Hello everyone:
I was upgrading my regression testing server, and bumping the version of
guests to RHEL 6.4. I started getting massive failures in tests, because
the main network interface could never get up. The following is from a
serial session when virt-test tried to repair the problem:
2014-04-11 15:55:29: [root@localhost ~]#
2014-04-11 15:55:30: ifconfig -a
2014-04-11 15:55:30: lo Link encap:Local Loopback
2014-04-11 15:55:30: inet addr:127.0.0.1 Mask:255.0.0.0
2014-04-11 15:55:30: inet6 addr: ::1/128 Scope:Host
2014-04-11 15:55:30: UP LOOPBACK RUNNING MTU:16436 Metric:1
2014-04-11 15:55:30: RX packets:4 errors:0 dropped:0
overruns:0 frame:0
2014-04-11 15:55:30: TX packets:4 errors:0 dropped:0
overruns:0 carrier:0
2014-04-11 15:55:30: collisions:0 txqueuelen:0
2014-04-11 15:55:30: RX bytes:240 (240.0 b) TX bytes:240
(240.0 b)
After looking at the problem, it seems to trace back to the following
lines in the *general kickstart of all RHEL6 guests*:
15:21:51 DEBUG| sed -i "/^HWADDR/d"
/etc/sysconfig/network-scripts/ifcfg-eth0
That leaves the eth0 related file invalid so the interface can't go up.
I'm downloading the 6.4 image to see if I can reproduce, but in the mean
time, did you ever see anything similar? We need to start having
different kickstarts for different versions of RHEL. The current one is
already super bloated, the base kickstart should lead to a lean guest
install.
As I don't have a lot of time to work on virt-test maintenance these
days, I'll have to revert back to 6.3.
Thanks,
Lucas
_______________________________________________
Virt-test-devel mailing list
[email protected]
https://www.redhat.com/mailman/listinfo/virt-test-devel
