minor issues on Cloudmin installation [#26241]

Submitted by ghomem on Mon, 04/08/2013 - 13:17

Following up a recent purchase I would like to report two minor issues on the installation:

1) According to https://www.virtualmin.com/documentation/cloudmin/virtualization/kvm for CentOS we must:

"In the /etc/sysconfig/network-scripts directory, copy ifcfg-eth0 to ifcfg-br0. [...] service network restart "

This is incomplete and the network won't come up after restart . You then have to remove the line with HWADDR that is present on ifcfg-br0 after the copy.

Edit: and even after that the interface only comes up if you fo

ifdown eth0; ifup eth0

"service network restart" doesn't make the br0 interface come up. That means the interface is not available after a reboot.

2) It appears that the Cloudmin installer needs the telnet command in order to check for the successfull startup of webmin. But the Centos minimal installation doesn't include the telnet command. Can you confirm this? If it is confirmed perhaps the installer can add "telnet" to the list of installed packages.

We will start testing Cloudmin and report other issues we may find.

Thanks for your support, so far.

Gustavo

Status:

Closed (fixed)

Comments

Submitted by ghomem on Mon, 04/08/2013 - 13:18 Comment #1

Submitted by ghomem on Mon, 04/08/2013 - 13:18 Comment #2

Submitted by JamieCameron on Mon, 04/08/2013 - 15:51 Comment #3

Thanks for pointing this out - I have updated the docs to mention removing the HWADDR line.

Regarding your second point about telnet - this shouldn't be used by the installer anywhere. Where are you seeing references to it exactly?

Submitted by ghomem on Mon, 04/08/2013 - 16:01 Comment #4

Jamie,

I added something to the first point meanwhile: the br0 interface doesn't come up unless we do "ifdown eth0; ifup eth0" after the configurations and after every boot as well.

Regarding telnet:

I didn't look at the code. I ran the installer once and it failed to detected the service running on port 10000, on the post install check that it runs. Then I installed telnet and re-ran the installer and if succeeded. So it was really just a guess.

Submitted by JamieCameron on Mon, 04/08/2013 - 16:03 Comment #5

Do you happen to recall what the exact error message was?

Submitted by ghomem on Mon, 04/08/2013 - 16:11 Comment #6

Before removing de HWADDR line the message was:

Apr 8 10:26:40 localhost /etc/sysconfig/network-scripts/ifdown-eth: Device br0 has MAC address 00:00:00:00:00:00, instead of configured address 00:25:90:83:93:04. Ignoring.

After removing it the error, during "service network restart" is

Apr 8 22:08:33 localhost /etc/sysconfig/network-scripts/ifup-eth: Device br0 does not seem to be present, delaying initialization.

Then you do

ifdown eth0 ifup eth0

and there is no error.

Something mustn't be well configured because the init script launch the network service and not ifup directly.

Submitted by JamieCameron on Mon, 04/08/2013 - 16:25 Comment #7

Actually, I was wondering about the error you got about the service not running on port 10000.

Submitted by ghomem on Tue, 04/09/2013 - 13:51 Comment #8

I don't have that error at the moment, but I can keep a copy after the next cloudmin installation.

Anyway, I'm less worried with this (since actually Cloudmin got correctly installed) than with the bridge interface configuration.

Submitted by JamieCameron on Tue, 04/09/2013 - 14:52 Comment #9

Would it be possible for you to attach the files ifcfg-eth0 and ifcfg-br0 to this bug report?

Submitted by ghomem on Tue, 04/09/2013 - 17:17 Comment #10

Here they are, after copying and removing the HWADDR line:

[root@localhost ~]# cat /etc/sysconfig/network-scripts/ifcfg-eth0 
DEVICE="eth0"
BOOTPROTO="static"
IPADDR="192.168.1.95"
NETMASK="255.255.255.0"
HWADDR="00:25:90:83:93:04"
NM_CONTROLLED="yes"
ONBOOT="yes"
#TYPE="Ethernet"
#UUID="80c3e2ae-63e6-4a7b-8d63-dcc6b82363a7"
BRIDGE=br0

[root@localhost ~]# cat /etc/sysconfig/network-scripts/ifcfg-br0 
DEVICE="br0"
BOOTPROTO="static"
IPADDR="192.168.1.95"
NETMASK="255.255.255.0"
NM_CONTROLLED="yes"
ONBOOT="yes"
#TYPE="Ethernet"
#UUID="80c3e2ae-63e6-4a7b-8d63-dcc6b82363a7"

Submitted by JamieCameron on Tue, 04/09/2013 - 18:36 Comment #11

I think the problem is that ifcfg-eth0 still contains IP addresses, when those should only be on br0. It should be more like :

IPV6INIT=yes
HWADDR=00:25:90:83:93:04
BOOTPROTO=none
NAME=eth0
TYPE=Ethernet
DEVICE=eth0
IPV4_FAILURE_FATAL=yes
ONBOOT=yes
USERCTL=no
BRIDGE=br0

Submitted by ghomem on Wed, 04/10/2013 - 12:52 Comment #12

There is more than that. If the TYPE variable of br0 is not set to "Bridge" the interfaces will come up in the wrong order, and therefore the bridge interface fail to start. These are the minimal files that work:

[root@localhost ~]# cat /etc/sysconfig/network-scripts/ifcfg-eth0

DEVICE=eth0

BOOTPROTO=none

TYPE=Ethernet

HWADDR=00:25:90:83:93:04

BRIDGE=br0

[root@localhost ~]# cat /etc/sysconfig/network-scripts/ifcfg-br0

GATEWAY=192.168.1.1

DEVICE=br0

BOOTPROTO=none

NETMASK=255.255.255.0

TYPE=Bridge

IPADDR=192.168.1.95

Please test and if confirmed to be correct update the instructions to reflect this.

Submitted by JamieCameron on Thu, 04/11/2013 - 01:27 Comment #13

Thanks - I will update the instructions to match.

Submitted by ghomem on Sun, 04/28/2013 - 16:07 Comment #14

I have documented this also on my side. It can be closed.