CentOS - Apr 2011 - CentOS 5.6 x86_64 hosting unreachable Virtualbox 4.0.6 bridged guests

Hi,

On a box running CentOS 5.6 x86_64 and Virtualbox 4.0.6, I have 
configured 2 guests (WinXP x86 + CentOS 5.6 x86_64) in bridged mode.
The hosting box and the two guests are dhcp enabled. The hosting box 
get its address normally, but the 2 hosted systems fail to get theirs.
The dhcp server log tells : DHCPDISCOVER from 08:00:xx:xx:xx:xx via 
em0: network xx.xx/16: no free leases.
When changing from dhcp to static ip in the CentOS guest, 'service 
network restart' does not report any errors, but the guest keeps 
unreachable from the outside : pinging the guest works some times, but 
trying to reach the guest with ssh, http, or ftp always fail, with or 
whitout iptables enabled on the host and/or the guests.

Please note that:
- Kernel and softwares on the hosting CentOS are up-to-date, the box 
has an Intel Corporation 82567LM-3 Gigabit Network Connection (rev 02) 
NIC, does not have an Intel VT compliant CPU (Pentium E5700), but 
VT/AMD-V enabled in bios.
- The host is not configured in bridged mode (does not have a br 
interface configured), Oracle telling that bridged mode on the host is 
not required to run Virtualbox guests in bridged mode.
- The 2 guests work perfectly in NAT mode, but this mode is not an 
option for me.
- The same 2 guests work perfectly in bridged mode on another same 
hosting box running Fedora 14.

Is anyone experimenting the same difficulties with Virtualbox on CentOS 
5.6 ? I don't know where to search to solve this problem.
Thanks for your help !

-- 
Philippe

On 04/26/2011 03:28 AM, Philippe wrote:
> Hi,
> 
> On a box running CentOS 5.6 x86_64 and Virtualbox 4.0.6, I have 
> configured 2 guests (WinXP x86 + CentOS 5.6 x86_64) in bridged mode.
> The hosting box and the two guests are dhcp enabled. The hosting box 
> get its address normally, but the 2 hosted systems fail to get theirs.
> The dhcp server log tells : DHCPDISCOVER from 08:00:xx:xx:xx:xx via 
> em0: network xx.xx/16: no free leases.
> When changing from dhcp to static ip in the CentOS guest, 'service 
> network restart' does not report any errors, but the guest keeps 
> unreachable from the outside : pinging the guest works some times, but 
> trying to reach the guest with ssh, http, or ftp always fail, with or 
> whitout iptables enabled on the host and/or the guests.
> 
> Please note that:
> - Kernel and softwares on the hosting CentOS are up-to-date, the box 
> has an Intel Corporation 82567LM-3 Gigabit Network Connection (rev 02) 
> NIC, does not have an Intel VT compliant CPU (Pentium E5700), but 
> VT/AMD-V enabled in bios.
> - The host is not configured in bridged mode (does not have a br 
> interface configured), Oracle telling that bridged mode on the host is 
> not required to run Virtualbox guests in bridged mode.
> - The 2 guests work perfectly in NAT mode, but this mode is not an 
> option for me.
> - The same 2 guests work perfectly in bridged mode on another same 
> hosting box running Fedora 14.
> 
> Is anyone experimenting the same difficulties with Virtualbox on CentOS 
> 5.6 ? I don't know where to search to solve this problem.
> Thanks for your help !
> 
Some things to check:

1.  Make sure the user that is running the VMs is in the vboxusers group.

2.  What is the status of SELinux on the host CentOS and the Guest
CentOS.  For troubleshooting purposes I would turn SELinux off on both
hosts .. if nothing changes, turn them back on.

Although, if they work correctly in NAT mode, then those are likely OK.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 253 bytes
Desc: OpenPGP digital signature
URL:
<http://lists.centos.org/pipermail/centos/attachments/20110426/98016e3d/attachment-0005.sig>

Philippe wrote:
> Hi,
>
> On a box running CentOS 5.6 x86_64 and Virtualbox 4.0.6, I have 
> configured 2 guests (WinXP x86 + CentOS 5.6 x86_64) in bridged mode.
> The hosting box and the two guests are dhcp enabled. The hosting box 
> get its address normally, but the 2 hosted systems fail to get theirs.
> The dhcp server log tells : DHCPDISCOVER from 08:00:xx:xx:xx:xx via 
> em0: network xx.xx/16: no free leases.
> When changing from dhcp to static ip in the CentOS guest, 'service 
> network restart' does not report any errors, but the guest keeps 
> unreachable from the outside : pinging the guest works some times, but 
> trying to reach the guest with ssh, http, or ftp always fail, with or 
> whitout iptables enabled on the host and/or the guests.
>
> Please note that:
> - Kernel and softwares on the hosting CentOS are up-to-date, the box 
> has an Intel Corporation 82567LM-3 Gigabit Network Connection (rev 02) 
> NIC, does not have an Intel VT compliant CPU (Pentium E5700), but 
> VT/AMD-V enabled in bios.
> - The host is not configured in bridged mode (does not have a br 
> interface configured), Oracle telling that bridged mode on the host is 
> not required to run Virtualbox guests in bridged mode.
> - The 2 guests work perfectly in NAT mode, but this mode is not an 
> option for me.
> - The same 2 guests work perfectly in bridged mode on another same 
> hosting box running Fedora 14.
>
> Is anyone experimenting the same difficulties with Virtualbox on CentOS 
> 5.6 ? I don't know where to search to solve this problem.
> Thanks for your help !
>   
I am running Virtualbox 3.1.2 in bridged mode with WinXP guest OS but 
use a fixed IP address as its in my home network, hardwired, and I want 
my DNS to be complete with forward and reverse lookups for my CentOS 
workstation host as well as my WinXP guest. I only use DHCP for my 
wireless clients (hard allocate IP and hostname based upon ethernet 
address) and true guests to my network. The bridged method works fine - 
no bridge on host just normal ethernet - bridged on Virtualbox. As it 
works I have not moved up to version 4 as yet - probably will when 
CentOS 6 gets released and I build a new workstation. There are just a 
few pesky apps that insist on using windows / IE and have no viable open 
source alternatives that keep me using a WinXP guest.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: rkampen.vcf
Type: text/x-vcard
Size: 322 bytes
Desc: not available
URL:
<http://lists.centos.org/pipermail/centos/attachments/20110426/7f378bde/attachment-0005.vcf>

--On Tuesday, April 26, 2011 10:28:40 AM +0200 Philippe <lists at
lrnx.ath.cx>
wrote:
> The hosting box 
> get its address normally, but the 2 hosted systems fail to get theirs.
> The dhcp server log tells : DHCPDISCOVER from 08:00:xx:xx:xx:xx via 
> em0: network xx.xx/16: no free leases.
Based on that message, I would suspect that probably either:
  - the DHCP server is not seeing distinct MAC addresses from your VMs
    vs your host, or
  - you have actually exhausted your DHCP address pool.

I'd look at those two first, based on what you see in your dhcp server
log.  You may have to get your dhcp server to tell you what leases
are currently in use.

Devin

Le 26/04/2011 11:26, Johnny Hughes a ?crit :
> On 04/26/2011 03:28 AM, Philippe wrote:
>> Hi,
>> On a box running CentOS 5.6 x86_64 and Virtualbox 4.0.6, I have 
>> configured 2 guests (WinXP x86 + CentOS 5.6 x86_64) in bridged mode.
>> The hosting box and the two guests are dhcp enabled. The hosting box 
>> get its address normally, but the 2 hosted systems fail to get theirs.
>> The dhcp server log tells : DHCPDISCOVER from 08:00:xx:xx:xx:xx via 
>> em0: network xx.xx/16: no free leases.
>> When changing from dhcp to static ip in the CentOS guest, 'service 
>> network restart' does not report any errors, but the guest keeps 
>> unreachable from the outside : pinging the guest works some times, but 
>> trying to reach the guest with ssh, http, or ftp always fail, with or 
>> whitout iptables enabled on the host and/or the guests.
>>
>> Please note that:
>> - Kernel and softwares on the hosting CentOS are up-to-date, the box 
>> has an Intel Corporation 82567LM-3 Gigabit Network Connection (rev 02) 
>> NIC, does not have an Intel VT compliant CPU (Pentium E5700), but 
>> VT/AMD-V enabled in bios.
>> - The host is not configured in bridged mode (does not have a br 
>> interface configured), Oracle telling that bridged mode on the host is 
>> not required to run Virtualbox guests in bridged mode.
>> - The 2 guests work perfectly in NAT mode, but this mode is not an 
>> option for me.
>> - The same 2 guests work perfectly in bridged mode on another same 
>> hosting box running Fedora 14.
>>
>> Is anyone experimenting the same difficulties with Virtualbox on CentOS
>> 5.6 ? I don't know where to search to solve this problem.
>> Thanks for your help !
>>
> 
> Some things to check:
> 
> 1.  Make sure the user that is running the VMs is in the vboxusers group.
> 
Yes, the user running the vms belongs to vboxusers group.
> 2.  What is the status of SELinux on the host CentOS and the Guest
> CentOS.  For troubleshooting purposes I would turn SELinux off on both
> hosts .. if nothing changes, turn them back on.
> 
SE in strict, permissive, or simply disabled does not change anything.
> Although, if they work correctly in NAT mode, then those are likely OK.
Thanks for your response.
-- 
Philippe