Topic: 6.03b USB entry in modules.conf breaks network

[Thony de Waal]

Dear fellow SME users,
I have a Dell Poweredge 400sc server with IDE RAID controller here that I want to install sme 6.03b on for a client.
I feared problems with the RAID controller, but that was detected and installed just fine as megaraid.
I can't use the installation though, because of an unexpected problem with the network.
The server has an internal Gigabit NIC and an extra 100 mb NIC that are recognised just fine. But the network will not start.
This is in the syslog:
Oct  2 14:03:01 secure network: Bringing up interface eth0:
Oct  2 14:03:01 secure ifup: modprobe: Invalid line 63 in /etc/modules.conf
Oct  2 14:03:01 secure ifup: ^Iehci-hcd
Oct  2 14:03:01 secure network: Bringing up interface eth0:  succeeded
Oct  2 14:03:01 secure network:
Oct  2 14:03:01 secure network: Bringing up interface eth1:
Oct  2 14:03:01 secure ifup: modprobe:
Oct  2 14:03:01 secure ifup: Invalid line 63 in /etc/modules.conf
Oct  2 14:03:01 secure ifup: ^Iehci-hcd
Oct  2 14:03:01 secure network: Bringing up interface eth1:  succeeded
Oct  2 14:03:01 secure network:
Oct  2 14:03:05 secure rc: Starting network:  succeeded

It says succeeded, but it really isn't.
When I remove the line that says: "ehci-hcd" in the /etc/modules.conf and issue a service network restart the network starts fine:
Oct  2 14:04:56 secure kernel: divert: allocating divert_blk for eth0
Oct  2 14:04:56 secure kernel: eth0: Intel(R) PRO/1000 Network Connection
Oct  2 14:04:57 secure kernel: e1000: eth0 NIC Link is Up 100 Mbps Full Duplex
Oct  2 14:04:58 secure network: Bringing up interface eth0:  succeeded
Oct  2 14:04:58 secure kernel: Intel(R) PRO/100 Network Driver - version 2.2.21-
k1
Oct  2 14:04:58 secure kernel: Copyright (c) 2003 Intel Corporation
Oct  2 14:04:58 secure kernel:
Oct  2 14:04:58 secure kernel: e100: selftest OK.
Oct  2 14:04:58 secure kernel: divert: allocating divert_blk for eth1
Oct  2 14:04:58 secure kernel: e100: eth1: Intel(R) PRO/100 Network Connection
Oct  2 14:04:58 secure kernel:   Hardware receive checksums enabled
Oct  2 14:04:58 secure kernel:   cpu cycle saver enabled
Oct  2 14:04:58 secure kernel:
Oct  2 14:05:00 secure kernel: e100: eth1 NIC Link is Up 100 Mbps Full duplex
Oct  2 14:05:00 secure network: Bringing up interface eth1:  succeeded

But at the first reboot the entry is placed once more in the /etc/modules.conf and the network fails to start.

I tried to disable the USB in the BIOS and reinstall SME-server, that didn't help.
The CD I used to install from worked OK on another server.  And KNOPPIX 3.2 on the Dell Poweredge works perfect.

Does anyone have some clue how I can fix this? My guess is that there is some template responsible for this.

I *did* search the forums extensibly for this, but came up with nothing.

I am totally stuck here, please help...
Kind greetings,
Thony

[Thony de Waal]

Hello all,
I haven't been waiting for an answer.
I found the a clue on Darrel May's howto for SCSI tape units.

lsmod on the console gave me a list of modules.
I removed the usb modules with the commandline:
rmmod ehci-hcd
rmmod uhci-hcd
removed all the usb-related entries from modules.conf.
rebooted

Now the list of modules is much greater, and the network now networks

Kind greetings,
Thony

[Charlie Brady]

> I *did* search the forums extensibly for this, but came up with nothing.

http://www.e-smith.org/bboard/search.php?search=6.0b3+smebugs

Charlie

[Thony de Waal]

Hi Charlie
You probably mean to say that I should have reported it as a bug to the fine people at Mitel.
That is what I would have done if there was no answer in the form of a "do this and that and reboot" in the coming day. I did not even know if it was a problem specific to the beta version.
I am a Windows-programs application manager, and in that capacity I find that most of these things come from doing things wrong rather than that they are bugs.
I am also not a paying customer, and would not want to stop Mitel from doing worthier things.
In addition, I find that the forum works better when problems are posted here as well, in that case others will benefit from it before a new release is launched. Or be warned not to try to test the beta until a problem is defined and a workaround available.

What I tried to find on the forums here and elsewhere is a reference to the specific usb problem, and that search didn't come up with an answer.

That said, thanks for the fine software and your efforts on the forum!

Kind greetings,
Thony
Charlie Brady wrote:
>
> > I *did* search the forums extensibly for this, but came up
> with nothing.
>
> http://www.e-smith.org/bboard/search.php?search=6.0b3+smebugs
>
> Charlie

[Charlie Brady]

Thony de Waal wrote:

> You probably mean to say that I should have reported it as a
> bug to the fine people at Mitel.

Indeed I do.

> That is what I would have done if there was no answer in the
> form of a "do this and that and reboot" in the coming day. I
> did not even know if it was a problem specific to the beta
> version.

That doesn't matter to us. We want all things which don't work as they should to be reported, and the sooner the better.

> I am a Windows-programs application manager, and in that
> capacity I find that most of these things come from doing
> things wrong rather than that they are bugs.

Clearly that's not the case here.

> In addition, I find that the forum works better when problems
> are posted here as well, in that case others will benefit
> from it before a new release is launched.

We don't mind you posting to the forum *after* reporting a problem.

> Or be warned not to
> try to test the beta until a problem is defined and a
> workaround available.

See comments from earlier today about not testing betas. If (lots of) people don't test them, then there's no point in them being released.

Please, as has been said many many times, report all problems to smebugs@mitel.com. Please report your problem and the fix that you have found. I haven't done it for you.

Charlie

[Thony de Waal]

Hi Charlie,
I have emailed the case to Mitel a minute ago.
Thanks again,
Thony