Re: [CentOS] Upgrade to 5.3?

[Sorin Srbu]

>-Original Message-
>From: centos-boun...@centos.org [mailto:centos-boun...@centos.org] On Behalf
Of
>James Bensley
>Sent: Tuesday, January 27, 2009 7:28 AM
>To: CentOS mailing list
>Subject: [CentOS] Upgrade to 5.3?
>
>I'm not that up on CentOS so I'd be curious to know if it is possible
>to upgrade CentOS 5.2 to 5.3 without reinstalling? Perhaps via Yum? Or
>can you get update RPMs?

It's generally considered that upgrading minor versions (eg from 5.2 to 5.3)
is *usually* quite easy and harmless, and normally doesn't pose any problems. 

Upgrading from major versions (eg from 5.9 to 6.0) *is* possible but not
recommended. If you chose this path you should expect major shindings, lib-
and yum dependencies-mayhem and all sorts of weird things happening and/or a
botched operating system which will require a complete reinstall anyway.

It's your choice really, and depends on how brave and much fingerspitzgefühl
you have, alternatively how much time you have on your hands. ;-)

With that said, I once chose to upgrade a Fedora Core 4 to Fedora Core 6. I
went fine finally but took me a week or so to solve all the dependicies
problems for yum. Since then I've never even considered upgrading major
versions.
-- 
/Sorin



smime.p7s
Description: S/MIME cryptographic signature
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos

Re: [CentOS] Bezerk they will go!

[Morten Torstensen]

drew einhorn wrote:
> Note the minimalist .sig
> Probably should fatten it up

You already did by adding a copy of your email as HTML ... ;)

-- 

//Morten Torstensen
//Email: mor...@mortent.org
//IM: morten.torsten...@gmail.com
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos

Re: [CentOS] yum problem

[Mad Unix]

On Tue, Jan 27, 2009 at 9:45 AM, James Bensley  wrote:
> Can you surf the web ok? Is it a DNS resolution problem, 'Temporary
> failure in name resolution' ?
>
> Have you tried another mirror?
>
>
> -BEGIN GEEK CODE BLOCK-
>  Version: 3.1
> GIT/MU/U dpu s: a--> C++>$ U+> L++> B-> P+> E?> W+++>$ N K W++ O M++>$ V-
> PS+++ PE++ Y+ PGP t 5 X+ R- tv+ b+> DI D+++ G+ e(+) h--(++) r++ z++
> --END GEEK CODE BLOCK--
> ___
> CentOS mailing list
> CentOS@centos.org
> http://lists.centos.org/mailman/listinfo/centos
>

no, I think the DNS issue, but i have included it in the resolv.conf

r...@king Tue Jan 27 ~ $ ftp ftp.centos.org
ftp: ftp.centos.org: unknown host
ftp> quit
r...@king Tue Jan 27 ~ $ ftp ftp.freebsd.org
ftp: ftp.freebsd.org: unknown host
ftp>
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos

[CentOS] How to add network driver for network installation

[MOKRANI Rachid]

Hi,

Is someone has an experience for adding a network driver when starting a
network installation.

My network internal card is : Intel 82567M Gigabit  (laptop HP 8530w)

This card is not recognized by my system : CentOS 5.2 x86_64 (kernel
2.6.9-92) 

How to add correctly this driver for a network installation ?

I can add the Intel driver  e1000e-0.5.11.2 after the installation and
the network card works fine, but I need to do network installation, and
this card need to be recognized before starting.

Any help will be very appreciated.

Best Regards.


__

Ce message (et toutes ses pièces jointes éventuelles) est confidentiel et 
établi à l'intention exclusive de ses destinataires. Toute utilisation de ce 
message non conforme à sa destination, toute diffusion ou toute publication, 
totale ou partielle, est interdite, sauf autorisation expresse. L'IFP décline 
toute responsabilité au titre de ce message.

This message and any attachments (the message) are confidential and intended 
solely for the addressees. Any unauthorised use or dissemination is prohibited. 
IFP should not be liable for this message.

Visitez notre site Web / Visit our web site : http://www.ifp.fr
__
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos

Re: [CentOS] How to add network driver for network installation

[Pintér Tibor]

> CentOS 5.2 x86_64 (kernel
> 2.6.9-92) 

that cant be true

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

Re: [CentOS] How to add network driver for network installation

[MOKRANI Rachid]

Sorry : It's 2.6.18-92 

 

-Message d'origine-
De : centos-boun...@centos.org [mailto:centos-boun...@centos.org] De la part de 
Pintér Tibor
Envoyé : mardi 27 janvier 2009 09:13
À : CentOS mailing list
Objet : Re: [CentOS] How to add network driver for network installation

> CentOS 5.2 x86_64 (kernel
> 2.6.9-92)

that cant be true

t
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos
__

Ce message (et toutes ses pièces jointes éventuelles) est confidentiel et 
établi à l'intention exclusive de ses destinataires. Toute utilisation de ce 
message non conforme à sa destination, toute diffusion ou toute publication, 
totale ou partielle, est interdite, sauf autorisation expresse. L'IFP décline 
toute responsabilité au titre de ce message.

This message and any attachments (the message) are confidential and intended 
solely for the addressees. Any unauthorised use or dissemination is prohibited. 
IFP should not be liable for this message.

Visitez notre site Web / Visit our web site : http://www.ifp.fr
__

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

Re: [CentOS] How to add network driver for network installation

[John R Pierce]

Pintér Tibor wrote:
>> CentOS 5.2 x86_64 (kernel
>> 2.6.9-92) 
>> 
>
> that cant be true
>   


indeed, CentOS 5.anything should be 2.6.18-xx, like 2.6.18-92.1.18.el5 
(on one of my boxes)

I believe 2.6.9 was CentOS 4.



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

Re: [CentOS] yum problem

[James Bensley]

It doesn't matter if its in resolv.conf you don't seem to have a
working DNS provider.

I would start there, use nslookup, can you solve anything? Check with
your resolv.conf, what name server are you using, can you ping it? Try
another name server?

2009/1/27 Mad Unix :
> On Tue, Jan 27, 2009 at 9:45 AM, James Bensley  wrote:
>> Can you surf the web ok? Is it a DNS resolution problem, 'Temporary
>> failure in name resolution' ?
>>
>> Have you tried another mirror?
>>
>>
>> -BEGIN GEEK CODE BLOCK-
>>  Version: 3.1
>> GIT/MU/U dpu s: a--> C++>$ U+> L++> B-> P+> E?> W+++>$ N K W++ O M++>$ V-
>> PS+++ PE++ Y+ PGP t 5 X+ R- tv+ b+> DI D+++ G+ e(+) h--(++) r++ z++
>> --END GEEK CODE BLOCK--
>> ___
>> CentOS mailing list
>> CentOS@centos.org
>> http://lists.centos.org/mailman/listinfo/centos
>>
>
> no, I think the DNS issue, but i have included it in the resolv.conf
>
> r...@king Tue Jan 27 ~ $ ftp ftp.centos.org
> ftp: ftp.centos.org: unknown host
> ftp> quit
> r...@king Tue Jan 27 ~ $ ftp ftp.freebsd.org
> ftp: ftp.freebsd.org: unknown host
> ftp>
> ___
> CentOS mailing list
> CentOS@centos.org
> http://lists.centos.org/mailman/listinfo/centos
>



-- 
-BEGIN GEEK CODE BLOCK-
  Version: 3.1
GIT/MU/U dpu s: a--> C++>$ U+> L++> B-> P+> E?> W+++>$ N K W++ O M++>$ V-
PS+++ PE++ Y+ PGP t 5 X+ R- tv+ b+> DI D+++ G+ e(+) h--(++) r++ z++
--END GEEK CODE BLOCK--
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos

Re: [CentOS] yum problem

[Anthony Kamau]

> -Original Message-
> From: centos-boun...@centos.org [mailto:centos-boun...@centos.org] On
> Behalf Of James Bensley
> Sent: Tuesday, 27 January 2009 7:49 PM
> To: CentOS mailing list
> Subject: Re: [CentOS] yum problem
>
> It doesn't matter if its in resolv.conf you don't seem to have a
> working DNS provider.
>
> I would start there, use nslookup, can you solve anything? Check with
> your resolv.conf, what name server are you using, can you ping it? Try
> another name server?
>

If all else fails, then you can use opendns.com - they provide their
services for free!

Cheers,
AK.



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

Re: [CentOS] Upgrade to 5.3?

[William L. Maltby]

On Tue, 2009-01-27 at 00:33 -0600, Barry Brimer wrote:
> > I'm not that up on CentOS so I'd be curious to know if it is possible
> > to upgrade CentOS 5.2 to 5.3 without reinstalling? Perhaps via Yum? Or
> > can you get update RPMs?
> 
> Once CentOS 5.3 is released, you can just type "yum upgrade" and you will 
> be upgraded from CentOS 5.2 to CentOS 5.3.

If your configuration is set normally, "yum update" will do it. Just
like a normal set of updates. Having said that, ...

IIRC there was a recent thread, which I can't find at the moment, that
discussed some updates to glib (I though it was glibc but Ralph(?)
corrected me in another thread) that might cause a problem in later
stages of the update process. I believe the tentative conclusion was
that some notification about it would be needed.

I'm guessing it might be along the lines of the sqlite update needed in
the dark past. If so, that would be a two-step: update the lib and then
do the normal update. But keep in mind that's all predicated on vague
memory of the thread, incomplete information and blissful ignorance.

Regardless, I'm remaining alert in case my memory and understanding are
correct.

> 

-- 
Bill

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

Re: [CentOS] Old Small Box

[William L. Maltby]

On Tue, 2009-01-27 at 08:28 +0100, Morten Torstensen wrote:
> Ross Walker wrote:
> > Well the only reason /boot isn't possible in LVM is because grub can't
> > of yet handle reading LVM volumes. As soon as it can though, there will
> > be no need for a separate /boot.
> 
> Then we just need BIOS support to boot from LVM, and we can create the 
> PV on /dev/sda and never care about the old partitions anymore. 
> Everything should be in LVM for ease of management.

Oh boy! You've heard rumors that the BIOS manufacturers are going to
begin supporting file-system-specific layouts? I find that hard to
swallow.

> 

-- 
Bill

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

Re: [CentOS] Old Small Box

[Vnpenguin]

On Thu, Jan 22, 2009 at 12:06 AM, Mike -- EMAIL IGNORED
 wrote:
> I have an old 400mHz Dell with a 20G hard drive
> and 125M ram.  Can I install and run CentOS on it?

I have a box PII 333MHz + 128 MB ram running (sshd, samba, httpd,...)
CentOS 4.x at home.

-- 
http://vnoss.org
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos

[CentOS] postfix startup issue - CentOS 5.2

[Tom Brown]

Hi

I am trying to get postfix going on a friends VPS however after 
installing on startup i see this

Jan 27 11:15:13 postfix/postfix-script: starting the Postfix mail system
Jan 27 11:15:13 postfix/master[28090]: fatal: pipe: Cannot allocate memory

and a top gives

Tasks:  23 total,   1 running,  22 sleeping,   0 stopped,   0 zombie
Cpu(s):  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  
0.0%st
Mem:524288k total,   160920k used,   363368k free,0k buffers
Swap:0k total,0k used,0k free,0k cached

any clue?

thanks

# uname -i
Linux 2.6.18-028stab059.6 #1 SMP Fri Nov 14 14:01:22 MSK 2008 i686 i686 
i386 GNU/Linux

# cat /etc/redhat-release
CentOS release 5.2 (Final)
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos

Re: [CentOS] postfix startup issue - CentOS 5.2

[Per Qvindesland]

Hi 

Have you checked /etc/hosts to see if it is correct?

Per


On 1/27/09 12:28 PM, "Tom Brown"  wrote:

> Jan 27 11:15:13 postfix/master[28090]: fatal: pipe: Cannot allocate memory


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

Re: [CentOS] postfix startup issue - CentOS 5.2

[Tom Brown]

>
> Have you checked /etc/hosts to see if it is correct?
>
>   

it does seem correct yes
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos

Re: [CentOS] Old Small Box

[Morten Torstensen]

William L. Maltby wrote:
> Oh boy! You've heard rumors that the BIOS manufacturers are going to
> begin supporting file-system-specific layouts? I find that hard to
> swallow.

Well, in theory we don't need BIOS support. The BIOS will check that the 
first sector of the device is signed with 55AAh and just jump into the 
code in the first sector. So if LVM leaves the first sector and track 
(so grub can hide there) alone, it should work just fine.

-- 

//Morten Torstensen
//Email: mor...@mortent.org
//IM: morten.torsten...@gmail.com

I can't listen to that much Wagner. I start getting the urge to conquer 
Poland.
-- Woody Allen
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos

Re: [CentOS] postfix startup issue - CentOS 5.2

[Ralph Angenendt]

Tom Brown wrote:
> Hi
> 
> I am trying to get postfix going on a friends VPS however after 
> installing on startup i see this
> 
> Jan 27 11:15:13 postfix/postfix-script: starting the Postfix mail system
> Jan 27 11:15:13 postfix/master[28090]: fatal: pipe: Cannot allocate memory

You are out of open files on the system. Or out of available sockets.

> # uname -i
> Linux 2.6.18-028stab059.6 #1 SMP Fri Nov 14 14:01:22 MSK 2008 i686 i686 
> i386 GNU/Linux

That sounds like a good idea. stab your vps provider, because your
system seems to be very limited regarding the number of files and/or
sockets you can open.

Ralph


pgpKhGCWNHfL2.pgp
Description: PGP signature
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos

Re: [CentOS] Old Small Box

[William L. Maltby]

On Tue, 2009-01-27 at 12:43 +0100, Morten Torstensen wrote:
> William L. Maltby wrote:
> > Oh boy! You've heard rumors that the BIOS manufacturers are going to
> > begin supporting file-system-specific layouts? I find that hard to
> > swallow.
> 
> Well, in theory we don't need BIOS support. The BIOS will check that the 
> first sector of the device is signed with 55AAh and just jump into the 
> code in the first sector. So if LVM leaves the first sector and track 
> (so grub can hide there) alone, it should work just fine.

As I thought (remembered). Which means no additional BIOS support is
needed. What's needed is that the first sector understands LVM (or
whatever the file system is if there's not LVM). So if we put the root
onto a "conventional" file system, the first sector would be the boot
code - no GRUB/LILO or anything else needed (IIRC). Of course, then no
parameters passed, no selection of alternate boots, etc.

Just like the old days! :-)

Simplicity has its strengths and weaknesses.

> 

-- 
Bill

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

Re: [CentOS] postfix startup issue - CentOS 5.2

[Ned Slider]

Tom Brown wrote:
> Hi
> 
> I am trying to get postfix going on a friends VPS however after 
> installing on startup i see this
> 



> 
> # uname -i
> Linux 2.6.18-028stab059.6 #1 SMP Fri Nov 14 14:01:22 MSK 2008 i686 i686 
> i386 GNU/Linux
> 

That's not a CentOS kernel.

> # cat /etc/redhat-release
> CentOS release 5.2 (Final)

Please see this page:

http://wiki.centos.org/TipsAndTricks/BrokenVserver

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

[CentOS] More than 2TB RAID...

[John Doe]

Hi,

I just received a new server (HP DL180G5) with 12x 1TB HDs and I bumped into 
fdisks 2TB limits...
Since this is an entry level server, I can't use the classic HP bootable 
utilities to create smaller volumes et can only create a big RAID6.
I found out that: using parted, labelling it gpt and creating the partitions 
would do the trick.
But, what about grub?  I read that it does not support gpt...
Or is there another way to do it?
By example, do you think I could boot on a live CentOS, install HP RAID tools 
on it (is it possible?) and then create volumes...?

Thx,
JD


  

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

Re: [CentOS] Upgrade to 5.3?

[Ned Slider]

Barry Brimer wrote:
>> I'm not that up on CentOS so I'd be curious to know if it is possible
>> to upgrade CentOS 5.2 to 5.3 without reinstalling? Perhaps via Yum? Or
>> can you get update RPMs?
> 
> Once CentOS 5.3 is released, you can just type "yum upgrade" and you will 
> be upgraded from CentOS 5.2 to CentOS 5.3.

Please READ the release notes first :)

yum update glibc
yum update

And if you're updating *remotely* over an ssh login, consider using a 
screen session for the update in case the ssh link fails during the 
update process.

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

Re: [CentOS] More than 2TB RAID...

[James Bensley]

I have this problem, I just boot from usb sticks?

A 512MB usb pen drive is (just about extinct!) only about £2.00 so get
three, make one an ms-dos disk with parted and format it ready. Then
pop-in CentOS and install it mounting the usb pen as /boot so grub
boots from that msdos partition and then use dd to copy the usb
pendrive to two more to have them knocking around (take one home and
put it in your sock draw or something) and just keep the image for the
pen drive somewhere like with your work files?

2009/1/27 John Doe :
> Hi,
>
> I just received a new server (HP DL180G5) with 12x 1TB HDs and I bumped into 
> fdisks 2TB limits...
> Since this is an entry level server, I can't use the classic HP bootable 
> utilities to create smaller volumes et can only create a big RAID6.
> I found out that: using parted, labelling it gpt and creating the partitions 
> would do the trick.
> But, what about grub?  I read that it does not support gpt...
> Or is there another way to do it?
> By example, do you think I could boot on a live CentOS, install HP RAID tools 
> on it (is it possible?) and then create volumes...?
>
> Thx,
> JD
>
>
>
>
> ___
> CentOS mailing list
> CentOS@centos.org
> http://lists.centos.org/mailman/listinfo/centos
>



-- 
-BEGIN GEEK CODE BLOCK-
  Version: 3.1
GIT/MU/U dpu s: a--> C++>$ U+> L++> B-> P+> E?> W+++>$ N K W++ O M++>$ V-
PS+++ PE++ Y+ PGP t 5 X+ R- tv+ b+> DI D+++ G+ e(+) h--(++) r++ z++
--END GEEK CODE BLOCK--
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos

Re: [CentOS] postfix startup issue - CentOS 5.2

[Kai Schaetzl]

Tom Brown wrote on Tue, 27 Jan 2009 11:28:06 +:

> Linux 2.6.18-028stab059.6

This is not CentOS 5.2, this is an OpenVZ kernel. You should contact the 
vendor with this problem, I'm sure it's related to it being a "VPS".

Kai

-- 
Kai Schätzl, Berlin, Germany
Get your web at Conactive Internet Services: http://www.conactive.com



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

Re: [CentOS] More than 2TB RAID...

[Morten Torstensen]

John Doe wrote:
> I just received a new server (HP DL180G5) with 12x 1TB HDs and I bumped into 
> fdisks 2TB limits...

I would create two raid logical volumes, one for centos (say, 20GB to 
100GB) and one with the rest of the space. Install centos and normal MBR 
on /dev/sda and then use lvm on the /dev/sdb directly with no partition 
table needed.

I would also strongly consider having two disks mirrored for the system 
in one lvm vg and the rest in another, but with 1TB disks it is kind of 
wasted space. Tho with 12 disks you can have 2 disks for RAID1, then 
8+P+1 in RAID5.

-- 

//Morten Torstensen
//Email: mor...@mortent.org
//IM: morten.torsten...@gmail.com

I can't listen to that much Wagner. I start getting the urge to conquer 
Poland.
-- Woody Allen
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos

Re: [CentOS] How to add network driver for network installation

[Fabian Arrotin]

On Tue, 27 Jan 2009 09:04:46 +0100, "MOKRANI Rachid"
 wrote:
> 
> Hi,
> 
> Is someone has an experience for adding a network driver when starting a
> network installation.
> 
> My network internal card is : Intel 82567M Gigabit  (laptop HP 8530w)
> 
> This card is not recognized by my system : CentOS 5.2 x86_64 (kernel
> 2.6.9-92)
> 
> How to add correctly this driver for a network installation ?
> 
> I can add the Intel driver  e1000e-0.5.11.2 after the installation and
> the network card works fine, but I need to do network installation, and
> this card need to be recognized before starting.
> 
> Any help will be very appreciated.
> 
> Best Regards.
> 

If you want to do a network install you have to modify your initrd.img to
include the specific network kmod.
That's what i did for the eeepc/CentOS 5.2 network setup
You can also create a repo then that can be used during setup so that the
network kmod will be installed on the filesystem too and that you have
network access after the first reboot without having to manually install
the driver/module

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

Re: [CentOS] More than 2TB RAID...

[John Doe]

From: James Bensley 
> I have this problem, I just boot from usb sticks?
> A 512MB usb pen drive is (just about extinct!) only about £2.00 so get
> three, make one an ms-dos disk with parted and format it ready. Then
> pop-in CentOS and install it mounting the usb pen as /boot so grub
> boots from that msdos partition and then use dd to copy the usb
> pendrive to two more to have them knocking around (take one home and
> put it in your sock draw or something) and just keep the image for the
> pen drive somewhere like with your work files?

I guess that would work but I don't think my boss will like having a usb key 
plugged in each servers...

From: Morten Torstensen 
> I would create two raid logical volumes, one for centos (say, 20GB to 
> 100GB) and one with the rest of the space. Install centos and normal MBR 
> on /dev/sda and then use lvm on the /dev/sdb directly with no partition 
> table needed.
> I would also strongly consider having two disks mirrored for the system 
> in one lvm vg and the rest in another, but with 1TB disks it is kind of 
> wasted space. Tho with 12 disks you can have 2 disks for RAID1, then 
> 8+P+1 in RAID5.

Yes but, as I said, I sadly don't have access (yet)to vendor utilities that 
would allow me to create logical volumes...
For this entry model of server, there is no offline/bootable utility to access 
extended RAID configuration (even if HP shipped the utility CD for this 
unsupported server ^_^) .
The BIOS raid setup is very basic, select drives, RAID level, and click...
I could do 2 HDs in RAID1, then 10 disks in RAID6, but that would indeed 
"waste" 1TB.

The best solution would be to be able somehow to run the avanced RAID utility...
For that, I need a running OS where I could install it.

Thx,
JD


  

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

[CentOS] OT : iptables/arptables question

[Fabian Arrotin]

I have a CentOS box that acts as a packet filter/firewall with iptables but
the box itself isn't able to reach internet : here why : 

Internet - public IP|ISP router|private IP  - private IP + public
IP/32 + public IP subnet/29|my CentOS fw|private network/dmz

As you can see my provider gave us a /29 public ip subnet but behind a
private IP subnet (192.168.X.X/24 - used for the routing between the ISP
router and the fw)
I've configured my iptables/routing correctly and machines from the DMZ
have no problems reaching the external world (use of SNAT in the nat table
of course).
The problem is that the firewall itself can't access the public network
because of his private ip 192.168.X.X used for the routing between ISP
router and itself.
I also received a /32 public ip for the fw itself and i've added to the
ethx:1 alias . Problem is that kernel always decide that (because of
default gw being on the private ip 192.168.X.X) he has to use the
192.168.X.X ip address as outbond interface. So every packet leaving (so
i'm talking about OUTPUT table and not about FORWARD nor nat table) the fw
comes from a 192.168.X.X ip and so never comes back (which is normal).
Question is : how can i "mangle" output packets to appear coming from
public ip and not from 192.168.X.X ? 
For example , at the application layer, i can produce icmp packets with
`ping -I my.public.ip/32 remote.host.on.internet` that come back but of
course nothing with a traditionnal `ping remote.host.on.internet`
I've had a look at arptables and tested ` arptables -A OUT -s 192.168.X.X !
-d  192.168.X.0/24 -o eth3 -j mangle --mangle-ip-s my.public.ip` but that
doesn't seem to do the trick ..

Any ideas ? 
I just hope that it was clear enough :-p
 
--
 Fabian Arrotin
 idea=`grep -i clue /dev/brain` ; test -z "$idea" && echo "sorry, init 6 in
progress" || sh ./answer.sh 

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

[CentOS] DNS

[adrian kok]

Hi all

I am setting dns and trying to ping my domain
the 1st reply is ok and solve the ip address

but 2nd line to end, it reponses to my upstream info
instead of www.mydomain.com

What wrong of my dns?

Thank you

linux ~ # ping www.mydomain.com
PING www.mydomain.com (171.203.20.173) 56(84) bytes of
data.
64 bytes from 171203020173.upstream.com
(171.203.20.173): icmp_seq=1 ttl=43 time=291 ms
64 bytes from 171203020173.upstream.com
(171.203.20.173): icmp_seq=2 ttl=43 time=290 ms


Send instant messages to your online friends http://uk.messenger.yahoo.com 
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos

Re: [CentOS] More than 2TB RAID...

[Morten Torstensen]

John Doe wrote:
> The best solution would be to be able somehow to run the avanced RAID 
> utility...
> For that, I need a running OS where I could install it.

They don't have a bootable CD image with the tools? I am only familiar 
with the IBM ServeRAID adapters where you have such a tool.

Another option is of course to create a number of 2TB partitions and 
create a VG with all of them as PVs. Of course, you would need 6 
partitions for this, but once you have them addded to a VG it is pretty 
seamless.

-- 

//Morten Torstensen
//Email: mor...@mortent.org
//IM: morten.torsten...@gmail.com

I can't listen to that much Wagner. I start getting the urge to conquer 
Poland.
-- Woody Allen
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos

Re: [CentOS] DNS

[Ralph Angenendt]

adrian kok wrote:
> linux ~ # ping www.mydomain.com
> PING www.mydomain.com (171.203.20.173) 56(84) bytes of
> data.
> 64 bytes from 171203020173.upstream.com
> (171.203.20.173): icmp_seq=1 ttl=43 time=291 ms
> 64 bytes from 171203020173.upstream.com
> (171.203.20.173): icmp_seq=2 ttl=43 time=290 ms

Something is severly broken on your side:

[angen...@shutdown ~]$host www.mydomain.com
www.mydomain.com is an alias for www.mydomain.com.akadns.net.
www.mydomain.com.akadns.net has address 66.150.120.131
www.mydomain.com.akadns.net has address 216.34.94.184
[angen...@shutdown ~]$

Ralph


pgpXiB8G2JGmq.pgp
Description: PGP signature
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos

Re: [CentOS] DNS

[Fajar Priyanto]

On Tue, Jan 27, 2009 at 10:51 PM, adrian kok  wrote:
> I am setting dns and trying to ping my domain
> the 1st reply is ok and solve the ip address
>
> but 2nd line to end, it reponses to my upstream info
> instead of www.mydomain.com
>
> What wrong of my dns?

It's rather difficult to know without any related configuration here.
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos

Re: [CentOS] DNS

[adrian kok]

Sorry. I use fake name

host command can return right ip address

My question is

I am curious why the 2nd line is 

171203020173.upstream.com instead of www.mydomain.com
?

It should be. right?

64 bytes from www.mydomain.com (171.203.20.173):
icmp_seq=1 ttl=43 time=291 ms

Thank you

--- Ralph Angenendt  wrote:

> adrian kok wrote:
> > linux ~ # ping www.mydomain.com
> > PING www.mydomain.com (171.203.20.173) 56(84)
> bytes of
> > data.
> > 64 bytes from 171203020173.upstream.com
> > (171.203.20.173): icmp_seq=1 ttl=43 time=291 ms
> > 64 bytes from 171203020173.upstream.com
> > (171.203.20.173): icmp_seq=2 ttl=43 time=290 ms
> 
> Something is severly broken on your side:
> 
> [angen...@shutdown ~]$host www.mydomain.com
> www.mydomain.com is an alias for
> www.mydomain.com.akadns.net.
> www.mydomain.com.akadns.net has address
> 66.150.120.131
> www.mydomain.com.akadns.net has address
> 216.34.94.184
> [angen...@shutdown ~]$
> 
> Ralph
> > ___
> CentOS mailing list
> CentOS@centos.org
> http://lists.centos.org/mailman/listinfo/centos
> 


Send instant messages to your online friends http://uk.messenger.yahoo.com 
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos

Re: [CentOS] DNS

[Joost Waversveld]

Isn't this just the Reverse DNS name of your ip (171.203.20.173)?

Do a check through "dig -x 171.203.20.173". Most likely you will get the 
17120320173.upstream.com back. If so, it's the Reverse DNS name...

-- 
Joost Waversveld

adrian kok wrote:
> Sorry. I use fake name
> 
> host command can return right ip address
> 
> My question is
> 
> I am curious why the 2nd line is 
> 
> 171203020173.upstream.com instead of www.mydomain.com
> ?
> 
> It should be. right?
> 
> 64 bytes from www.mydomain.com (171.203.20.173):
> icmp_seq=1 ttl=43 time=291 ms
> 
> Thank you
> 
> --- Ralph Angenendt  wrote:
> 
>> adrian kok wrote:
>>> linux ~ # ping www.mydomain.com
>>> PING www.mydomain.com (171.203.20.173) 56(84)
>> bytes of
>>> data.
>>> 64 bytes from 171203020173.upstream.com
>>> (171.203.20.173): icmp_seq=1 ttl=43 time=291 ms
>>> 64 bytes from 171203020173.upstream.com
>>> (171.203.20.173): icmp_seq=2 ttl=43 time=290 ms
>> Something is severly broken on your side:
>>
>> [angen...@shutdown ~]$host www.mydomain.com
>> www.mydomain.com is an alias for
>> www.mydomain.com.akadns.net.
>> www.mydomain.com.akadns.net has address
>> 66.150.120.131
>> www.mydomain.com.akadns.net has address
>> 216.34.94.184
>> [angen...@shutdown ~]$
>>
>> Ralph
>>> ___
>> CentOS mailing list
>> CentOS@centos.org
>> http://lists.centos.org/mailman/listinfo/centos
>>
> 
> 
> Send instant messages to your online friends http://uk.messenger.yahoo.com 
> ___
> CentOS mailing list
> CentOS@centos.org
> http://lists.centos.org/mailman/listinfo/centos


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

Re: [CentOS] How to add network driver for network installation

[MOKRANI Rachid]

But it's not working :-(

1 - After installing CentOS 5.2 from scratch, I need to add the following rpm 
for recognized my network card :
http://centos.toracat.org/ajb/CentOS-5/e1000e/kmod-e1000e-0.5.11.2_NAPI-1.el5.x86_64.rpm
 

OK it's working after install the following files. I have the network. Good !
rpm -qlp kmod-e1000e-0.5.11.2_NAPI-1.el5.x86_64.rpm
/lib/modules/2.6.18-92.el5
/lib/modules/2.6.18-92.el5/extra
/lib/modules/2.6.18-92.el5/extra/e1000e
/lib/modules/2.6.18-92.el5/extra/e1000e/e1000e.ko



2 - for my network installation I do the following for trying to add the new 
e1000e.ko module.
Something is mistake :-( but I don't know what . a dependency ? please HELP !



mkdir /my_isolinux
% cp -r /my_centos_DVD/isolinux /my_isolinux
% mkdir /tmp/my_initrd
% cd /my_isolinux/isolinux
% ls -1
boot.cat
boot.msg
general.msg
initrd.img
isolinux.bin
isolinux.cfg
memtest
options.msg
param.msg
rescue.msg
splash.lss
TRANS.TBL
vmlinuz

%mkdir INITRD
%mv initrd.img INITRD
%cd INITRD
%ls -1
initrd.img
%gunzip < initrd.img  | cpio -i --make-directories
%ls -1
bin
dev
etc
init
initrd.img
modules
proc
sbin
selinux
sys
tmp
var
%rm initrd.img
%cd modules
%ls -1
module-info
modules.alias
modules.cgz
modules.dep
pci.ids
%mkdir MODULES
%cp modules.cgz MODULES
%cd MODULES
%gunzip < modules.cgz   | cpio -i --make-directories
%ls -1
2.6.18-92.el5
modules.cgz
%rm modules.cgz
%ls -al 2.6.18-92.el5/x86_64/e100*
-rwxr--r-- 1 root root 167808 jan 27 16:42 2.6.18-92.el5/x86_64/e1000e.ko
-rwxr--r-- 1 root root 204616 jan 27 16:42 2.6.18-92.el5/x86_64/e1000.ko
-rwxr--r-- 1 root root  84872 jan 27 16:42 2.6.18-92.el5/x86_64/e100.ko
%rm 2.6.18-92.el5/x86_64/e1000e.ko

"I Change my module "

%cp /lib/modules/2.6.18-92.el5/extra/e1000e/e1000e.ko 
2.6.18-92.el5/x86_64/e1000e.ko
find 2.6.18-92.el5 -print | cpio -o -Hnewc | gzip -c9 > modules.cgz
%mv modules.cgz  ../
%cd ../
%rm -rf MODULES
%cd ../
%pwd
/my_isolinux/isolinux/INITRD
%find . -print | cpio -o -Hnewc | gzip -c9 > ../initrd.img_NEW
%cd ..
%rm -rf INITRD
%mv initrd.img_NEW initrd.img
%ls -1
boot.cat
boot.msg
general.msg
initrd.img
isolinux.bin
isolinux.cfg
memtest
options.msg
param.msg
rescue.msg
splash.lss
TRANS.TBL
vmlinuz


Burn an ISO and  boot on the new CD. The intel network card is always not 
recognized :-(



-Message d'origine-
De : centos-boun...@centos.org [mailto:centos-boun...@centos.org] De la part de 
Fabian Arrotin
Envoyé : mardi 27 janvier 2009 14:00
À : CentOS mailing list
Objet : Re: [CentOS] How to add network driver for network installation

If you want to do a network install you have to modify your initrd.img to 
include the specific network kmod.
That's what i did for the eeepc/CentOS 5.2 network setup You can also create a 
repo then that can be used during setup so that the network kmod will be 
installed on the filesystem too and that you have network access after the 
first reboot without having to manually install the driver/module

__

Ce message (et toutes ses pièces jointes éventuelles) est confidentiel et 
établi à l'intention exclusive de ses destinataires. Toute utilisation de ce 
message non conforme à sa destination, toute diffusion ou toute publication, 
totale ou partielle, est interdite, sauf autorisation expresse. L'IFP décline 
toute responsabilité au titre de ce message.

This message and any attachments (the message) are confidential and intended 
solely for the addressees. Any unauthorised use or dissemination is prohibited. 
IFP should not be liable for this message.

Visitez notre site Web / Visit our web site : http://www.ifp.fr
__

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

Re: [CentOS] More than 2TB RAID...

[Peter Kjellstrom]

On Tuesday 27 January 2009, John Doe wrote:
> Hi,
>
> I just received a new server (HP DL180G5) with 12x 1TB HDs and I bumped
> into fdisks 2TB limits... Since this is an entry level server, I can't use
> the classic HP bootable utilities to create smaller volumes et can only
> create a big RAID6.

If it does raid6 then you have a p400 or p800 controller, right? If so then 
hpacucli or similar can easily give you a small logical drive for the OS and 
then a large one for data.

On the small one partition normally and install grub.

On the big one do one of:

1) partition with parted and gpt
2) use lvm and put a pv directly on the cciss device
3) put the filesystem directly on the cciss device

Also since you have that hardware. Make sure your controller firmware is at 
5.22, drives at HPG6 and backplane at 2.00

Cheers,
 Peter

> I found out that: using parted, labelling it gpt and 
> creating the partitions would do the trick. But, what about grub?  I read
> that it does not support gpt...
> Or is there another way to do it?
> By example, do you think I could boot on a live CentOS, install HP RAID
> tools on it (is it possible?) and then create volumes...?
>
> Thx,
> JD


signature.asc
Description: This is a digitally signed message part.
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos

Re: [CentOS] More than 2TB RAID...

[Peter Kjellstrom]

On Tuesday 27 January 2009, Morten Torstensen wrote:
> John Doe wrote:
> > I just received a new server (HP DL180G5) with 12x 1TB HDs and I bumped
> > into fdisks 2TB limits...
>
> I would create two raid logical volumes, one for centos (say, 20GB to
> 100GB) and one with the rest of the space. Install centos and normal MBR
> on /dev/sda and then use lvm on the /dev/sdb directly with no partition
> table needed.

Smartarray controllers user cciss.ko which is one of the few block I/O thingys 
that doesn't use the linux scsi layer. Devices will be /dev/cciss/c0d0, 
c0d1, ...

> I would also strongly consider having two disks mirrored for the system
> in one lvm vg and the rest in another, but with 1TB disks it is kind of
> wasted space. Tho with 12 disks you can have 2 disks for RAID1, then
> 8+P+1 in RAID5.

Smartarray controllers don't work like "normal" raid. You can have the first 
part of all your 12 drives be a logical raid1+0 and then the rest be a raid6. 
In this way you won't "waste" two drive-slots for OS.

/Peter


signature.asc
Description: This is a digitally signed message part.
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos

Re: [CentOS] How to add network driver for network installation

[MOKRANI Rachid]

Sorry I make a mistake in my last email : MORE EASY FOR READING  



But it's not working :-(

1 - After installing CentOS 5.2 from scratch, I need to add the
following rpm for recognized my network card :
http://centos.toracat.org/ajb/CentOS-5/e1000e/kmod-e1000e-0.5.11.2_NAPI-
1.el5.x86_64.rpm 

OK it's working after install the following files. I have the network.
Good !
rpm -qlp kmod-e1000e-0.5.11.2_NAPI-1.el5.x86_64.rpm
/lib/modules/2.6.18-92.el5
/lib/modules/2.6.18-92.el5/extra
/lib/modules/2.6.18-92.el5/extra/e1000e
/lib/modules/2.6.18-92.el5/extra/e1000e/e1000e.ko



2 - for my network installation I do the following for trying to add the
new e1000e.ko module.
Something is mistake :-( but I don't know what . a dependency ? please
HELP !


% mkdir /my_isolinux
% cp -r /my_centos_DVD/isolinux /my_isolinux 
% cd /my_isolinux/isolinux 
% ls -1 
boot.cat boot.msg general.msg initrd.img isolinux.bin isolinux.cfg
memtest options.msg param.msg rescue.msg splash.lss TRANS.TBL vmlinuz


% mkdir INITRD
% mv initrd.img INITRD
% cd INITRD
% ls -1
initrd.img
% gunzip < initrd.img  | cpio -i --make-directories 
% ls -1 
bin dev etc init initrd.img modules proc sbin selinux sys tmp var 
% rm initrd.img 
% cd modules 
% ls -1 
module-info modules.alias modules.cgz modules.dep pci.ids 
% mkdir MODULES 
% cp modules.cgz MODULES 
% cd MODULES
% gunzip < modules.cgz | cpio -i --make-directories
% ls -1
2.6.18-92.el5
modules.cgz
% rm modules.cgz
% ls -al 2.6.18-92.el5/x86_64/e100*
-rwxr--r-- 1 root root 167808 jan 27 16:42
2.6.18-92.el5/x86_64/e1000e.ko
-rwxr--r-- 1 root root 204616 jan 27 16:42 2.6.18-92.el5/x86_64/e1000.ko
-rwxr--r-- 1 root root  84872 jan 27 16:42 2.6.18-92.el5/x86_64/e100.ko
%rm 2.6.18-92.el5/x86_64/e1000e.ko

"I Change my module "

% cp /lib/modules/2.6.18-92.el5/extra/e1000e/e1000e.ko
2.6.18-92.el5/x86_64/e1000e.ko 
% find 2.6.18-92.el5 -print | cpio -o -Hnewc | gzip -c9 > modules.cgz 
% mv modules.cgz  ../ 
% cd ../ 
% rm -rf MODULES 
% cd ../ 
% pwd 
/my_isolinux/isolinux/INITRD 
% find . -print | cpio -o -Hnewc | gzip -c9 > ../initrd.img_NEW 
% cd ..
% rm -rf INITRD
% mv initrd.img_NEW initrd.img
% ls -1
boot.cat
boot.msg
general.msg
initrd.img
isolinux.bin
isolinux.cfg
memtest
options.msg
param.msg
rescue.msg
splash.lss
TRANS.TBL
vmlinuz


Burn an ISO and  boot on the new CD. The intel network card is always
not recognized :-(



__

Ce message (et toutes ses pièces jointes éventuelles) est confidentiel et 
établi à l'intention exclusive de ses destinataires. Toute utilisation de ce 
message non conforme à sa destination, toute diffusion ou toute publication, 
totale ou partielle, est interdite, sauf autorisation expresse. L'IFP décline 
toute responsabilité au titre de ce message.

This message and any attachments (the message) are confidential and intended 
solely for the addressees. Any unauthorised use or dissemination is prohibited. 
IFP should not be liable for this message.

Visitez notre site Web / Visit our web site : http://www.ifp.fr
__
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos

Re: [CentOS] OT : iptables/arptables question

[Filipe Brandenburger]

Hi Fabian:

On Tue, Jan 27, 2009 at 08:16, Fabian Arrotin  wrote:
> Question is : how can i "mangle" output packets to appear coming from
> public ip and not from 192.168.X.X ?

Found this that might help you (google for: linux default outgoing ip):

"""
> On a machine with multiple interfaces, is it possible to set the default
> outgoing IP address to something other than the address for the interface
> on the outgoing route?

Yes.

ip route add 10.1.1.0/24 via 192.168.1.1 src 172.16.1.1
  ^^^
The src parameter tells the routing code to use this address when sending
packets. The address only needs to be on the system. IE:

ip addr add 172.16.1.1/32 dev dummy0

And send the packets out of eth0.
"""
From: http://lkml.indiana.edu/hypermail/linux/kernel/0112.1/0359.html

Just make sure you keep a separate route for your ISP's side of the
private network (maybe the one created when your interface goes up
will do), otherwise your routing protocol might fail.

HTH,
Filipe
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos

Re: [CentOS] DNS

[Ralph Angenendt]

adrian kok wrote:
> Sorry. I use fake name

No, you didn't. www.mydomain.com does exist, so it cannot be fake.

> host command can return right ip address

I don't really like people who misuse existing domains for questions
like this, as it really makes it harder to debug things.

> My question is
> 
> I am curious why the 2nd line is 
> 
> 171203020173.upstream.com instead of www.mydomain.com

Because you don't have a correct reverse DNS setting for your domain.
But as you don't say what it is, I won't look further.

Ralph


pgp5XHlo7p5tA.pgp
Description: PGP signature
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos

Re: [CentOS] yum problem

[Lanny Marcus]

On Tue, Jan 27, 2009 at 4:04 AM, Anthony Kamau  wrote:
>> -Original Message-
>> From: centos-boun...@centos.org [mailto:centos-boun...@centos.org] On
>> Behalf Of James Bensley
>> Sent: Tuesday, 27 January 2009 7:49 PM
>> To: CentOS mailing list
>> Subject: Re: [CentOS] yum problem
>>
>> It doesn't matter if its in resolv.conf you don't seem to have a
>> working DNS provider.

> If all else fails, then you can use opendns.com - they provide their
> services for free!

I second that recommendation. I have been using OpenDns.com for about
six months now. Eliminated all the problems we had, when the DNS
servers at our ISP were down or slow.
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos

Re: [CentOS] DNS

[Filipe Brandenburger]

Hi,

On Tue, Jan 27, 2009 at 11:48, Ralph Angenendt  wrote:
> adrian kok wrote:
>> Sorry. I use fake name
>
> No, you didn't. www.mydomain.com does exist, so it cannot be fake.

Interestingly enough, there are domains specially reserved for "fake"
names in examples: example.com, example.net and example.org.

There is even an RFC for that:
http://www.faqs.org/rfcs/rfc2606.html

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

Re: [CentOS] possible CentOS 5.2 yum RPM checksum issue upstream?

[Scott Silva]

on 1-26-2009 5:12 PM Justin M. Hunter spake the following:
> hey guys..
> 
> I'm continually seeing the following error although I've got the
> latest/greatest yum RPM from various CentOS 5.2 repos
> (mirrors.kernel.org  most recently). is there
> possibly something wrong with the RPM's checksum somewhere upstream? 
> 
> note: I've already regenerated our own mrepo server various times and
> refreshed the metadata on my yum clients as well.
> 
> ***
> 
> ~# yum clean all; yum update
> 
> ...
> 
> Transaction Summary
> =
> Install 2 Package(s) 
> Update 57 Package(s) 
> Remove 0 Package(s) 
> 
> Total download size: 97 M
> Is this ok [y/N]: y
> Downloading Packages:
> (1/2): yum-3.2.8-9.el5.ce 100% |=| 582 kB 00:00 
> http://oak-ss-admin001.sixapart.net/mrepo/centos5.2-x86_64/RPMS.updates/yum-3.2.8-9.el5.centos.2.1.noarch.rpm:
> [Errno -1] Package does not match intended download
> Trying other mirror.
> (2/2): yum-fastestmirror- 100% |=| 13 kB 00:00 
> 
> ~# md5sum /var/tmp/yum-3.2.8-9.el5.centos.2.1.noarch.rpm 
> d776014eb266b87299465cb38e0f6b96
>  /var/tmp/yum-3.2.8-9.el5.centos.2.1.noarch.rpm
> 
> ***
> 
Or it could be getting corrupted by an upstream proxy if you have one.

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



signature.asc
Description: OpenPGP digital signature
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos

Re: [CentOS] How to add network driver for network installation

[Scott Silva]

on 1-27-2009 12:04 AM MOKRANI Rachid spake the following:
> Hi,
> 
> Is someone has an experience for adding a network driver when starting a
> network installation.
> 
> My network internal card is : Intel 82567M Gigabit  (laptop HP 8530w)
> 
> This card is not recognized by my system : CentOS 5.2 x86_64 (kernel
> 2.6.9-92) 
> 
> How to add correctly this driver for a network installation ?
> 
> I can add the Intel driver  e1000e-0.5.11.2 after the installation and
> the network card works fine, but I need to do network installation, and
> this card need to be recognized before starting.
> 
> Any help will be very appreciated.
> 
> Best Regards.

If you can wait a few weeks,  the 5.3 installer might have this driver in it.

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



signature.asc
Description: OpenPGP digital signature
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos

Re: [CentOS] OT : iptables/arptables question

[Fabian Arrotin]

Filipe Brandenburger wrote:
> Hi Fabian:
> 
> On Tue, Jan 27, 2009 at 08:16, Fabian Arrotin  
> wrote:
>> Question is : how can i "mangle" output packets to appear coming from
>> public ip and not from 192.168.X.X ?
> 
> Found this that might help you (google for: linux default outgoing ip):
> 
> """
>> On a machine with multiple interfaces, is it possible to set the default
>> outgoing IP address to something other than the address for the interface
>> on the outgoing route?
> 
> Yes.
> 
> ip route add 10.1.1.0/24 via 192.168.1.1 src 172.16.1.1
>   ^^^
> The src parameter tells the routing code to use this address when sending
> packets. The address only needs to be on the system. IE:
> 
> ip addr add 172.16.1.1/32 dev dummy0
> 
> And send the packets out of eth0.
> """
> From: http://lkml.indiana.edu/hypermail/linux/kernel/0112.1/0359.html
> 
> Just make sure you keep a separate route for your ISP's side of the
> private network (maybe the one created when your interface goes up
> will do), otherwise your routing protocol might fail.
> 
> HTH,
> Filipe

Hi Filipe,

thanks for the link, i completely missed that point from the 'ip route' 
command.
On the other hand, 10 minutes after i had sent my mail (and 3 coffee 
later to be precise) i saw also a picture from wikipedia 
(http://upload.wikimedia.org/wikipedia/fr/thumb/3/3e/Netfilter_schema.png/400px-Netfilter_schema.png)
 
explaining how packets went through the different ip tables and it was 
directly clear : even packets leaving the local box (and being processed 
in the OUTPUT filter) are still processed in the nat table (postrouting 
filter) so a simple SNAT rule did the job perfectly too ;-) In fact it's 
the first time that i have to modify packets leaving a linux gateway and 
i thought that only packets being forwarded (and so traversing the 
FORWARD filter) could also being modified in the nat table ...
I've also had a look in the sysconfig.txt file to see how your solution 
could be applied but it's still not very clear how that can be done. But 
using GATEWAYDEV=eth3 (eth3 having my public-ip/32 while eth3:1 having 
my 192.168.X.X/24 ip) in the /etc/sysconfig/network and declaring a 
GATEWAY=192.168.X.X (isp router ip) in the ifcfg-eth3:1 does also the 
job. But a `route -n` is strange though : " 0.0.0.0 0.0.0.0 
 0.0.0.0 U 0  00 eth3" like for the old ppp 
stuff

So multiple ways to solve the initial question ...
--
Fabian Arrotin
  idea=`grep -i clue /dev/brain` ; test -z "$idea" && echo "sorry, init 
6 in progress" || sh ./answer.sh

-- 
--
Fabian Arrotin
  idea=`grep -i clue /dev/brain` ; test -z "$idea" && echo "sorry, init 
6 in progress" || sh ./answer.sh
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos

[CentOS] network question

[ann kok]

Hi

How can I clear the arp cache?

and 

ls it possible to setup 2 gateways in centos

Thank you for your help


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

Re: [CentOS] possible CentOS 5.2 yum RPM checksum issue upstream?

[Justin M. Hunter]

hrm. I've got nothing sitting between me and my upstream providers (that I
know of).

On Tue, Jan 27, 2009 at 9:45 AM, Scott Silva  wrote:

> on 1-26-2009 5:12 PM Justin M. Hunter spake the following:
> > hey guys..
> >
> > I'm continually seeing the following error although I've got the
> > latest/greatest yum RPM from various CentOS 5.2 repos
> > (mirrors.kernel.org  most recently). is there
> > possibly something wrong with the RPM's checksum somewhere upstream?
> >
> > note: I've already regenerated our own mrepo server various times and
> > refreshed the metadata on my yum clients as well.
> >
> > ***
> >
> > ~# yum clean all; yum update
> >
> > ...
> >
> > Transaction Summary
> >
> =
> > Install 2 Package(s)
> > Update 57 Package(s)
> > Remove 0 Package(s)
> >
> > Total download size: 97 M
> > Is this ok [y/N]: y
> > Downloading Packages:
> > (1/2): yum-3.2.8-9.el5.ce 100% |=| 582 kB 00:00
> >
> http://oak-ss-admin001.sixapart.net/mrepo/centos5.2-x86_64/RPMS.updates/yum-3.2.8-9.el5.centos.2.1.noarch.rpm
> :
> > [Errno -1] Package does not match intended download
> > Trying other mirror.
> > (2/2): yum-fastestmirror- 100% |=| 13 kB 00:00
> >
> > ~# md5sum /var/tmp/yum-3.2.8-9.el5.centos.2.1.noarch.rpm
> > d776014eb266b87299465cb38e0f6b96
> >  /var/tmp/yum-3.2.8-9.el5.centos.2.1.noarch.rpm
> >
> > ***
> >
> Or it could be getting corrupted by an upstream proxy if you have one.
>
> --
> 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
>
>
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos

Re: [CentOS] network question

[nate]

ann kok wrote:
> Hi
>
> How can I clear the arp cache?

I don't see a way to clear it but you can delete the entries
with arp -d


>
> and
>
> ls it possible to setup 2 gateways in centos

2 gateways or 2 default gateways? If you need more than one
default gateway you likely will need to use iproute2 to set
that up.

If you have 1 default gateway and want to have another route(s)
pointing to another gateway add them to /etc/sysconfig/static-routes

e.g.

any net 192.168.0.0 netmask 255.255.0.0 gw 10.254.113.1


not sure why the syntax is not the same as the route command,
it's similar but not exactly the same.

nate


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

Re: [CentOS] OT : iptables/arptables question

[Fabian Arrotin]

Fabian Arrotin wrote:
> Filipe Brandenburger wrote:
>> Hi Fabian:
>>
>> On Tue, Jan 27, 2009 at 08:16, Fabian Arrotin  
>> wrote:
>>> Question is : how can i "mangle" output packets to appear coming from
>>> public ip and not from 192.168.X.X ?
>> Found this that might help you (google for: linux default outgoing ip):
>>
>> """
>>> On a machine with multiple interfaces, is it possible to set the default
>>> outgoing IP address to something other than the address for the interface
>>> on the outgoing route?
>> Yes.
>>
>> ip route add 10.1.1.0/24 via 192.168.1.1 src 172.16.1.1
>>   ^^^
>> The src parameter tells the routing code to use this address when sending
>> packets. The address only needs to be on the system. IE:
>>
>> ip addr add 172.16.1.1/32 dev dummy0
>>
>> And send the packets out of eth0.
>> """
>> From: http://lkml.indiana.edu/hypermail/linux/kernel/0112.1/0359.html
>>
>> Just make sure you keep a separate route for your ISP's side of the
>> private network (maybe the one created when your interface goes up
>> will do), otherwise your routing protocol might fail.
>>
>> HTH,
>> Filipe
> 
> Hi Filipe,
> 
> thanks for the link, i completely missed that point from the 'ip route' 
> command.
> On the other hand, 10 minutes after i had sent my mail (and 3 coffee 
> later to be precise) i saw also a picture from wikipedia 
> (http://upload.wikimedia.org/wikipedia/fr/thumb/3/3e/Netfilter_schema.png/400px-Netfilter_schema.png)
>  
> explaining how packets went through the different ip tables and it was 
> directly clear : even packets leaving the local box (and being processed 
> in the OUTPUT filter) are still processed in the nat table (postrouting 
> filter) so a simple SNAT rule did the job perfectly too ;-) In fact it's 
> the first time that i have to modify packets leaving a linux gateway and 
> i thought that only packets being forwarded (and so traversing the 
> FORWARD filter) could also being modified in the nat table ...
> I've also had a look in the sysconfig.txt file to see how your solution 
> could be applied but it's still not very clear how that can be done. But 
> using GATEWAYDEV=eth3 (eth3 having my public-ip/32 while eth3:1 having 
> my 192.168.X.X/24 ip) in the /etc/sysconfig/network and declaring a 
> GATEWAY=192.168.X.X (isp router ip) in the ifcfg-eth3:1 does also the 
> job. But a `route -n` is strange though : " 0.0.0.0 0.0.0.0 
>  0.0.0.0 U 0  00 eth3" like for the old ppp 
> stuff
> 
> So multiple ways to solve the initial question ...
>

one coffee cup later i see in sysconfig.txt documentation file for the 
paramaters of ifcfg- : "SRCADDR=  use the specified 
source address for outgoing packets" .. so definitely resolved by 
sysconfig files (so a clean solution)


-- 
--
Fabian Arrotin
  idea=`grep -i clue /dev/brain` ; test -z "$idea" && echo "sorry, init 
6 in progress" || sh ./answer.sh
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos

Re: [CentOS] network question

[John R Pierce]

ann kok wrote:
> How can I clear the arp cache?
>
>   

man arp  shows the options.   arp caches expire fairly quickly 
though, I don't think I've ever needed to manually expire one.

> and 
>
> ls it possible to setup 2 gateways in centos
>   

you can have various static routes to different gateways for specific 
subnets.only one 'default' gateway will effectively work, if more 
than one is defined, its somewhat indeterminate as to what happens.

now, you -can- play games with 'ip route' and 'ip table' and have 
alternate route tables, and use tagging to identify packets you want to 
route through an alternate gateway, this gets kinda messy.

example...  3 interfaces, eth0 -> one internet connection with a 
static subnet, eth1 -> another internet connection with a static subnet, 
and eth2 -> LAN as 10.0.0.0/16

The regular CentOS networknig is setup to use eth0's default gateway, 
and have all 3 subnets properly configured.NAT rules are setup so 
any host on the LAN on 10.0.0.* is to be routed to eth0, and any host on 
10.0.1.* is to be routed to eth1.  futher, a bunch of IPs on eth1 are 
mapped to specific server hosts on 10.0.1.* ... that part is fairly 
straight forward.   the actually tricky part is to ensure that packets 
to/from these eth1 mapped hosts only exit via eth1...  thats done as 
follows...

# network of eth1 interface
eth1net=100.100.100.0/24
# internet gateway of eth1 interface
eth1gate=100.100.100.1
ip rule add from 10.0.1.0/24 table 200
ip route add default via $eth1gate dev eth1 table 200
ip route flush cache

(where 100.100.100.* is the internet address of this eth1 subnet) 
the ip rule command 'tags' any packets from 10.0.1.0/24 to use this 
alternate table '200' (if you prefer, you can name these tables via 
/etc/iproute2/rt_tables)
the ip route add command sets an alternate gateway only for packets 
using this special table, and the flush command makes sure any route 
caching is cleaned.   this `ip` command (actually /sbin/ip) is part of 
the iproute2 package.  

this stuff gets quite tricky fast, and requires you to really understand 
what you're doing.

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

Re: [CentOS] OT : iptables/arptables question

[Filipe Brandenburger]

Hi,

On Tue, Jan 27, 2009 at 14:14, Fabian Arrotin  wrote:
> On the other hand, 10 minutes after i had sent my mail (and 3 coffee
> later to be precise) i saw also a picture from wikipedia
> (http://upload.wikimedia.org/wikipedia/fr/thumb/3/3e/Netfilter_schema.png/400px-Netfilter_schema.png)

Interesting diagram! I tried to look for the Wikipedia page that links
to it, but no luck. Where did you find it? Is there an article with
that?

> explaining how packets went through the different ip tables and it was
> directly clear : even packets leaving the local box (and being processed
> in the OUTPUT filter) are still processed in the nat table (postrouting
> filter) so a simple SNAT rule did the job perfectly too ;-)

That was my first thought, configuring a NAT for it. I actually
thought that you would be able to solve it by adding a SNAT entry in
the OUTPUT chain of the nat table, but when I checked the man page I
saw that SNAT can only be used in POSTROUTING. As I assumed you
already had a POSTROUTING SNAT rule catching all and you still had the
problem with the firewall itself, I did not mention it, but now I see
that it makes sense as you need a separate rule for that as the
interface is different.

> I've also had a look in the sysconfig.txt file to see how your solution
> could be applied but it's still not very clear how that can be done. But
> using GATEWAYDEV=eth3 (eth3 having my public-ip/32 while eth3:1 having
> my 192.168.X.X/24 ip) in the /etc/sysconfig/network and declaring a
> GATEWAY=192.168.X.X (isp router ip) in the ifcfg-eth3:1 does also the
> job. But a `route -n` is strange though : " 0.0.0.0 0.0.0.0
> 0.0.0.0 U 0  00 eth3" like for the old ppp
> stuff

> one coffee cup later i see in sysconfig.txt documentation file for the
> paramaters of ifcfg- : "SRCADDR=  use the specified
> source address for outgoing packets" .. so definitely resolved by
> sysconfig files (so a clean solution)

Cool! That's real easy! Glad to know that.

> So multiple ways to solve the initial question ...

Great! I learned something today! :-)

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

Re: [CentOS] network question

[Robert Moskowitz]

John R Pierce wrote:
> ann kok wrote:
>   
>> How can I clear the arp cache?
>>
>>   
>> 
>
> man arp  shows the options.   arp caches expire fairly quickly 
> though, I don't think I've ever needed to manually expire one.
>   

Per Dr. Larry Peterson who was 'there' as a grad student at Purdue under 
Dr. Comer

ARP time outs were set at 10 min as Dr. Comer set up 2 teams of grad 
students that had to power down one of the DECs, swap the ethernet card, 
and bring the system back up. The fastest team did this in 10 min, thus 
ARP time out was set to 10 min...

Oh, Dr. Noel Chiappa 'created' ARP as a hack while he was a post-doc 
researcher at MIT.

>   
>> and 
>>
>> ls it possible to setup 2 gateways in centos
>>   
>> 
>
> you can have various static routes to different gateways for specific 
> subnets.only one 'default' gateway will effectively work, if more 
> than one is defined, its somewhat indeterminate as to what happens.
>   

There is extensive work in the IETF around multihoming systems. This is 
becoming more critical as we are interfacing systems to multiple network 
types, and standards like IEEE 802.21 handoff.

Basically it is HARD and current stuff does it wrong, and it will be a 
while before we agree on how to do something that has a chance of success.

Tread lightly here unless you are a researcher or are desperate

And my HIP protocol is one of the technologies being used as a model for 
how we COULD get multihoming working.

> now, you -can- play games with 'ip route' and 'ip table' and have 
> alternate route tables, and use tagging to identify packets you want to 
> route through an alternate gateway, this gets kinda messy.
>
> example...  3 interfaces, eth0 -> one internet connection with a 
> static subnet, eth1 -> another internet connection with a static subnet, 
> and eth2 -> LAN as 10.0.0.0/16
>
> The regular CentOS networknig is setup to use eth0's default gateway, 
> and have all 3 subnets properly configured.NAT rules are setup so 
> any host on the LAN on 10.0.0.* is to be routed to eth0, and any host on 
> 10.0.1.* is to be routed to eth1.  futher, a bunch of IPs on eth1 are 
> mapped to specific server hosts on 10.0.1.* ... that part is fairly 
> straight forward.   the actually tricky part is to ensure that packets 
> to/from these eth1 mapped hosts only exit via eth1...  thats done as 
> follows...
>
> # network of eth1 interface
> eth1net=100.100.100.0/24
> # internet gateway of eth1 interface
> eth1gate=100.100.100.1
> ip rule add from 10.0.1.0/24 table 200
> ip route add default via $eth1gate dev eth1 table 200
> ip route flush cache
>
> (where 100.100.100.* is the internet address of this eth1 subnet) 
> the ip rule command 'tags' any packets from 10.0.1.0/24 to use this 
> alternate table '200' (if you prefer, you can name these tables via 
> /etc/iproute2/rt_tables)
> the ip route add command sets an alternate gateway only for packets 
> using this special table, and the flush command makes sure any route 
> caching is cleaned.   this `ip` command (actually /sbin/ip) is part of 
> the iproute2 package.  
>
> this stuff gets quite tricky fast, and requires you to really understand 
> what you're doing.
>
> ___
> CentOS mailing list
> CentOS@centos.org
> http://lists.centos.org/mailman/listinfo/centos
>
>   
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos

Re: [CentOS] OT : iptables/arptables question

[Fabian Arrotin]

Filipe Brandenburger wrote:
> Hi,
> 
> On Tue, Jan 27, 2009 at 14:14, Fabian Arrotin  
> wrote:
>> On the other hand, 10 minutes after i had sent my mail (and 3 coffee
>> later to be precise) i saw also a picture from wikipedia
>> (http://upload.wikimedia.org/wikipedia/fr/thumb/3/3e/Netfilter_schema.png/400px-Netfilter_schema.png)
> 
> Interesting diagram! I tried to look for the Wikipedia page that links
> to it, but no luck. Where did you find it? Is there an article with
> that?

In fact that's a diagram i found through google that pointed me to the 
french version of wikipedia for netfilter : 
http://fr.wikipedia.org/wiki/Netfilter .. strange that the english 
version doesn't have it because such picture talks more than a paragraph ...

-- 
--
Fabian Arrotin
  idea=`grep -i clue /dev/brain` ; test -z "$idea" && echo "sorry, init 
6 in progress" || sh ./answer.sh
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos

Re: [CentOS] More than 2TB RAID...

[Jake]

On Tue, Jan 27, 2009 at 7:20 AM, John Doe  wrote:

> Hi,
>
> I just received a new server (HP DL180G5) with 12x 1TB HDs and I bumped
> into fdisks 2TB limits...
> Since this is an entry level server, I can't use the classic HP bootable
> utilities to create smaller volumes et can only create a big RAID6.
> I found out that: using parted, labelling it gpt and creating the
> partitions would do the trick.
> But, what about grub?  I read that it does not support gpt...
> Or is there another way to do it?
> By example, do you think I could boot on a live CentOS, install HP RAID
> tools on it (is it possible?) and then create volumes...?
>
> Thx,
> JD



I came across this article you may find useful:
http://www.unixgods.org/~tilo/linux_larger_2TB.html

I should say that I STRONGLY recommend not creating ext3 file systems in the
2TB+ range - fsck takes too long and you'd hate to get hit by one of those
in what is supposed to be a "quick" reboot...and disabling them on the file
system isn't a good idea either.

-- 
Jake Paulus
jakepau...@gmail.com
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos

Re: [CentOS] More than 2TB RAID...

[Joshua Baker-LePain]

On Tue, 27 Jan 2009 at 6:43pm, Jake wrote

> I should say that I STRONGLY recommend not creating ext3 file systems in the
> 2TB+ range - fsck takes too long and you'd hate to get hit by one of those
> in what is supposed to be a "quick" reboot...and disabling them on the file
> system isn't a good idea either.

On the other hand, nothing is as well supported on RHEL/CentOS as is ext3. 
So if you're data is really important to you, think hard about using 
another FS.

-- 
Joshua Baker-LePain
QB3 Shared Cluster Sysadmin
UCSF
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos

[CentOS] Missing latest kmod-xfs and kmod-drbd82?

[Alan Sparks]

Since December there have been several kernel updates to the centosplus
repository for centos4... but no updates to kmod-xfs or kmod-drbd82. 
Latest kernel is 2.6.9-78.0.13.plus.c4, but latest kmod-xfs is
2.6.9_78.0.5.plus.c4.  Latest drbd82 kmod is also for 78.0.5.

I know there was a mention on the list that these modules are not kernel
dependent anymore... but yum and rpm seem to indicate otherwise, and
trying to install kmod-xfs-smp or update kmod-drbd82-smp on my servers
is not possible.

$ rpm -qRp kmod-xfs-smp-0.4-1.el4.2.6.9_78.0.5.plus.c4.x86_64.rpm
/bin/sh
/bin/sh
/sbin/depmod
/sbin/depmod
kernel-smp-x86_64 = 2.6.9-78.0.5.plus.c4
rpmlib(CompressedFileNames) <= 3.0.4-1
rpmlib(PayloadFilesHavePrefix) <= 4.0-1
rpmlib(VersionedDependencies) <= 3.0.3-1

Above, the last available version of the module shows a dependency on
the 78.0.5 kernel.  These modules will of course also install in the
wrong /lib/modules/ location.

I've over the last two months opened tickets on bugs.centos.org on the
issue (tickets 3360, 3316, 3291).  However, these tickets are still new
and have had no comments.

Anyone have any advice on dealing with the missing RPMs or the
dependency issue?  Thanks in advance.
-Alan


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

Re: [CentOS] ftp and iptables

[Agile Aspect]

Robert Spangler wrote:
> Do you have a rule like this:
>
> -A OUTPUT --m state --state RELATED,ESTABLISHED -j ACCEPT 
>   
No I don't.

It doesn't work under CentOS 5.2. But it works on my laptop
which is running Fedora 9.
> If not you should place this in your rules.  This rule eleminates the need to 
> continuesly add rules to allow out going connection for allowed incoming 
> connection.
>
> If you do then you should not need the OUTPUT rules you listed above.
>
>   
Thanks for the reply!

-- 
Article. VI. Clause 3 of the constitution of the United States states: 

"The Senators and Representatives before mentioned, and the Members of 
the several State Legislatures, and all executive and judicial Officers, 
both of the United States and of the several States, shall be bound by 
Oath or Affirmation, to support this Constitution; but no religious Test 
shall ever be required as a Qualification to any Office or public Trust 
under the United States." 


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

[CentOS] getting Centos on "used" rackable systems 1U with dual Opteron 248 HE

[RobertH]

hi there.

anyone have any tips on params to pass to the kernel etc in order to install
Centos4 or Centos5 32bit or 64 bit installed on this unit

i can get it close to install, yet no cigar.

it is a tyan MB and ill try to get more info if needed...

for this case, please assume for now that the hardware itself is not "bad"

thanks in advance...

 - rh

Rackable Systems Dual Opteron 2.0 GHz 1U Half-Depth Server 

General Features: 
No Server Operating System 
Dual AMD Opteron 246 HE (OSK246CMP5AU) 2.0 GHz processors 
4 GB DDR RAM  
120 GB IDE hard drive 
ATI Rage XL integrated video 
Broadcom BCM5704 Gigabit Ethernet 


Motherboard Features: 
AMD-8111 chipset 
One (1) PCI-X riser slot 
Eight (8) DDR DIMM sockets  


Front I/O Ports: 
Two (2) PS/2 ports 
One (1) 15-pin VGA port 
One (1) 9-pin serial port  
Two (2) USB ports 
Two (2) RJ-45 Ethernet jacks 


Case Features: 
1U half-depth server chassis 
One (1) 3.5-inch drive bay 
Four rear cooling fans 
400-watt power supply 


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

Re: [CentOS] getting Centos on "used" rackable systems 1U with dualOpteron 248 HE

[bruce]

hi robert...

what issues are you having? what params are you using as of now...



-Original Message-
From: centos-boun...@centos.org [mailto:centos-boun...@centos.org]on
Behalf Of RobertH
Sent: Wednesday, January 28, 2009 7:04 PM
To: 'CentOS mailing list'
Subject: [CentOS] getting Centos on "used" rackable systems 1U with
dualOpteron 248 HE


hi there.

anyone have any tips on params to pass to the kernel etc in order to install
Centos4 or Centos5 32bit or 64 bit installed on this unit

i can get it close to install, yet no cigar.

it is a tyan MB and ill try to get more info if needed...

for this case, please assume for now that the hardware itself is not "bad"

thanks in advance...

 - rh

Rackable Systems Dual Opteron 2.0 GHz 1U Half-Depth Server

General Features:
No Server Operating System
Dual AMD Opteron 246 HE (OSK246CMP5AU) 2.0 GHz processors
4 GB DDR RAM
120 GB IDE hard drive
ATI Rage XL integrated video
Broadcom BCM5704 Gigabit Ethernet


Motherboard Features:
AMD-8111 chipset
One (1) PCI-X riser slot
Eight (8) DDR DIMM sockets


Front I/O Ports:
Two (2) PS/2 ports
One (1) 15-pin VGA port
One (1) 9-pin serial port
Two (2) USB ports
Two (2) RJ-45 Ethernet jacks


Case Features:
1U half-depth server chassis
One (1) 3.5-inch drive bay
Four rear cooling fans
400-watt power supply


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

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

Re: [CentOS] More than 2TB RAID...

[Ross Walker]

On Jan 27, 2009, at 6:46 PM, Joshua Baker-LePain  wrote:

> On Tue, 27 Jan 2009 at 6:43pm, Jake wrote
>
>> I should say that I STRONGLY recommend not creating ext3 file  
>> systems in the
>> 2TB+ range - fsck takes too long and you'd hate to get hit by one  
>> of those
>> in what is supposed to be a "quick" reboot...and disabling them on  
>> the file
>> system isn't a good idea either.
>
> On the other hand, nothing is as well supported on RHEL/CentOS as is  
> ext3.
> So if you're data is really important to you, think hard about using
> another FS.

This is sound advice.

Best way is probably to use lvm and create multiple 2TB LVs with ext3.  
Disable auto-fsck and look at doing proactive online fsck. There was a  
post about a month ago on performing online fscks with snapshots.

For that kind of storage I might look at Solaris and ZFS though.

-Ross

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

Re: [CentOS] ftp and iptables

[Chaz Sliger]

Have you loaded the ftp modules?
  modprobe ip_conntrack_ftp
  modprobe ip_nat_ftp

-chaz


-Original Message-
From: centos-boun...@centos.org [mailto:centos-boun...@centos.org] On Behalf
Of Agile Aspect
Sent: Tuesday, January 27, 2009 6:45 PM
To: CentOS mailing list
Subject: Re: [CentOS] ftp and iptables

Robert Spangler wrote:
> Do you have a rule like this:
>
> -A OUTPUT --m state --state RELATED,ESTABLISHED -j ACCEPT 
>   
No I don't.

It doesn't work under CentOS 5.2. But it works on my laptop
which is running Fedora 9.
> If not you should place this in your rules.  This rule eleminates the need
to 
> continuesly add rules to allow out going connection for allowed incoming 
> connection.
>
> If you do then you should not need the OUTPUT rules you listed above.
>
>   
Thanks for the reply!

-- 
Article. VI. Clause 3 of the constitution of the United States states: 

"The Senators and Representatives before mentioned, and the Members of 
the several State Legislatures, and all executive and judicial Officers, 
both of the United States and of the several States, shall be bound by 
Oath or Affirmation, to support this Constitution; but no religious Test 
shall ever be required as a Qualification to any Office or public Trust 
under the United States." 


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


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

Re: [CentOS] More than 2TB RAID...

[Ian Forde]

On Tue, 2009-01-27 at 18:46 -0500, Joshua Baker-LePain wrote:
> On Tue, 27 Jan 2009 at 6:43pm, Jake wrote
> 
> > I should say that I STRONGLY recommend not creating ext3 file systems in the
> > 2TB+ range - fsck takes too long and you'd hate to get hit by one of those
> > in what is supposed to be a "quick" reboot...and disabling them on the file
> > system isn't a good idea either.
> 
> On the other hand, nothing is as well supported on RHEL/CentOS as is ext3. 
> So if you're data is really important to you, think hard about using 
> another FS.

Actually, on RHEL, the *only* filesystems that upstream *officially*
supports are ext2/3 and GFS.  Not XFS, nor reiser, nor JFS.  Nada...

Well, maybe FAT for USB-attached storage... ;)

But if you're using CentOS, it's entirely up to you... If I were in
RHEL-land (meaning: at a company willing to pony up for licenses), I'd
consider a GFS2 cluster shared out via NFS.  Or maybe an OCFS2 NFS
cluster.  If at a company using CentOS, I'd consider an OFCS2/NFS
cluster or heartbeat/XFS/NFS.  For home? XFS (or JFS if you like).  But
then, I'm willing (and capable) of supporting the mess I create.  It all
depends upon one's comfort level with getting out of a jam when one
strays out of the "sweet spots" of available help...

-I

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

Re: [CentOS] getting Centos on "used" rackable systems 1U withdualOpteron 248 HE

[RobertH]

 

> Subject: Re: [CentOS] getting Centos on "used" rackable 
> systems 1U withdualOpteron 248 HE
> 
> hi robert...
> 
> what issues are you having? what params are you using as of now...
> 
> 
> 

u basically with centos 4 or 5 64 bit, it is freezing when i get to the
partitioning and want to continue forward or after partitioning and moving
towards the um place where one chooses software.

4.7 seems to work better than 5.2

would not work in graphical install mode. hangs

so i went to

linux text

...and i am not using any kernel params *yet* other than one time i tried

nousb nousbstorage skipddc

just for the heck of it, which worked on old compaq DL380 servers hehehehe
;-)

what i will try next is to turn off power mgmt and apic or whatever

i was wondering if others had bought any of these boxes and played with them
yet. they are pretty cheap and i have been looking for some cheap rackmount
boxes that i can throw large ata ide drives in to do some things with

 - rh

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

Re: [CentOS] yum problem

[Mad Unix]

it works now added the following to the profile
export http_proxy="http://foo:b...@10.0.0.1:8080";


On Tue, Jan 27, 2009 at 6:57 PM, Lanny Marcus  wrote:
> On Tue, Jan 27, 2009 at 4:04 AM, Anthony Kamau  wrote:
>>> -Original Message-
>>> From: centos-boun...@centos.org [mailto:centos-boun...@centos.org] On
>>> Behalf Of James Bensley
>>> Sent: Tuesday, 27 January 2009 7:49 PM
>>> To: CentOS mailing list
>>> Subject: Re: [CentOS] yum problem
>>>
>>> It doesn't matter if its in resolv.conf you don't seem to have a
>>> working DNS provider.
> 
>> If all else fails, then you can use opendns.com - they provide their
>> services for free!
>
> I second that recommendation. I have been using OpenDns.com for about
> six months now. Eliminated all the problems we had, when the DNS
> servers at our ISP were down or slow.
> ___
> CentOS mailing list
> CentOS@centos.org
> http://lists.centos.org/mailman/listinfo/centos
>
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos

Re: [CentOS] DNS

[James Bensley]

I don't quite follow this thread?

Are your pinging your web server, its is hosted with a hosting company
(i.e. you don't personal host it on a server you manage?)

It could just be the details for their set up (i.e. multiple servers
with load balancing across multiple IPs?)

-BEGIN GEEK CODE BLOCK-
  Version: 3.1
GIT/MU/U dpu s: a--> C++>$ U+> L++> B-> P+> E?> W+++>$ N K W++ O M++>$ V-
PS+++ PE++ Y+ PGP t 5 X+ R- tv+ b+> DI D+++ G+ e(+) h--(++) r++ z++
--END GEEK CODE BLOCK--
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos