From Bugzilla Helper:
User-Agent: Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.0; T312461; Q312461)
Description of problem:
The RHL documents don't explain how to configure the /etc/sysconfig/network-
scripts/ifcfg-eth? files to override the MAC address of network adapters. The
scripts seem not to work properly but, without documentation, I can't be sure
that I'm modifying them correctly. It's necessary to be able to override MAC
addreses because some major manufacturers of network adapters seem to be unable
to avoid assigning duplicate MAC addresses. The systems in question are servers
and therefore don't have X installed; so, using netconfig isn't an attractive
option. Moreover, I see a bug report claiming that netconfig doesn't seem to
work right either.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
Search documentation for explanation of /etc/sysconfig/network-scripts/ifcfg-
eth? and how those scripts are used by /etc/sysconfig/network-scripts/ifup
Actual Results: I found no explanation. I then proceeded to debug the related
scripts, without succeeding in understanding how they SHOULD work in this
situation. They appear NOT to work. But, without documentation, I can't draw
any reliable conclusion in that regard.
Expected Results: I expected to find a list of the environment variables used
in these scripts and the functions associated with those variables, including
examples for common situations, including the need to override a MAC address.
Workaround is to add an ifconfig command to /etc/rc.d/rc.sysinit. But, that's
not how this should work, is it? Based on my study of
the /etc/sysconfig/network-scripts/ifup script, it is clearly intended to
handle overriding of MAC addresses.
Since that script is part of the initscripts package, I filed this bug report
against initscripts, rather than some more obviously related component.
will change the MAC address
Well, my hope was that specifying MACADDR would change the MAC address.
However, the code tried to match the specified value with the current actual
value and terminated when it found the mismatch.
So, to be frank, I expect that a bug lurks here. But, given the absence of
documentation, I thought it best to verify exactly how the scripts should work
before assuming that my way -- and ChrisMcc's -- was correct. At the least, I
suspect something more is necessary than merely specifying MACADDR.
Closing bugs on older, no longer supported, releases. Apologies for any lack of
If this persists on a current release, such as Fedora Core 4, please open a new bug.