linux/selinux_netlink.h
[root@server1 linux-2.6.16.19]#

[root@server1 selinux]# ls
avc.c hooks.c include Kconfig Makefile netif.c netlink.c nlmsgtab.c selinuxfs.c ss xfrm.c
[root@server1 selinux]#

Selinux apears to be there..

http://img437.imageshack.us/img437/6069/selinux1kz.jpg


morcego,

Is there another part to selinux that I am missing?

do you happen to have ext3 compiled in there somewhere?

you can also disable selinux by adding selinux=0 to your kernel command line
or at /etc/selinux/config if you need to

I fixed the problem now. It was probem that i have compiled kernel with ext3 as module and not directly. Now it is ok. Tnx to all that helped me.

Is it neccessary to upgrade kernel to latest ones? What does it fixed usually. Please teach me because i want to learn more about linux =)

Is it neccessary to upgrade kernel to latest ones? What does it fixed usually. Please teach me because i want to learn more about linux =)

Is it necessary? This depends on your definition of "necessary", and your uses for the related server.

If the server is going to be online, then yes, it is advisable to update the kernel, as often as your distro updates it. Why? Simply put: bug fixes and exploits closed.

Each kernel release will put out bugfixes (usually a few), close a security exploit or two, and add new features. The only way to know FOR SURE what has been changed/added/patched/fixed in the new release? Read the README/DOC files.

With manually compiled kernels, especially, it's quite important to keep up to date with what's going on in the new releases. With distro released (rpm, dep, etc) kernels, you will probably want to upgrade those when they come out, because usually it's pretty important.


Please teach me because i want to learn more about linux

The kernel is the LAST place you want to start whyen learning about linux. The first place? Your local bookstore. Get yourself a few linux books, read them, understand them. Then, live with linux only for a few years (no windows, strictly linux). By that time, you might have caught up to a few of the gurus around here ;)

I'm using FreeBSD latest Version and i'm usign the system for Gaming purposes .. So it is better to upgrade to the latest kernels? Thanks ^^ For teaching me

Freebsd is not linux. You cannot upgrade the kernel in this way.

Ok first of all this is abit off topic. But can you teach me how to upgrade to latest KERNEL for freebsd thanks ;D

Ok first of all this is abit off topic. But can you teach me how to upgrade to latest KERNEL for freebsd thanks ;D
Before you ask such off-topic question, will you mind to spend couple minutes to read in their handbook?
http://www.freebsd.org/doc/en_US.ISO8859-1/books/handbook/
especially the below?
http://www.freebsd.org/doc/en_US.ISO8859-1/books/handbook/kernelconfig.html

Sorry for the delay. Missed this post of yours.

[root@server1 linux-2.6.16.19]# find . -name "*selinux*"
./security/selinux
./security/selinux/selinuxfs.c
./include/linux/selinux_netlink.h
[root@server1 linux-2.6.16.19]#
morcego,

Is there another part to selinux that I am missing?

Aparently not.
My mistake. I stand corrected.

Just a FYI Morcego,

Selinux has been in mainline kernel since before 2.6.9, so adding selinux wasent just a centos thing considering 2.6.9 is what centos is based off, they mainly added patches that were not in the mainline kernel at the time but released by selinux.

Lets get further into the precision then :)
CentOS doesn't include (or remove) anything on the kernel. That is done by RedHat.

Lets get further into the precision then :)
CentOS doesn't include (or remove) anything on the kernel. That is done by RedHat.

Maybe not the kernel. But centos has removed things in the past from other packages such as the centos 3.x project removing things from certian packages that relied on NTPL (ex. BIND) due to them not working with out hacking up the /lib/tls directory when using a non-stock kernel. Really wouldn't surprise me if they did add their own patches to the kernel.

Hummm, CentOS still has a "custom" line of packages, but that stays at a different repository (CentOSPlus), which is disabled by default.
And, of course, there are always Karan's and Dag's repositories, which are recomended as standard fare even on the centos list (specially since Karan is one of the hands behind CentOS).

Yeah I got it working. I just had to enable all instances of xtables in the config file. I think in a page or two back someone posted how they go iptables working. I also have APF installed and it's working fine. My only problem is that when I go service iptables status it states that my firewall is not working when it is.
Hi

I have done. I have recompiled and enabled all the netfilter and ip_tables instances and I then enabled monokern but it still fails to load.

All I get is:

iptables v1.2.11: can't initialize iptables table `filter': Table does not exist (do you need to insmod?) Perhaps iptables or your kernel needs to be upgraded.

Intel Pentium D 930
>6. Pentium-Pro (M686)
10. Pentium-4/Celeron(P4-based)/Pentium-4 M/Xeon (MPENTIUM4)

which one ?

The Pentium D series is a dual-core P4 that uses the Pentium-M microarchitecture (according to http://www.intel.com/design/pentium4/manuals/248966.htm). so choose #10.

Don't forget you want the SMP option too.

For Some reason, On my AMD machines, When setting the User_hZ to 1000 the server seems to be running at 250hZ. Rather than the Intel actually runs at 1000hZ user space. I tested this by my game servers running at 333FPS rather on intel which is 500FPS.

I've tried to upgrade the kernel twice, but both didnt work. It wont boot, where can I see the logs for the reason why it wont boot?
I just left everything on default, except cpu and forced preemption . Are there more settings where I should pay more attention to it?
My processor: AMD64 3500+

Have you installed everything nessicary, (I don't know if thats spelled right, I'm tired), to build the kernel? I know some of my AMD machines didn't build it because it was missing somethings that were needed to compile.

I believe I did, I followed the howto on page 4.

I've tried to upgrade the kernel twice, but both didnt work. It wont boot, where can I see the logs for the reason why it wont boot?
I just left everything on default, except cpu and forced preemption . Are there more settings where I should pay more attention to it?
My processor: AMD64 3500+

with the newer kernels, you cannot really jsut copy the kernel over and have it work. there have been hundreds of changes you may need to account for.

According to the tutorial, I can leave everything on default except CPU settings.

Hi!

I've been lurking around for weeks on this forum but thought I'd join in the fun! I am also a member over at Geek/Talk (s.herzig), which most should know.

I ordered a server about 2 or 3 weeks ago, which replaced one of the older ones I had before. The new one is an AMD 64 Athlon 3200, with 1 Gig of RAM and CentOS 4.x installed (kernel version 2.6.9-34.0.1.EL, can you guess the problem I am about to describe already?) The site that I am hosting on this server has had no problems in terms of load before (the load went up to 3.0 at max), but ever since I switched over to the new server I am experiencing loads of up to 60.0! CPU usage is relatively low, considering its average is at about 20-25%, and the memory usage is at 172MB (for some weird reason it's at +-300MB for the first day after a reboot, but then it drops down to 172MB). The system seems to love to swap though, as the swap is at about 178MB. As described in many other replies in this topic, the I/O-Wait value that I see when looking at "top", is at about 90-98% max, which is obviously the source of the high load. This happens especially when a memory consuming php script is run directly, which then sends out a newsletter.

Now, I've googled and yahooed around a bit for days, trying to find a solution to this problem, even spent sleepless nights. One solution was to upgrade the kernel to version 2.6.9-39.EL (testing stage) (this solution came directly from CentOS bug central), but that did not work. I tried installing the given RPM package, but when I rebooted, a kernel panic message appeared, saying something about init/rc not being able to initialize (?). So that didnt work. I then tried the next solution I found, downgrading the kernel to version 2.6.16.1, and compiling it from source. I tried that, but again, the system didn't come up again. I also tried "yum upgrade kernel" from ssh, but alas, it found no upgrade.

Sooo, I'm basically stuck now with a freshly set up server, with a buggy kernel. I've talked to the hosting company I am with, but apparently this is my own problem, unless I want to pay them $75 for punshing in a few lines of code and making a cup of coffee while they wait for the source to compile. No thanks.

So, has anyone got any further ideas? Or faced this very same problem and found a solution to it? Any comments are greatly appreciated, as I am loosing customers at the moment because of this...

Cheers,

-Sebastian

Wow, I just ran "yum upgrade kernel*" again, and see what it did:

[root@server1 src]# yum upgrade kernel*
Repository extras is listed more than once in the configuration
Repository updates is listed more than once in the configuration
Repository base is listed more than once in the configuration
Repository addons is listed more than once in the configuration
Setting up Upgrade Process
Setting up repositories
extras 100% |=========================| 1.1 kB 00:00
updates 100% |=========================| 951 B 00:00
base 100% |=========================| 1.1 kB 00:00
addons 100% |=========================| 951 B 00:00
Reading repository metadata in from local files
Examining kernel-2.6.9-39.EL.i686.rpm: kernel - 2.6.9-39.EL.i686
Examining kernel-2.6.9-39.EL.x86_64.rpm: kernel - 2.6.9-39.EL.x86_64
Examining kernel-devel-2.6.9-34.19.EL.x86_64.rpm: kernel-devel - 2.6.9-34.19.EL.x86_64
Examining kernel-devel-2.6.9-39.EL.i686.rpm: kernel-devel - 2.6.9-39.EL.i686
Examining kernel-devel-2.6.9-39.EL.x86_64.rpm: kernel-devel - 2.6.9-39.EL.x86_64
Marking kernel-2.6.9-39.EL.i686.rpm as an update to kernel - 2.6.9-34.0.1.EL.i686
Marking kernel-2.6.9-39.EL.i686.rpm as an update to kernel - 2.6.9-34.EL.i686
Marking kernel-devel-2.6.9-34.19.EL.x86_64.rpm as an update to kernel-devel - 2.6.9-34.EL.i686
Marking kernel-devel-2.6.9-34.19.EL.x86_64.rpm as an update to kernel-devel - 2.6.9-34.0.1.EL.i686
Marking kernel-devel-2.6.9-39.EL.i686.rpm as an update to kernel-devel - 2.6.9-34.EL.i686
Marking kernel-devel-2.6.9-39.EL.i686.rpm as an update to kernel-devel - 2.6.9-34.0.1.EL.i686
Marking kernel-devel-2.6.9-39.EL.x86_64.rpm as an update to kernel-devel - 2.6.9-34.EL.i686
Marking kernel-devel-2.6.9-39.EL.x86_64.rpm as an update to kernel-devel - 2.6.9-34.0.1.EL.i686
kernel-2.6.9-39.EL.i686.rpm: does not update installed package.
kernel-2.6.9-39.EL.x86_64.rpm: does not update installed package.
kernel-2.6.9-39.EL.x86_64.rpm: does not update installed package.
kernel-2.6.9-39.EL.x86_64.rpm: does not update installed package.
kernel-devel-2.6.9-34.19.EL.x86_64.rpm: does not update installed package.
kernel-devel-2.6.9-39.EL.i686.rpm: does not update installed package.
kernel-devel-2.6.9-39.EL.x86_64.rpm: does not update installed package.
Could not find update match for kernels
Resolving Dependencies
--> Populating transaction set with selected packages. Please wait.
---> Package kernel-devel.x86_64 0:2.6.9-39.EL set to be installed
---> Package kernel.i686 0:2.6.9-39.EL set to be installed
---> Package kernel-devel.x86_64 0:2.6.9-34.19.EL set to be installed
warning: package kernel-devel = 2.6.9-39.EL was already added, replacing with kernel-devel <= 2.6.9-39.EL
---> Package kernel-devel.i686 0:2.6.9-39.EL set to be installed
--> Running transaction check
Dependencies Resolved
=============================================================================
Package Arch Version Repository Size
=============================================================================
Installing:
kernel i686 2.6.9-39.EL kernel-2.6.9-39.EL.i686.rpm 27 M
kernel-devel x86_64 2.6.9-39.EL kernel-devel-2.6.9-39.EL.x86_64.rpm 11 M
kernel-devel x86_64 2.6.9-34.19.EL kernel-devel-2.6.9-34.19.EL.x86_64.rpm 11 M
kernel-devel i686 2.6.9-39.EL kernel-devel-2.6.9-39.EL.i686.rpm 11 M
Transaction Summary
=============================================================================
Install 4 Package(s)
Update 0 Package(s)
Remove 0 Package(s)
Total download size: 61 M
Is this ok [y/N]: y
Downloading Packages:
Running Transaction Test
warning: package kernel-devel = 2.6.9-39.EL was already added, replacing with kernel-devel <= 2.6.9-39.EL
Finished Transaction Test

Transaction Check Error: package kernel-devel-2.6.9-39.EL is already installed
package kernel-2.6.9-39.EL is already installed
package kernel-devel-2.6.9-34.19.EL is intended for a x86_64 architecture

(discard last post)

OK, just an update, I upgraded from kernel 2.6.9-34.0.1.EL to 2.6.9-34.0.2.EL using yum BUT after a reboot the server didnt come up again! This is too weird for me...

cat /etc/modprobe.conf

Has anyone come up with a guide for RHEL 4 and the newer kernels? Quite a bit has changed.

@Steven

Thanks! I punched it into SSH, it showed a couple of lines, but uhm... what it exactly did I cant tell you. I've posted the output below. Will the new kernel work now :o)?

# cat /etc/modprobe.conf
alias eth0 tg3
alias scsi_hostadapter sata_sil
alias snd-card-0 snd-atiixp
options snd-card-0 index=0
install snd-atiixp /sbin/modprobe --ignore-install snd-atiixp && /usr/sbin/alsactl restore >/dev/null 2>&1 || :
remove snd-atiixp { /usr/sbin/alsactl store >/dev/null 2>&1 || : ; }; /sbin/modprobe -r --ignore-remove snd-atiixp
alias usb-controller ehci-hcd
alias usb-controller1 ohci-hcd

based off that output the kernels should work fine.

They should or they most probably will now?

Any idea what would cause this panic error? I am guessing its related to the file system options I've been trying to compile 2.6.17.6 on an RHEL 4 test box.

Using IPI Shortcut mode
Freeing unused kernel memory: 184k freed
Red Hat nash version 4.2.1.6 starting


Kernel panic - not syncing: Attempted to kill init!

ounting sysfs
Creating /dev

pixelized

please paste:

lspci
cat /etc/modprobe.conf
df -h

Steven:
Thanks for your reply, the kernels should work fine, yes, but they don't. The datacenter still reports a "Kernel Panic", whenever the system tries to boot.

I guess the best solution would be to hire someone to do it!

-Sebastian

root@dev2 [~]# lspci
00:01.0 PCI bridge: Broadcom HT1000 PCI/PCI-X bridge
00:02.0 Host bridge: Broadcom HT1000 Legacy South Bridge
00:02.1 IDE interface: Broadcom HT1000 Legacy IDE controller
00:02.2 ISA bridge: Broadcom HT1000 LPC Bridge
00:03.0 USB Controller: Broadcom HT1000 USB Controller (rev 01)
00:03.1 USB Controller: Broadcom HT1000 USB Controller (rev 01)
00:03.2 USB Controller: Broadcom HT1000 USB Controller (rev 01)
00:05.0 VGA compatible controller: ATI Technologies Inc Rage XL (rev 27)
00:18.0 Host bridge: Advanced Micro Devices [AMD] K8 [Athlon64/Opteron] HyperTrn
00:18.1 Host bridge: Advanced Micro Devices [AMD] K8 [Athlon64/Opteron] Addressp
00:18.2 Host bridge: Advanced Micro Devices [AMD] K8 [Athlon64/Opteron] DRAM Cor
00:18.3 Host bridge: Advanced Micro Devices [AMD] K8 [Athlon64/Opteron] Miscelll
01:0d.0 PCI bridge: Broadcom HT1000 PCI/PCI-X bridge (rev b2)
01:0e.0 RAID bus controller: Broadcom BCM5785 (HT1000) SATA Native SATA Mode
02:03.0 Ethernet controller: Broadcom Corporation NetXtreme BCM5704 Gigabit Eth)
02:03.1 Ethernet controller: Broadcom Corporation NetXtreme BCM5704 Gigabit Eth)

root@edge2 [/etc]# cat modprobe.conf
Cannot find old version of modprobe. Giving up.

root@edge2 [/etc]# cat modprobe.conf~
alias eth0 tg3
alias eth1 tg3
alias scsi_hostadapter sata_svw
alias usb-controller ehci-hcd

root@edge2 [/etc]# df -h
Filesystem Size Used Avail Use% Mounted on
/dev/sda5 9.9G 683M 8.7G 8% /
/dev/sda1 99M 14M 80M 15% /boot
none 506M 0 506M 0% /dev/shm
/dev/sda7 197G 5.5G 182G 3% /home
/dev/sda8 1012M 57M 904M 6% /tmp
/dev/sda3 9.9G 3.5G 6.0G 37% /usr
/dev/sda2 9.9G 601M 8.8G 7% /var
/dev/sdb1 230G 11G 208G 5% /backup
/tmp 1012M 57M 904M 6% /var/tmp

root@edge2 [/etc]# cat modprobe.conf
Cannot find old version of modprobe. Giving up.


There is one of your problems.

Try this:

cp /etc/modprobe.conf~ /etc/modprobe.conf
wget http://kernel.org/pub/linux/kernel/v2.6/linux-2.6.17.6.tar.bz2
tar jxf linux-2.6.17.6.tar.bz2
cd linux-2.6.17.6
wget layer1.rack911.com/pixelized-kernel -O .config
make oldconfig
make bzImage
make modules
make modules_install
make install

setup your bootloader

reboot




KEEP in mind this is setup for a dual xeon, you may need to make menuconfig and change the processor type and SMP settings in the menu named "Processor Type and Features"

Alright Steve I'll give that a shot. Am I supposed to use the Elan SubArchitecture for Opteron processors?

Dont worry about that.

Awesome it booted this time. Server seems to be using barely any swap as compared with before.

top - 10:25:27 up 3 min, 1 user, load average: 0.24, 0.32, 0.14
Tasks: 113 total, 1 running, 112 sleeping, 0 stopped, 0 zombie
Cpu0 : 0.7% us, 0.0% sy, 0.0% ni, 99.3% id, 0.0% wa, 0.0% hi, 0.0% si
Cpu1 : 0.3% us, 0.0% sy, 0.0% ni, 99.7% id, 0.0% wa, 0.0% hi, 0.0% si
Mem: 1036388k total, 340028k used, 696360k free, 22032k buffers
Swap: 2096440k total, 0k used, 2096440k free, 137972k cached

Awesome it booted this time. Server seems to be using barely any swap as compared with before.

top - 10:25:27 up 3 min, 1 user, load average: 0.24, 0.32, 0.14
Tasks: 113 total, 1 running, 112 sleeping, 0 stopped, 0 zombie
Cpu0 : 0.7% us, 0.0% sy, 0.0% ni, 99.3% id, 0.0% wa, 0.0% hi, 0.0% si
Cpu1 : 0.3% us, 0.0% sy, 0.0% ni, 99.7% id, 0.0% wa, 0.0% hi, 0.0% si
Mem: 1036388k total, 340028k used, 696360k free, 22032k buffers
Swap: 2096440k total, 0k used, 2096440k free, 137972k cached

it shoudl use less swap, and have better memory management and load.

WOW, compare this to my latest "top" result:

Cpu(s): 2.6% us, 2.6% sy, 0.0% ni, 0.7% id, 93.4% wa, 0.7% hi, 0.0% si
Mem: 1018520k total, 195944k used, 822576k free, 10476k buffers
Swap: 1959920k total, 145296k used, 1814624k free, 25036k cached

I dont think it is necessary to compile a custom kernel on centos. Im sure some people who did were got by these rash of latest exploits for the new kernels. They fixed the exploit in proc but there is still up to 2 other locals that still work on the new one.
I am running the latest centos kernel, when all; those exploits came out I tried each one and they didnt work, they easily work on the new kernels. I reccomend anyone using redhat or centos to stick with the enterprise kernels. New doesnt always mean better and these kernels already have extensive patching.

Well, it's still not working here! We now have tried upgrading and downgrading the kernel, changing the configs, nearly everything possible, but the system is still only using <200MB of RAM and is swapping about as much all the time. Loads are relatively OK, but I don't want the hard drive to crash because of too much I/O! (The I/O wait value is at 28% average, but at high load times at 98%)

Even after a "yum upgrade kernel", which was our last, very desperate attempt at getting the server back to normal, the system didn't reboot again and a datacenter technician had to manually change the grub config again! (Upgrade from kernel 2.6.9-34.0.1.EL to 2.6.9-34.0.2.EL, didnt work, now running 2.6.9-34.EL! Kernel 2.6.9-42.EL failed to boot as well, so did Kernel 2.6.16.1... but the compilation, resp. the rpm packages were compiled/extracted/installed just fine!)

Help ^^!

-Sebastian

sounds like you got hardware issues

But the server is brand new? Not even a month old?

The RHEL/CentOS high load problem was identified here:

https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=188141 (https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=188141)


And applied to the kernel development stream here:

Wed Apr 19 2006 Jason Baron <jbaron@redhat.com (jbaron@redhat.com)> [2.6.9-34.20]

"-revert: vm: wired pagetables patch (Larry Woodman) [188141 185110]"


(notice this is version 2.6.9-34.20 NOT 2.6.9-34.2)

If you want to compile your "supported" 2.6.9-34.2 kernel by hand with this fix, the patch info is in this section specifically (comment #7):

https://bugzilla.redhat.com/bugzilla...i?id=188141#c7 (https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=188141#c7)

OK, Here's the revised edition. Updated to reflect the latest kernel at this time (2.6.16.1) and removed the "make mrproper" command since it was wiping out the old config and causing big issues. :smash: Also some info on "make menuconfig".


cd /usr/src/
http://www.kernel.org/pub/linux/kernel/v2.6/linux-2.6.16.1.tar.gz
tar -zxf linux-2.6.16.1.tar.gz
cd linux-2.6.16.1

cp /boot/config-2.6.9-22.0.2.EL .config
make clean (only necessary if you compile more than once)
make oldconfig


keep hitting enter and take the defaults until you get to the CPU type. Select your CPU. Also in the SMP section select SMP if you have multiple CPUs. Oh, and select "(No Forced Preemption (server))" under that section. Everything else, just hit enter to take the defaults.

(this is the "quick and dirty" method. You can also run "make menuconfig" instead, go to the bottom of the menu, load your old config file, and then select everything in a menu individually if you wish. For most systems, the "make oldconfig" method works fine.)


Then to compile and install:

make bzImage
make modules
make modules_install
make install


Then to configure your boot loader in case of disaster:

nano /boot/grub/grub.conf

default=1
timeout=5
splashimage=(hd0,0)/grub/splash.xpm.gz
hiddenmenu
title CentOS (2.6.16.1)
root (hd0,0)
kernel /vmlinuz-2.6.16.1 ro root=/dev/VolGroup00/LogVol00
initrd /initrd-2.6.16.1.img
title CentOS (2.6.9-22.0.2.EL)
root (hd0,0)
kernel /vmlinuz-2.6.9-22.0.2.EL ro root=/dev/VolGroup00/LogVol00
initrd /initrd-2.6.9-22.0.2.EL.img
title CentOS_ServerCD (2.6.9-22.EL)
root (hd0,0)
kernel /vmlinuz-2.6.9-22.EL ro root=/dev/VolGroup00/LogVol00
initrd /initrd-2.6.9-22.EL.img
..make sure it looks something like this with default=1 and the new kernel (2.6.16.1) on top. (Grub starts at 0, so "default=1" is still your old kernel. You change this later if it works out ok)




..then configure grub to boot to the new kernel one time. This way if it panicks or you can't get to the machine because the network service blows up, you can have the data center do a reboot and it will default back to the last kernel. If it does boot ok and everything looks good, edit grub.conf again and change it to "default=0" and it will boot the new kernel by default after that.

grub
savedefault --default=0 --once
quit

Reboot and see what happens. If it hangs, reboot again and you'll be back to your previous kernel.
Good luck.
Hi,
when i going to /boot , i dont see anything files on this like /boot/config-2.6.9-22.0.2.EL ...
there is no file on this folder. what is the problem?

I dont think it is necessary to compile a custom kernel on centos. Im sure some people who did were got by these rash of latest exploits for the new kernels. They fixed the exploit in proc but there is still up to 2 other locals that still work on the new one.
I am running the latest centos kernel, when all; those exploits came out I tried each one and they didnt work, they easily work on the new kernels. I reccomend anyone using redhat or centos to stick with the enterprise kernels. New doesnt always mean better and these kernels already have extensive patching.

Are you sure about this? Can you provide proof instead of just speculation?

Well I preffer grsecurity now because I have shell users. I did find out I was wrong on one thing then, that kernel was still vuln to the proc race unless you mounted proc as nosuid.
The current centos kernel is suposed to be patched against that now.
As far as novice users I wouldnt reccomend it. If someone is gonna experiment with kernels I advise a datacenter that doesnt care to boot back up for you in case you have a kernel panic.

Oh and yeah 2.6.17.4 and earlier could be rooted in a heartbeat then with the h0olyshit proc race and the raptor exploit

I'm pretty sure. the kernel development mailing list describes the security bugs that have been fixed in each release. If you're not running the very latest, you will be exposed to some security bug that is now documented :(

The redhat kernels do get all the security fixes ported into their kernel, so if you use them, you do not get the latest features, but you also do not get the latest bugs either.

Redhat are as open about which security issues are fixed in their kernels as the kernel team are in their release notes.

(root@myserver)root@myserver [/usr/src/linux-2.6.16.1]# make oldconfig
.config:1: *** missing separator. Stop.
make: *** [scripts_basic] Error 2
(root@myserver)root@myserver [/usr/src/linux-2.6.16.1]# (root@myserver)

How can i solve this problem ? :eek:
thanks

why would you be compiling such and old kernel anyway? You are compiling something that is already vulnerable to all the latest local root exploits.
make sure you have your copied config in there and then try make menuconfig.
Read the instructions carefully on this tut, you do not copy and paste word for word, you replace some things with what you have or installing
Hope that helps. And dont compile that kernel there unless you want your server to be rooted if someone ever gets shell

I'm currently getting one of these errors :(

CC drivers/ide/pci/cy82c693.o
CC drivers/ide/pci/hpt34x.o
CC drivers/ide/pci/hpt366.o
CC drivers/ide/pci/pdc202xx_old.o
CC drivers/ide/pci/pdc202xx_new.o
CC drivers/ide/pci/piix.o
CC drivers/ide/pci/rz1000.o
CC drivers/ide/pci/serverworks.o
CC drivers/ide/pci/siimage.o
CC drivers/ide/pci/sis5513.o
CC drivers/ide/pci/slc90e66.o
CC drivers/ide/pci/triflex.o
CC drivers/ide/pci/via82cxxx.o
drivers/ide/pci/via82cxxx.c:85: error: `PCI_DEVICE_ID_VIA_8237A' undeclared here (not in a function)
drivers/ide/pci/via82cxxx.c:85: error: initializer element is not constant
drivers/ide/pci/via82cxxx.c:85: error: (near initialization for `via_isa_bridges[3].id')
drivers/ide/pci/via82cxxx.c:85: error: initializer element is not constant
drivers/ide/pci/via82cxxx.c:85: error: (near initialization for `via_isa_bridges[3]')
drivers/ide/pci/via82cxxx.c:86: error: initializer element is not constant
drivers/ide/pci/via82cxxx.c:86: error: (near initialization for `via_isa_bridges[4]')
drivers/ide/pci/via82cxxx.c:87: error: initializer element is not constant
drivers/ide/pci/via82cxxx.c:87: error: (near initialization for `via_isa_bridges[5]')
drivers/ide/pci/via82cxxx.c:88: error: initializer element is not constant
drivers/ide/pci/via82cxxx.c:88: error: (near initialization for `via_isa_bridges[6]')
drivers/ide/pci/via82cxxx.c:89: error: initializer element is not constant
drivers/ide/pci/via82cxxx.c:89: error: (near initialization for `via_isa_bridges[7]')
drivers/ide/pci/via82cxxx.c:90: error: initializer element is not constant
drivers/ide/pci/via82cxxx.c:90: error: (near initialization for `via_isa_bridges[8]')
drivers/ide/pci/via82cxxx.c:91: error: initializer element is not constant
drivers/ide/pci/via82cxxx.c:91: error: (near initialization for `via_isa_bridges[9]')
drivers/ide/pci/via82cxxx.c:92: error: initializer element is not constant
drivers/ide/pci/via82cxxx.c:92: error: (near initialization for `via_isa_bridges[10]')
drivers/ide/pci/via82cxxx.c:93: error: initializer element is not constant
drivers/ide/pci/via82cxxx.c:93: error: (near initialization for `via_isa_bridges[11]')
drivers/ide/pci/via82cxxx.c:94: error: initializer element is not constant
drivers/ide/pci/via82cxxx.c:94: error: (near initialization for `via_isa_bridges[12]')
drivers/ide/pci/via82cxxx.c:95: error: initializer element is not constant
drivers/ide/pci/via82cxxx.c:95: error: (near initialization for `via_isa_bridges[13]')
drivers/ide/pci/via82cxxx.c:96: error: initializer element is not constant
drivers/ide/pci/via82cxxx.c:96: error: (near initialization for `via_isa_bridges[14]')
drivers/ide/pci/via82cxxx.c:97: error: initializer element is not constant
drivers/ide/pci/via82cxxx.c:97: error: (near initialization for `via_isa_bridges[15]')
drivers/ide/pci/via82cxxx.c:98: error: initializer element is not constant
drivers/ide/pci/via82cxxx.c:98: error: (near initialization for `via_isa_bridges[16]')
drivers/ide/pci/via82cxxx.c:99: error: initializer element is not constant
drivers/ide/pci/via82cxxx.c:99: error: (near initialization for `via_isa_bridges[17]')
drivers/ide/pci/via82cxxx.c:100: error: initializer element is not constant
drivers/ide/pci/via82cxxx.c:100: error: (near initialization for `via_isa_bridges[18]')
drivers/ide/pci/via82cxxx.c:101: error: initializer element is not constant
drivers/ide/pci/via82cxxx.c:101: error: (near initialization for `via_isa_bridges[19]')
drivers/ide/pci/via82cxxx.c:102: error: initializer element is not constant
drivers/ide/pci/via82cxxx.c:102: error: (near initialization for `via_isa_bridges[20]')
drivers/ide/pci/via82cxxx.c:103: error: initializer element is not constant
drivers/ide/pci/via82cxxx.c:103: error: (near initialization for `via_isa_bridges[21]')
make[3]: *** [drivers/ide/pci/via82cxxx.o] Error 1
make[2]: *** [drivers/ide/pci] Error 2
make[1]: *** [drivers/ide] Error 2
make: *** [drivers] Error 2

Please help.

i tried to install kernel 2.6.20 on centos 4 and in last step i got following message:

[root@cc]# make install
sh /usr/src/linux-2.6.20/arch/i386/boot/install.sh 2.6.20 arch/i386/boot/bzImage System.map "/boot"
WARNING: No module ata_piix found for kernel 2.6.20, continuing anyway

what this mean ?

In the later kernels SATA drivers got split into a new menu section.

run make menuconfig
go down to device drivers
then SATA drivers
set libata as a module
find ata_piix set it as a module or it shows up as Intel piixn in menuconfig
recompile and you're good to go

modprobe trips that message because it can't find that module in your configuration

make sure you properly inventory your system using lspci and lsmod

i tried to install kernel 2.6.20 on centos 4 and in last step i got following message:

[root@cc]# make install
sh /usr/src/linux-2.6.20/arch/i386/boot/install.sh 2.6.20 arch/i386/boot/bzImage System.map "/boot"
WARNING: No module ata_piix found for kernel 2.6.20, continuing anyway

what this mean ?

I'm getting the same thing..

Hello,

the upgrade way on post#6 don't work with servers from SoftLayer why ?

Could be that you need to get a custom driver for the 3ware raid card. It might be in the kernel source now.

You're welcome to use my more up to date guide http://syntechnical.com/index.php/topic,12.0.html

Use at your own risk.

i installed kernel on centos but server is using lilo instead of grub
So how can i configure lilo to boot 1 time using new kernel ?

like following in grub ?

grub
savedefault --default=0 --once
quit


Also my current lilo.conf is this , is everything ok ?


message = /boot/message
timeout = 40
prompt
lba32
boot = /dev/hda

default=2.6.9-55.ELsmp
image = /boot/vmlinuz-2.6.21.5
label = 2.6.21.5
initrd = /boot/initrd-2.6.21.5.img
append="root=LABEL=/"

image = /boot/vmlinuz-2.6.9-55.ELsmp
label = 2.6.9-55.ELsmp
initrd = /boot/initrd-2.6.9-55.ELsmp.img
append="root=LABEL=/"

image = /boot/vmlinuz-2.6.9-55.EL
label = 2.6.9-55.EL
initrd = /boot/initrd-2.6.9-55.EL.img
append="root=LABEL=/"

image = /boot/vmlinuz-2.6.9-34.ELsmp
label = Linux
initrd = /boot/initrd-2.6.9-34.ELsmp.img
root = /dev/hda1

image = /boot/vmlinuz-2.6.9-34.0.2.EL
label = 2.6.9-34.0.2.EL
initrd = /boot/initrd-2.6.9-34.0.2.EL.img
root = /dev/hda1

ok i found it :D


#/sbin/lilo -v -v
#/sbin/lilo -R 2.6.21.5