Re: [Testbed-admins] TMCC exited with status 256! *** WARNING: Could not get bossinfo from tmcc!

[Ben Burnett <bburnett@atcorp.com>]

Leigh,
Boss was unable to resolve pc1.hwdown.emulab-ops.emulab.naresky so I
ran named_setup and then the ping succeeded to the address.

Next we rebooted PC1 and was able to log in successfully with ALL the
daemons and services (including eth99). The only errors on boot were:

cp: cannot stat '/proj/emulab-ops/exp/hwdown/tbdata/topomap.gz': No
such file or directory
cp: cannot stat '/proj/emulab-ops/exp/hwdown/tbdata/ltmap.gz': No such
file or directory

Looks like we're taking steps in the right direction. The experimental
interfaces still do not load on PC1. I am trying to modprobe the
driver from Redhat, but this could be a faux pas considering the
Redhat is different kernel version from PC1.

Can someone point me to the wiki page (it was missed in the last
post...)? I'm not positive that I have seen it.

it's looking like a kernel update might be in order to get this driver going.


Ben Burnett
  R&D Software Engineer: wired, wireless, ad hoc, mesh networking,
    network monitoring and data visualization
  ATC (Architecture Technology Corporation)   http://www.atcorp.com/
  bburnett@atcorp.com                         (952) 829-5864 x167
  --------Views expressed are mine and not those of ATC--------
  Personal:
    ben@burnett.ws                            http://www.burnett.ws/
http://ben.burnett.ws/
http://www.linkedin.com/in/benburnett



On Wed, Jul 15, 2009 at 3:04 PM, Leigh Stoller<stoller@flux.utah.edu> wrote:
>> On FreeBSD, the control interface is bce1 and the experimental
>> interfaces are em0-3.
>>
>> On Fedora we can only see two interfaces, eth0 and eth1. After
>> manually running ifcfg-eth99, the control interface is resolved with
>> an IP address on eth0.  Running eth99 on startup causes problems
>> logging on as previously mentioned, and this might be a domain
>> issue.
>
> Hmm, good point. First off, I assume that you have only a single DHCP
> server (boss) on the network, right.
>
> Next, make sure that pc1.hwdown.emulab-ops.emulab.naresky resolves
> on boss (ping it and see it resolves to an IP). If not, then run
> named_setup on boss.
>
>> There seems to be an inconsistency between FreeBSD and Fedora, as
>> the control interface is bce1 on FreeBSD (making it the second
>> interface) and is eth0 on Fedora (making it the first interface).
>
> This is okay; we do not rely on the ordering the OS assigns.
>
>> Fedora will boot as long at the eth99 script is not started on boot.
>>
>> FreeBSD and Fedora (after manually running eth99) will resolve an IP
>> address for the control network with no troubles. FreeBSD can load the
>> experimental interfaces, but Fedora is unable to see the experimental
>> interfaces. Could this be a driver issue (or lack of a driver) for the
>> experimental NICs in Fedora?
>
> Almost certainly a driver issue ... you might need to add to that
> modprobe line in ifcfg-eth99 and then recreate the image. I'll point
> you to the wiki page that describes how to do that, if you have not
> seen it already.
>
>> can we setup the DHCP setting on the experimental PC's to use a
>> different NIC for the DHCP?
>
> Not sure what you mean here; you want only the control interface (the
> thing wired to the control switch) to get a DHCP response.
>
> Lbs
>
>
>