Red Hat Bugzilla – Bug 682365
assumes all ethernet devices are named ethX
Last modified: 2014-03-16 23:26:41 EDT
Description of problem:
Network devices can have arbitrary names, and due to
http://fedoraproject.org/wiki/Features/ConsistentNetworkDeviceNaming, will have
different names in Fedora 15.
strnmov(ifr.ifr_name, "eth0", sizeof(ifr.ifr_name) - 1);
if (ioctl(fd, SIOCGIFHWADDR, &ifr) >= 0)
memcpy(to, &ifr.ifr_hwaddr.sa_data, ETHER_ADDR_LEN);
res= memcmp(to, zero_array, ETHER_ADDR_LEN) ? 0 : 1;
} while (res && (errno == 0 || errno == ENODEV) && ifr.ifr_name++ < '6');
Now, since this is only used to initialize the UUID (afaict), and there is fallback code for that, it may not be critical to fix this. Filing for completeness.
Version-Release number of selected component (if applicable):
[ raised eyebrow... ] Changing that naming convention is going to break boatloads of stuff, much of it user-written scripts that we couldn't fix if we wanted to. You *sure* this is a good idea?
Tom: yes. The feature page describes how to disable the new naming convention, but the approved feature is enabled by default in F15.
Created attachment 529924 [details]
proposed patch that creates a "sum" of all interfaces
Since hw address is needed only for initiate a seed for pseudo-random number generator, we can return a value, that is unique enough to the machine, rather then the mac address. This patch loops through all interfaces and creates byte-by-byte sum of all devices.
If real hw address should be returned, we'd need find out if the interface is a real one or just a virtual (loopback, virtual bridge, ...). One way (not sure how reliable it is) is compare the address with 00:00:00:00:00:00, since virtual interfaces seems to have zero address.
Created attachment 531117 [details]
proposed patch that takes the first non-zero HW address
Since there is a comment in mysys/my_gethwaddr.c, that we can use any of non-zero HW address, this patch fixes the issue that way.
The patch has also been sent to upstream: http://bugs.mysql.com/bug.php?id=63055
Since this bug is against F15, I was planning to keep it open until the patch can be back-patched. That's waiting on some other stuff though ...
mysql-5.5.18-1.fc15 has been submitted as an update for Fedora 15.
* should fix your issue,
* was pushed to the Fedora 15 testing repository,
* should be available at your local mirror within two days.
Update it with:
# su -c 'yum update --enablerepo=updates-testing mysql-5.5.18-1.fc15'
as soon as you are able to.
Please go to the following url:
then log in and leave karma (feedback).
mysql-5.5.18-1.fc15 has been pushed to the Fedora 15 stable repository. If problems still persist, please make note of it in this bug report.